5 Changes between 1.0.x and 1.1.0 [xx XXX xxxx]
7 *) New option -brief for s_client and s_server to print out a brief summary
8 of connection parameters.
11 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
15 *) New Suite B modes for TLS code. These use and enforce the requirements
16 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
17 only use Suite B curves. The Suite B modes can be set by using the
18 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
21 *) New chain verification flags for Suite B levels of security. Check
22 algorithms are acceptable when flags are set in X509_verify_cert.
25 *) Make tls1_check_chain return a set of flags indicating checks passed
26 by a certificate chain. Add additional tests to handle client
27 certificates: checks for matching certificate type and issuer name
31 *) If an attempt is made to use a signature algorithm not in the peer
32 preference list abort the handshake. If client has no suitable
33 signature algorithms in response to a certificate request do not
37 *) If server EC tmp key is not in client preference list abort handshake.
40 *) Add support for certificate stores in CERT structure. This makes it
41 possible to have different stores per SSL structure or one store in
42 the parent SSL_CTX. Include distint stores for certificate chain
43 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
44 to build and store a certificate chain in CERT structure: returing
45 an error if the chain cannot be built: this will allow applications
46 to test if a chain is correctly configured.
48 Note: if the CERT based stores are not set then the parent SSL_CTX
49 store is used to retain compatibility with existing behaviour.
53 *) New function ssl_set_client_disabled to set a ciphersuite disabled
54 mask based on the current session, check mask when sending client
55 hello and checking the requested ciphersuite.
58 *) New ctrls to retrieve and set certificate types in a certificate
59 request message. Print out received values in s_client. If certificate
60 types is not set with custom values set sensible values based on
61 supported signature algorithms.
64 *) Support for distinct client and server supported signature algorithms.
67 *) Add certificate callback. If set this is called whenever a certificate
68 is required by client or server. An application can decide which
69 certificate chain to present based on arbitrary criteria: for example
70 supported signature algorithms. Add very simple example to s_server.
71 This fixes many of the problems and restrictions of the existing client
72 certificate callback: for example you can now clear an existing
73 certificate and specify the whole chain.
76 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
77 the certificate can be used for (if anything). Set valid_flags field
78 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
79 to have similar checks in it.
81 Add new "cert_flags" field to CERT structure and include a "strict mode".
82 This enforces some TLS certificate requirements (such as only permitting
83 certificate signature algorithms contained in the supported algorithms
84 extension) which some implementations ignore: this option should be used
85 with caution as it could cause interoperability issues.
88 *) Update and tidy signature algorithm extension processing. Work out
89 shared signature algorithms based on preferences and peer algorithms
90 and print them out in s_client and s_server. Abort handshake if no
91 shared signature algorithms.
94 *) Add new functions to allow customised supported signature algorithms
95 for SSL and SSL_CTX structures. Add options to s_client and s_server
99 *) New function SSL_certs_clear() to delete all references to certificates
100 from an SSL structure. Before this once a certificate had been added
101 it couldn't be removed.
104 *) Initial SSL tracing code. This parses out SSL/TLS records using the
105 message callback and prints the results. Needs compile time option
106 "enable-ssl-trace". New options to s_client and s_server to enable
110 *) New functions to retrieve certificate signature and signature
114 *) Print out deprecated issuer and subject unique ID fields in
118 *) Update fips_test_suite to support multiple command line options. New
119 test to induce all self test errors in sequence and check expected
123 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
124 sign or verify all in one operation.
127 *) Add fips_algvs: a multicall fips utility incorporaing all the algorithm
128 test programs and fips_test_suite. Includes functionality to parse
129 the minimal script output of fipsalgest.pl directly.
132 *) Add authorisation parameter to FIPS_module_mode_set().
135 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
138 *) Use separate DRBG fields for internal and external flags. New function
139 FIPS_drbg_health_check() to perform on demand health checking. Add
140 generation tests to fips_test_suite with reduced health check interval to
141 demonstrate periodic health checking. Add "nodh" option to
142 fips_test_suite to skip very slow DH test.
145 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
149 *) More extensive health check for DRBG checking many more failure modes.
150 New function FIPS_selftest_drbg_all() to handle every possible DRBG
151 combination: call this in fips_test_suite.
154 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
155 and POST to handle Dual EC cases.
158 *) Add support for canonical generation of DSA parameter 'g'. See
161 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
162 POST to handle HMAC cases.
165 *) Add functions FIPS_module_version() and FIPS_module_version_text()
166 to return numberical and string versions of the FIPS module number.
169 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
170 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implmeneted
171 outside the validated module in the FIPS capable OpenSSL.
174 *) Minor change to DRBG entropy callback semantics. In some cases
175 there is no mutiple of the block length between min_len and
176 max_len. Allow the callback to return more than max_len bytes
177 of entropy but discard any extra: it is the callback's responsibility
178 to ensure that the extra data discarded does not impact the
179 requested amount of entropy.
182 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
183 information in FIPS186-3, SP800-57 and SP800-131A.
186 *) CCM support via EVP. Interface is very similar to GCM case except we
187 must supply all data in one chunk (i.e. no update, final) and the
188 message length must be supplied if AAD is used. Add algorithm test
192 *) Initial version of POST overhaul. Add POST callback to allow the status
193 of POST to be monitored and/or failures induced. Modify fips_test_suite
194 to use callback. Always run all selftests even if one fails.
197 *) XTS support including algorithm test driver in the fips_gcmtest program.
198 Note: this does increase the maximum key length from 32 to 64 bytes but
199 there should be no binary compatibility issues as existing applications
200 will never use XTS mode.
203 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
204 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
205 performs algorithm blocking for unapproved PRNG types. Also do not
206 set PRNG type in FIPS_mode_set(): leave this to the application.
207 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
208 the standard OpenSSL PRNG: set additional data to a date time vector.
211 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
212 This shouldn't present any incompatibility problems because applications
213 shouldn't be using these directly and any that are will need to rethink
214 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
217 *) Extensive self tests and health checking required by SP800-90 DRBG.
218 Remove strength parameter from FIPS_drbg_instantiate and always
219 instantiate at maximum supported strength.
222 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
225 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
228 *) New function DH_compute_key_padded() to compute a DH key and pad with
229 leading zeroes if needed: this complies with SP800-56A et al.
232 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
233 anything, incomplete, subject to change and largely untested at present.
236 *) Modify fipscanisteronly build option to only build the necessary object
237 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
240 *) Add experimental option FIPSSYMS to give all symbols in
241 fipscanister.o and FIPS or fips prefix. This will avoid
242 conflicts with future versions of OpenSSL. Add perl script
243 util/fipsas.pl to preprocess assembly language source files
244 and rename any affected symbols.
247 *) Add selftest checks and algorithm block of non-fips algorithms in
248 FIPS mode. Remove DES2 from selftests.
251 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
252 return internal method without any ENGINE dependencies. Add new
253 tiny fips sign and verify functions.
256 *) New build option no-ec2m to disable characteristic 2 code.
259 *) New build option "fipscanisteronly". This only builds fipscanister.o
260 and (currently) associated fips utilities. Uses the file Makefile.fips
261 instead of Makefile.org as the prototype.
264 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
265 Update fips_gcmtest to use IV generator.
268 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
269 setting output buffer to NULL. The *Final function must be
270 called although it will not retrieve any additional data. The tag
271 can be set or retrieved with a ctrl. The IV length is by default 12
272 bytes (96 bits) but can be set to an alternative value. If the IV
273 length exceeds the maximum IV length (currently 16 bytes) it cannot be
277 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
278 underlying do_cipher function handles all cipher semantics itself
279 including padding and finalisation. This is useful if (for example)
280 an ENGINE cipher handles block padding itself. The behaviour of
281 do_cipher is subtly changed if this flag is set: the return value
282 is the number of characters written to the output buffer (zero is
283 no longer an error code) or a negative error code. Also if the
284 input buffer is NULL and length 0 finalisation should be performed.
287 *) If a candidate issuer certificate is already part of the constructed
288 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
291 *) Improve forward-security support: add functions
293 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
294 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
296 for use by SSL/TLS servers; the callback function will be called whenever a
297 new session is created, and gets to decide whether the session may be
298 cached to make it resumable (return 0) or not (return 1). (As by the
299 SSL/TLS protocol specifications, the session_id sent by the server will be
300 empty to indicate that the session is not resumable; also, the server will
301 not generate RFC 4507 (RFC 5077) session tickets.)
303 A simple reasonable callback implementation is to return is_forward_secure.
304 This parameter will be set to 1 or 0 depending on the ciphersuite selected
305 by the SSL/TLS server library, indicating whether it can provide forward
307 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
309 *) New function OPENSSL_gmtime_diff to find the difference in days
310 and seconds between two tm structures. This will be used to provide
311 additional functionality for ASN1_TIME.
314 *) Add -trusted_first option which attempts to find certificates in the
315 trusted store even if an untrusted chain is also supplied.
318 *) Initial experimental support for explicitly trusted non-root CAs.
319 OpenSSL still tries to build a complete chain to a root but if an
320 intermediate CA has a trust setting included that is used. The first
321 setting is used: whether to trust or reject.
324 *) New -verify_name option in command line utilities to set verification
328 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
329 Add CMAC pkey methods.
332 *) Experimental regnegotiation in s_server -www mode. If the client
333 browses /reneg connection is renegotiated. If /renegcert it is
334 renegotiated requesting a certificate.
337 *) Add an "external" session cache for debugging purposes to s_server. This
338 should help trace issues which normally are only apparent in deployed
339 multi-process servers.
342 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
343 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
344 BIO_set_cipher() and some obscure PEM functions were changed so they
345 can now return an error. The RAND changes required a change to the
346 RAND_METHOD structure.
349 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
350 a gcc attribute to warn if the result of a function is ignored. This
351 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
352 whose return value is often ignored.
355 Changes between 1.0.1 and 1.0.2 [xx XXX xxxx]
357 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
358 the right response is stapled. Also change current certificate to
359 the certificate actually sent.
360 See http://rt.openssl.org/Ticket/Display.html?id=2836.
361 [Rob Stradling <rob.stradling@comodo.com>]
363 *) Support for linux-x32, ILP32 environment in x86_64 framework.
367 [Emilia Kasper, Adam Langley, Ben Laurie (Google)]
369 *) Experimental multi-implementation support for FIPS capable OpenSSL.
370 When in FIPS mode the approved implementations are used as normal,
371 when not in FIPS mode the internal unapproved versions are used instead.
372 This means that the FIPS capable OpenSSL isn't forced to use the
373 (often lower perfomance) FIPS implementations outside FIPS mode.
376 *) Transparently support X9.42 DH parameters when calling
377 PEM_read_bio_DHparameters. This means existing applications can handle
378 the new parameter format automatically.
381 *) Initial experimental support for X9.42 DH parameter format: mainly
382 to support use of 'q' parameter for RFC5114 parameters.
385 *) Add DH parameters from RFC5114 including test data to dhtest.
388 *) Support for automatic EC temporary key parameter selection. If enabled
389 the most preferred EC parameters are automatically used instead of
390 hardcoded fixed parameters. Now a server just has to call:
391 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
392 support ECDH and use the most appropriate parameters.
395 *) Enhance and tidy EC curve and point format TLS extension code. Use
396 static structures instead of allocation if default values are used.
397 New ctrls to set curves we wish to support and to retrieve shared curves.
398 Print out shared curves in s_server. New options to s_server and s_client
399 to set list of supported curves.
402 *) New ctrls to retrieve supported signature algorithms and
403 supported curve values as an array of NIDs. Extend openssl utility
404 to print out received values.
407 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
408 between NIDs and the more common NIST names such as "P-256". Enhance
409 ecparam utility and ECC method to recognise the NIST names for curves.
412 *) Enhance SSL/TLS certificate chain handling to support different
413 chains for each certificate instead of one chain in the parent SSL_CTX.
416 *) Support for fixed DH ciphersuite client authentication: where both
417 server and client use DH certificates with common parameters.
420 *) Support for fixed DH ciphersuites: those requiring DH server
424 Changes between 1.0.1c and 1.0.1d [xx XXX xxxx]
426 *) Fix possible deadlock when decoding public keys.
429 *) Don't use TLS 1.0 record version number in initial client hello
433 Changes between 1.0.1b and 1.0.1c [10 May 2012]
435 *) Sanity check record length before skipping explicit IV in TLS
436 1.2, 1.1 and DTLS to avoid DoS attack.
438 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
439 fuzzing as a service testing platform.
443 *) Initialise tkeylen properly when encrypting CMS messages.
444 Thanks to Solar Designer of Openwall for reporting this issue.
447 *) In FIPS mode don't try to use composite ciphers as they are not
451 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
453 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
454 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
455 mean any application compiled against OpenSSL 1.0.0 headers setting
456 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
457 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
458 0x10000000L Any application which was previously compiled against
459 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
460 will need to be recompiled as a result. Letting be results in
461 inability to disable specifically TLS 1.1 and in client context,
462 in unlike event, limit maximum offered version to TLS 1.0 [see below].
465 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
466 disable just protocol X, but all protocols above X *if* there are
467 protocols *below* X still enabled. In more practical terms it means
468 that if application wants to disable TLS1.0 in favor of TLS1.1 and
469 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
470 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
474 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
476 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
477 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
478 in CRYPTO_realloc_clean.
480 Thanks to Tavis Ormandy, Google Security Team, for discovering this
481 issue and to Adam Langley <agl@chromium.org> for fixing it.
483 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
485 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
488 *) Workarounds for some broken servers that "hang" if a client hello
489 record length exceeds 255 bytes:
491 1. Do not use record version number > TLS 1.0 in initial client
492 hello: some (but not all) hanging servers will now work.
493 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
494 the number of ciphers sent in the client hello. This should be
495 set to an even number, such as 50, for example by passing:
496 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
497 Most broken servers should now work.
498 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
499 TLS 1.2 client support entirely.
502 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
505 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
507 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
508 STRING form instead of a DigestInfo.
511 *) The format used for MDC2 RSA signatures is inconsistent between EVP
512 and the RSA_sign/RSA_verify functions. This was made more apparent when
513 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
514 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
515 the correct format in RSA_verify so both forms transparently work.
518 *) Some servers which support TLS 1.0 can choke if we initially indicate
519 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
520 encrypted premaster secret. As a workaround use the maximum pemitted
521 client version in client hello, this should keep such servers happy
522 and still work with previous versions of OpenSSL.
525 *) Add support for TLS/DTLS heartbeats.
526 [Robin Seggelmann <seggelmann@fh-muenster.de>]
528 *) Add support for SCTP.
529 [Robin Seggelmann <seggelmann@fh-muenster.de>]
531 *) Improved PRNG seeding for VOS.
532 [Paul Green <Paul.Green@stratus.com>]
534 *) Extensive assembler packs updates, most notably:
536 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
537 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
538 - x86_64: bit-sliced AES implementation;
539 - ARM: NEON support, contemporary platforms optimizations;
540 - s390x: z196 support;
541 - *: GHASH and GF(2^m) multiplication implementations;
545 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
546 (removal of unnecessary code)
547 [Peter Sylvester <peter.sylvester@edelweb.fr>]
549 *) Add TLS key material exporter from RFC 5705.
552 *) Add DTLS-SRTP negotiation from RFC 5764.
555 *) Add Next Protocol Negotiation,
556 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
557 disabled with a no-npn flag to config or Configure. Code donated
559 [Adam Langley <agl@google.com> and Ben Laurie]
561 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
562 NIST-P256, NIST-P521, with constant-time single point multiplication on
563 typical inputs. Compiler support for the nonstandard type __uint128_t is
564 required to use this (present in gcc 4.4 and later, for 64-bit builds).
565 Code made available under Apache License version 2.0.
567 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
568 line to include this in your build of OpenSSL, and run "make depend" (or
569 "make update"). This enables the following EC_METHODs:
571 EC_GFp_nistp224_method()
572 EC_GFp_nistp256_method()
573 EC_GFp_nistp521_method()
575 EC_GROUP_new_by_curve_name() will automatically use these (while
576 EC_GROUP_new_curve_GFp() currently prefers the more flexible
578 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
580 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
581 all platforms. Move ssize_t definition from e_os.h to the public
582 header file e_os2.h as it now appears in public header file cms.h
585 *) New -sigopt option to the ca, req and x509 utilities. Additional
586 signature parameters can be passed using this option and in
590 *) Add RSA PSS signing function. This will generate and set the
591 appropriate AlgorithmIdentifiers for PSS based on those in the
592 corresponding EVP_MD_CTX structure. No application support yet.
595 *) Support for companion algorithm specific ASN1 signing routines.
596 New function ASN1_item_sign_ctx() signs a pre-initialised
597 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
598 the appropriate parameters.
601 *) Add new algorithm specific ASN1 verification initialisation function
602 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
603 handling will be the same no matter what EVP_PKEY_METHOD is used.
604 Add a PSS handler to support verification of PSS signatures: checked
605 against a number of sample certificates.
608 *) Add signature printing for PSS. Add PSS OIDs.
609 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
611 *) Add algorithm specific signature printing. An individual ASN1 method
612 can now print out signatures instead of the standard hex dump.
614 More complex signatures (e.g. PSS) can print out more meaningful
615 information. Include DSA version that prints out the signature
619 *) Password based recipient info support for CMS library: implementing
623 *) Split password based encryption into PBES2 and PBKDF2 functions. This
624 neatly separates the code into cipher and PBE sections and is required
625 for some algorithms that split PBES2 into separate pieces (such as
629 *) Session-handling fixes:
630 - Fix handling of connections that are resuming with a session ID,
631 but also support Session Tickets.
632 - Fix a bug that suppressed issuing of a new ticket if the client
633 presented a ticket with an expired session.
634 - Try to set the ticket lifetime hint to something reasonable.
635 - Make tickets shorter by excluding irrelevant information.
636 - On the client side, don't ignore renewed tickets.
637 [Adam Langley, Bodo Moeller (Google)]
639 *) Fix PSK session representation.
642 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
644 This work was sponsored by Intel.
647 *) Add GCM support to TLS library. Some custom code is needed to split
648 the IV between the fixed (from PRF) and explicit (from TLS record)
649 portions. This adds all GCM ciphersuites supported by RFC5288 and
650 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
651 add a special AESGCM string for GCM only.
654 *) Expand range of ctrls for AES GCM. Permit setting invocation
655 field on decrypt and retrieval of invocation field only on encrypt.
658 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
659 As required by RFC5289 these ciphersuites cannot be used if for
660 versions of TLS earlier than 1.2.
663 *) For FIPS capable OpenSSL interpret a NULL default public key method
664 as unset and return the appopriate default but do *not* set the default.
665 This means we can return the appopriate method in applications that
666 swicth between FIPS and non-FIPS modes.
669 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
670 ENGINE is used then we cannot handle that in the FIPS module so we
671 keep original code iff non-FIPS operations are allowed.
674 *) Add -attime option to openssl utilities.
675 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
677 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
680 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
681 FIPS EC methods unconditionally for now.
684 *) New build option no-ec2m to disable characteristic 2 code.
687 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
688 all cases can be covered as some introduce binary incompatibilities.
691 *) Redirect RSA operations to FIPS module including keygen,
692 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
695 *) Add similar low level API blocking to ciphers.
698 *) Low level digest APIs are not approved in FIPS mode: any attempt
699 to use these will cause a fatal error. Applications that *really* want
700 to use them can use the private_* version instead.
703 *) Redirect cipher operations to FIPS module for FIPS builds.
706 *) Redirect digest operations to FIPS module for FIPS builds.
709 *) Update build system to add "fips" flag which will link in fipscanister.o
710 for static and shared library builds embedding a signature if needed.
713 *) Output TLS supported curves in preference order instead of numerical
714 order. This is currently hardcoded for the highest order curves first.
715 This should be configurable so applications can judge speed vs strength.
718 *) Add TLS v1.2 server support for client authentication.
721 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
725 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
726 FIPS modules versions.
729 *) Add TLS v1.2 client side support for client authentication. Keep cache
730 of handshake records longer as we don't know the hash algorithm to use
731 until after the certificate request message is received.
734 *) Initial TLS v1.2 client support. Add a default signature algorithms
735 extension including all the algorithms we support. Parse new signature
736 format in client key exchange. Relax some ECC signing restrictions for
737 TLS v1.2 as indicated in RFC5246.
740 *) Add server support for TLS v1.2 signature algorithms extension. Switch
741 to new signature format when needed using client digest preference.
742 All server ciphersuites should now work correctly in TLS v1.2. No client
743 support yet and no support for client certificates.
746 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
747 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
748 ciphersuites. At present only RSA key exchange ciphersuites work with
749 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
750 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
751 and version checking.
754 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
755 with this defined it will not be affected by any changes to ssl internal
756 structures. Add several utility functions to allow openssl application
757 to work with OPENSSL_NO_SSL_INTERN defined.
761 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
763 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
766 *) Permit abbreviated handshakes when renegotiating using the function
767 SSL_renegotiate_abbreviated().
768 [Robin Seggelmann <seggelmann@fh-muenster.de>]
770 *) Add call to ENGINE_register_all_complete() to
771 ENGINE_load_builtin_engines(), so some implementations get used
772 automatically instead of needing explicit application support.
775 *) Add support for TLS key exporter as described in RFC5705.
776 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
778 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
779 a few changes are required:
781 Add SSL_OP_NO_TLSv1_1 flag.
783 Update version checking logic to handle version 1.1.
784 Add explicit IV handling (ported from DTLS code).
785 Add command line options to s_client/s_server.
788 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
790 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
791 in CMS and PKCS7 code. When RSA decryption fails use a random key for
792 content decryption and always return the same error. Note: this attack
793 needs on average 2^20 messages so it only affects automated senders. The
794 old behaviour can be reenabled in the CMS code by setting the
795 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
796 an MMA defence is not necessary.
797 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
798 this issue. (CVE-2012-0884)
801 *) Fix CVE-2011-4619: make sure we really are receiving a
802 client hello before rejecting multiple SGC restarts. Thanks to
803 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
806 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
808 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
809 Thanks to Antonio Martin, Enterprise Secure Access Research and
810 Development, Cisco Systems, Inc. for discovering this bug and
811 preparing a fix. (CVE-2012-0050)
814 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
816 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
817 of the Vaudenay padding oracle attack on CBC mode encryption
818 which enables an efficient plaintext recovery attack against
819 the OpenSSL implementation of DTLS. Their attack exploits timing
820 differences arising during decryption processing. A research
821 paper describing this attack can be found at:
822 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
823 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
824 Security Group at Royal Holloway, University of London
825 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
826 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
827 for preparing the fix. (CVE-2011-4108)
828 [Robin Seggelmann, Michael Tuexen]
830 *) Clear bytes used for block padding of SSL 3.0 records.
832 [Adam Langley (Google)]
834 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
835 Kadianakis <desnacked@gmail.com> for discovering this issue and
836 Adam Langley for preparing the fix. (CVE-2011-4619)
837 [Adam Langley (Google)]
839 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
840 [Andrey Kulikov <amdeich@gmail.com>]
842 *) Prevent malformed RFC3779 data triggering an assertion failure.
843 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
844 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
845 [Rob Austein <sra@hactrn.net>]
847 *) Improved PRNG seeding for VOS.
848 [Paul Green <Paul.Green@stratus.com>]
850 *) Fix ssl_ciph.c set-up race.
851 [Adam Langley (Google)]
853 *) Fix spurious failures in ecdsatest.c.
854 [Emilia Käsper (Google)]
856 *) Fix the BIO_f_buffer() implementation (which was mixing different
857 interpretations of the '..._len' fields).
858 [Adam Langley (Google)]
860 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
861 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
862 threads won't reuse the same blinding coefficients.
864 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
865 lock to call BN_BLINDING_invert_ex, and avoids one use of
866 BN_BLINDING_update for each BN_BLINDING structure (previously,
867 the last update always remained unused).
868 [Emilia Käsper (Google)]
870 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
871 [Bob Buckholz (Google)]
873 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
875 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
876 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
877 [Kaspar Brand <ossl@velox.ch>]
879 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
880 for multi-threaded use of ECDH. (CVE-2011-3210)
881 [Adam Langley (Google)]
883 *) Fix x509_name_ex_d2i memory leak on bad inputs.
886 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
887 signature public key algorithm by using OID xref utilities instead.
888 Before this you could only use some ECC ciphersuites with SHA1 only.
891 *) Add protection against ECDSA timing attacks as mentioned in the paper
892 by Billy Bob Brumley and Nicola Tuveri, see:
894 http://eprint.iacr.org/2011/232.pdf
896 [Billy Bob Brumley and Nicola Tuveri]
898 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
900 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
901 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
903 *) Fix bug in string printing code: if *any* escaping is enabled we must
904 escape the escape character (backslash) or the resulting string is
908 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
910 *) Disable code workaround for ancient and obsolete Netscape browsers
911 and servers: an attacker can use it in a ciphersuite downgrade attack.
912 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
915 *) Fixed J-PAKE implementation error, originally discovered by
916 Sebastien Martini, further info and confirmation from Stefan
917 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
920 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
922 *) Fix extension code to avoid race conditions which can result in a buffer
923 overrun vulnerability: resumed sessions must not be modified as they can
924 be shared by multiple threads. CVE-2010-3864
927 *) Fix WIN32 build system to correctly link an ENGINE directory into
931 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
933 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
935 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
937 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
939 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
940 context. The operation can be customised via the ctrl mechanism in
941 case ENGINEs want to include additional functionality.
944 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
947 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
948 output hashes compatible with older versions of OpenSSL.
949 [Willy Weisz <weisz@vcpc.univie.ac.at>]
951 *) Fix compression algorithm handling: if resuming a session use the
952 compression algorithm of the resumed session instead of determining
953 it from client hello again. Don't allow server to change algorithm.
956 *) Add load_crls() function to apps tidying load_certs() too. Add option
957 to verify utility to allow additional CRLs to be included.
960 *) Update OCSP request code to permit adding custom headers to the request:
961 some responders need this.
964 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
966 [Julia Lawall <julia@diku.dk>]
968 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
969 needlessly dereferenced structures, used obsolete functions and
970 didn't handle all updated verify codes correctly.
973 *) Disable MD2 in the default configuration.
976 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
977 indicate the initial BIO being pushed or popped. This makes it possible
978 to determine whether the BIO is the one explicitly called or as a result
979 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
980 it handles reference counts correctly and doesn't zero out the I/O bio
981 when it is not being explicitly popped. WARNING: applications which
982 included workarounds for the old buggy behaviour will need to be modified
983 or they could free up already freed BIOs.
986 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
987 renaming to all platforms (within the 0.9.8 branch, this was
988 done conditionally on Netware platforms to avoid a name clash).
989 [Guenter <lists@gknw.net>]
991 *) Add ECDHE and PSK support to DTLS.
992 [Michael Tuexen <tuexen@fh-muenster.de>]
994 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
998 *) Add "missing" function EVP_MD_flags() (without this the only way to
999 retrieve a digest flags is by accessing the structure directly. Update
1000 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
1001 or cipher is registered as in the "from" argument. Print out all
1002 registered digests in the dgst usage message instead of manually
1003 attempting to work them out.
1006 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
1007 this allows the use of compression and extensions. Change default cipher
1008 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
1009 by default unless an application cipher string requests it.
1012 *) Alter match criteria in PKCS12_parse(). It used to try to use local
1013 key ids to find matching certificates and keys but some PKCS#12 files
1014 don't follow the (somewhat unwritten) rules and this strategy fails.
1015 Now just gather all certificates together and the first private key
1016 then look for the first certificate that matches the key.
1019 *) Support use of registered digest and cipher names for dgst and cipher
1020 commands instead of having to add each one as a special case. So now
1027 openssl dgst -sha256 foo
1029 and this works for ENGINE based algorithms too.
1033 *) Update Gost ENGINE to support parameter files.
1034 [Victor B. Wagner <vitus@cryptocom.ru>]
1036 *) Support GeneralizedTime in ca utility.
1037 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
1039 *) Enhance the hash format used for certificate directory links. The new
1040 form uses the canonical encoding (meaning equivalent names will work
1041 even if they aren't identical) and uses SHA1 instead of MD5. This form
1042 is incompatible with the older format and as a result c_rehash should
1043 be used to rebuild symbolic links.
1046 *) Make PKCS#8 the default write format for private keys, replacing the
1047 traditional format. This form is standardised, more secure and doesn't
1048 include an implicit MD5 dependency.
1051 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
1052 committed to OpenSSL should pass this lot as a minimum.
1055 *) Add session ticket override functionality for use by EAP-FAST.
1056 [Jouni Malinen <j@w1.fi>]
1058 *) Modify HMAC functions to return a value. Since these can be implemented
1059 in an ENGINE errors can occur.
1062 *) Type-checked OBJ_bsearch_ex.
1065 *) Type-checked OBJ_bsearch. Also some constification necessitated
1066 by type-checking. Still to come: TXT_DB, bsearch(?),
1067 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
1071 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
1072 seconds to a tm structure directly, instead of going through OS
1073 specific date routines. This avoids any issues with OS routines such
1074 as the year 2038 bug. New *_adj() functions for ASN1 time structures
1075 and X509_time_adj_ex() to cover the extended range. The existing
1076 X509_time_adj() is still usable and will no longer have any date issues.
1079 *) Delta CRL support. New use deltas option which will attempt to locate
1080 and search any appropriate delta CRLs available.
1082 This work was sponsored by Google.
1085 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
1086 code and add additional score elements. Validate alternate CRL paths
1087 as part of the CRL checking and indicate a new error "CRL path validation
1088 error" in this case. Applications wanting additional details can use
1089 the verify callback and check the new "parent" field. If this is not
1090 NULL CRL path validation is taking place. Existing applications wont
1091 see this because it requires extended CRL support which is off by
1094 This work was sponsored by Google.
1097 *) Support for freshest CRL extension.
1099 This work was sponsored by Google.
1102 *) Initial indirect CRL support. Currently only supported in the CRLs
1103 passed directly and not via lookup. Process certificate issuer
1104 CRL entry extension and lookup CRL entries by bother issuer name
1105 and serial number. Check and process CRL issuer entry in IDP extension.
1107 This work was sponsored by Google.
1110 *) Add support for distinct certificate and CRL paths. The CRL issuer
1111 certificate is validated separately in this case. Only enabled if
1112 an extended CRL support flag is set: this flag will enable additional
1113 CRL functionality in future.
1115 This work was sponsored by Google.
1118 *) Add support for policy mappings extension.
1120 This work was sponsored by Google.
1123 *) Fixes to pathlength constraint, self issued certificate handling,
1124 policy processing to align with RFC3280 and PKITS tests.
1126 This work was sponsored by Google.
1129 *) Support for name constraints certificate extension. DN, email, DNS
1130 and URI types are currently supported.
1132 This work was sponsored by Google.
1135 *) To cater for systems that provide a pointer-based thread ID rather
1136 than numeric, deprecate the current numeric thread ID mechanism and
1137 replace it with a structure and associated callback type. This
1138 mechanism allows a numeric "hash" to be extracted from a thread ID in
1139 either case, and on platforms where pointers are larger than 'long',
1140 mixing is done to help ensure the numeric 'hash' is usable even if it
1141 can't be guaranteed unique. The default mechanism is to use "&errno"
1142 as a pointer-based thread ID to distinguish between threads.
1144 Applications that want to provide their own thread IDs should now use
1145 CRYPTO_THREADID_set_callback() to register a callback that will call
1146 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
1148 Note that ERR_remove_state() is now deprecated, because it is tied
1149 to the assumption that thread IDs are numeric. ERR_remove_state(0)
1150 to free the current thread's error state should be replaced by
1151 ERR_remove_thread_state(NULL).
1153 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
1154 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
1155 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
1156 application was previously providing a numeric thread callback that
1157 was inappropriate for distinguishing threads, then uniqueness might
1158 have been obtained with &errno that happened immediately in the
1159 intermediate development versions of OpenSSL; this is no longer the
1160 case, the numeric thread callback will now override the automatic use
1162 [Geoff Thorpe, with help from Bodo Moeller]
1164 *) Initial support for different CRL issuing certificates. This covers a
1165 simple case where the self issued certificates in the chain exist and
1166 the real CRL issuer is higher in the existing chain.
1168 This work was sponsored by Google.
1171 *) Removed effectively defunct crypto/store from the build.
1174 *) Revamp of STACK to provide stronger type-checking. Still to come:
1175 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
1176 ASN1_STRING, CONF_VALUE.
1179 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
1180 RAM on SSL connections. This option can save about 34k per idle SSL.
1183 *) Revamp of LHASH to provide stronger type-checking. Still to come:
1184 STACK, TXT_DB, bsearch, qsort.
1187 *) Initial support for Cryptographic Message Syntax (aka CMS) based
1188 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
1189 support for data, signedData, compressedData, digestedData and
1190 encryptedData, envelopedData types included. Scripts to check against
1191 RFC4134 examples draft and interop and consistency checks of many
1192 content types and variants.
1195 *) Add options to enc utility to support use of zlib compression BIO.
1198 *) Extend mk1mf to support importing of options and assembly language
1199 files from Configure script, currently only included in VC-WIN32.
1200 The assembly language rules can now optionally generate the source
1201 files from the associated perl scripts.
1204 *) Implement remaining functionality needed to support GOST ciphersuites.
1205 Interop testing has been performed using CryptoPro implementations.
1206 [Victor B. Wagner <vitus@cryptocom.ru>]
1208 *) s390x assembler pack.
1211 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
1215 *) Implement Opaque PRF Input TLS extension as specified in
1216 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
1217 official specification yet and no extension type assignment by
1218 IANA exists, this extension (for now) will have to be explicitly
1219 enabled when building OpenSSL by providing the extension number
1220 to use. For example, specify an option
1222 -DTLSEXT_TYPE_opaque_prf_input=0x9527
1224 to the "config" or "Configure" script to enable the extension,
1225 assuming extension number 0x9527 (which is a completely arbitrary
1226 and unofficial assignment based on the MD5 hash of the Internet
1227 Draft). Note that by doing so, you potentially lose
1228 interoperability with other TLS implementations since these might
1229 be using the same extension number for other purposes.
1231 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
1232 opaque PRF input value to use in the handshake. This will create
1233 an interal copy of the length-'len' string at 'src', and will
1234 return non-zero for success.
1236 To get more control and flexibility, provide a callback function
1239 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
1240 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
1244 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
1247 Callback function 'cb' will be called in handshakes, and is
1248 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
1249 Argument 'arg' is for application purposes (the value as given to
1250 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
1251 be provided to the callback function). The callback function
1252 has to return non-zero to report success: usually 1 to use opaque
1253 PRF input just if possible, or 2 to enforce use of the opaque PRF
1254 input. In the latter case, the library will abort the handshake
1255 if opaque PRF input is not successfully negotiated.
1257 Arguments 'peerinput' and 'len' given to the callback function
1258 will always be NULL and 0 in the case of a client. A server will
1259 see the client's opaque PRF input through these variables if
1260 available (NULL and 0 otherwise). Note that if the server
1261 provides an opaque PRF input, the length must be the same as the
1262 length of the client's opaque PRF input.
1264 Note that the callback function will only be called when creating
1265 a new session (session resumption can resume whatever was
1266 previously negotiated), and will not be called in SSL 2.0
1267 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
1268 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
1269 for applications that need to enforce opaque PRF input.
1273 *) Update ssl code to support digests other than SHA1+MD5 for handshake
1276 [Victor B. Wagner <vitus@cryptocom.ru>]
1278 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1279 RFC4507bis. The encrypted ticket format is an encrypted encoded
1280 SSL_SESSION structure, that way new session features are automatically
1283 If a client application caches session in an SSL_SESSION structure
1284 support is transparent because tickets are now stored in the encoded
1287 The SSL_CTX structure automatically generates keys for ticket
1288 protection in servers so again support should be possible
1289 with no application modification.
1291 If a client or server wishes to disable RFC4507 support then the option
1292 SSL_OP_NO_TICKET can be set.
1294 Add a TLS extension debugging callback to allow the contents of any client
1295 or server extensions to be examined.
1297 This work was sponsored by Google.
1300 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1301 OpenSSL should now compile cleanly on gcc 4.2
1302 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1304 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1305 support including streaming MAC support: this is required for GOST
1306 ciphersuite support.
1307 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1309 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1310 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1311 to output in BER and PEM format.
1314 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1315 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1316 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1317 ENGINE support for HMAC keys which are unextractable. New -mac and
1318 -macopt options to dgst utility.
1321 *) New option -sigopt to dgst utility. Update dgst to use
1322 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1323 alternative signing paramaters such as X9.31 or PSS in the dgst
1327 *) Change ssl_cipher_apply_rule(), the internal function that does
1328 the work each time a ciphersuite string requests enabling
1329 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1330 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1331 the order of disabled ciphersuites such that those ciphersuites
1332 that most recently went from enabled to disabled not only stay
1333 in order with respect to each other, but also have higher priority
1334 than other disabled ciphersuites the next time ciphersuites are
1337 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1338 the same ciphersuites as with "HIGH" alone, but in a specific
1339 order where the PSK ciphersuites come first (since they are the
1340 most recently disabled ciphersuites when "HIGH" is parsed).
1342 Also, change ssl_create_cipher_list() (using this new
1343 funcionality) such that between otherwise identical
1344 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
1348 *) Change ssl_create_cipher_list() so that it automatically
1349 arranges the ciphersuites in reasonable order before starting
1350 to process the rule string. Thus, the definition for "DEFAULT"
1351 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
1352 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
1353 This makes it much easier to arrive at a reasonable default order
1354 in applications for which anonymous ciphers are OK (meaning
1355 that you can't actually use DEFAULT).
1356 [Bodo Moeller; suggested by Victor Duchovni]
1358 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
1359 processing) into multiple integers instead of setting
1360 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
1361 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
1362 (These masks as well as the individual bit definitions are hidden
1363 away into the non-exported interface ssl/ssl_locl.h, so this
1364 change to the definition of the SSL_CIPHER structure shouldn't
1365 affect applications.) This give us more bits for each of these
1366 categories, so there is no longer a need to coagulate AES128 and
1367 AES256 into a single algorithm bit, and to coagulate Camellia128
1368 and Camellia256 into a single algorithm bit, which has led to all
1371 Thus, among other things, the kludge introduced in 0.9.7m and
1372 0.9.8e for masking out AES256 independently of AES128 or masking
1373 out Camellia256 independently of AES256 is not needed here in 0.9.9.
1375 With the change, we also introduce new ciphersuite aliases that
1376 so far were missing: "AES128", "AES256", "CAMELLIA128", and
1380 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
1381 Use the leftmost N bytes of the signature input if the input is
1382 larger than the prime q (with N being the size in bytes of q).
1385 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1386 it yet and it is largely untested.
1389 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1392 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1393 some compilers (gcc 4.2 and later) reject their use. Safestack is
1394 reimplemented. Update ASN1 to avoid use of legacy functions.
1397 *) Win32/64 targets are linked with Winsock2.
1400 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1401 to external functions. This can be used to increase CRL handling
1402 efficiency especially when CRLs are very large by (for example) storing
1403 the CRL revoked certificates in a database.
1406 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1407 new CRLs added to a directory can be used. New command line option
1408 -verify_return_error to s_client and s_server. This causes real errors
1409 to be returned by the verify callback instead of carrying on no matter
1410 what. This reflects the way a "real world" verify callback would behave.
1413 *) GOST engine, supporting several GOST algorithms and public key formats.
1414 Kindly donated by Cryptocom.
1417 *) Partial support for Issuing Distribution Point CRL extension. CRLs
1418 partitioned by DP are handled but no indirect CRL or reason partitioning
1419 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
1420 selected via a scoring technique which handles IDP and AKID in CRLs.
1423 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1424 will ultimately be used for all verify operations: this will remove the
1425 X509_STORE dependency on certificate verification and allow alternative
1426 lookup methods. X509_STORE based implementations of these two callbacks.
1429 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1430 Modify get_crl() to find a valid (unexpired) CRL if possible.
1433 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1434 this would be called X509_CRL_cmp() but that name is already used by
1435 a function that just compares CRL issuer names. Cache several CRL
1436 extensions in X509_CRL structure and cache CRLDP in X509.
1439 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1440 this maps equivalent X509_NAME structures into a consistent structure.
1441 Name comparison can then be performed rapidly using memcmp().
1444 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1448 *) Allow digests to supply their own micalg string for S/MIME type using
1449 the ctrl EVP_MD_CTRL_MICALG.
1452 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1453 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1454 ctrl. It can then customise the structure before and/or after signing
1458 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1459 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1460 to free up any added signature OIDs.
1463 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1464 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1465 digest and cipher tables. New options added to openssl utility:
1466 list-message-digest-algorithms and list-cipher-algorithms.
1469 *) Change the array representation of binary polynomials: the list
1470 of degrees of non-zero coefficients is now terminated with -1.
1471 Previously it was terminated with 0, which was also part of the
1472 value; thus, the array representation was not applicable to
1473 polynomials where t^0 has coefficient zero. This change makes
1474 the array representation useful in a more general context.
1477 *) Various modifications and fixes to SSL/TLS cipher string
1478 handling. For ECC, the code now distinguishes between fixed ECDH
1479 with RSA certificates on the one hand and with ECDSA certificates
1480 on the other hand, since these are separate ciphersuites. The
1481 unused code for Fortezza ciphersuites has been removed.
1483 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1484 (not "ECDHE"). For consistency with the code for DH
1485 certificates, use of ECDH certificates is now considered ECDH
1486 authentication, not RSA or ECDSA authentication (the latter is
1487 merely the CA's signing algorithm and not actively used in the
1490 The temporary ciphersuite alias "ECCdraft" is no longer
1491 available, and ECC ciphersuites are no longer excluded from "ALL"
1492 and "DEFAULT". The following aliases now exist for RFC 4492
1493 ciphersuites, most of these by analogy with the DH case:
1495 kECDHr - ECDH cert, signed with RSA
1496 kECDHe - ECDH cert, signed with ECDSA
1497 kECDH - ECDH cert (signed with either RSA or ECDSA)
1498 kEECDH - ephemeral ECDH
1499 ECDH - ECDH cert or ephemeral ECDH
1505 AECDH - anonymous ECDH
1506 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1510 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1511 Use correct micalg parameters depending on digest(s) in signed message.
1514 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1515 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1518 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1519 an engine to register a method. Add ENGINE lookups for methods and
1520 functional reference processing.
1523 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1524 EVP_{Sign,Verify}* which allow an application to customise the signature
1528 *) New -resign option to smime utility. This adds one or more signers
1529 to an existing PKCS#7 signedData structure. Also -md option to use an
1530 alternative message digest algorithm for signing.
1533 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1534 create PKCS7 structures containing multiple signers. Update smime
1535 application to support multiple signers.
1538 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1542 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1543 Reorganize PBE internals to lookup from a static table using NIDs,
1544 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1545 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1546 PRF which will be automatically used with PBES2.
1549 *) Replace the algorithm specific calls to generate keys in "req" with the
1553 *) Update PKCS#7 enveloped data routines to use new API. This is now
1554 supported by any public key method supporting the encrypt operation. A
1555 ctrl is added to allow the public key algorithm to examine or modify
1556 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1560 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1561 a default digest type to use. In most cases this will be SHA1 but some
1562 algorithms (such as GOST) need to specify an alternative digest. The
1563 return value indicates how strong the prefernce is 1 means optional and
1564 2 is mandatory (that is it is the only supported type). Modify
1565 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1566 use the default md. Update openssl utilities to use the default digest
1567 type for signing if it is not explicitly indicated.
1570 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1571 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1572 signing method from the key type. This effectively removes the link
1573 between digests and public key types.
1576 *) Add an OID cross reference table and utility functions. Its purpose is to
1577 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1578 rsaEncryption. This will allow some of the algorithm specific hackery
1579 needed to use the correct OID to be removed.
1582 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1583 structures for PKCS7_sign(). They are now set up by the relevant public
1587 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1590 *) Add support for key derivation (agreement) in the API, DH method and
1594 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1595 public and private key formats. As a side effect these add additional
1596 command line functionality not previously available: DSA signatures can be
1597 generated and verified using pkeyutl and DH key support and generation in
1602 [Oliver Tappe <zooey@hirschkaefer.de>]
1604 *) New make target "install_html_docs" installs HTML renditions of the
1606 [Oliver Tappe <zooey@hirschkaefer.de>]
1608 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1609 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1610 support key and parameter generation and add initial key generation
1611 functionality for RSA.
1614 *) Add functions for main EVP_PKEY_method operations. The undocumented
1615 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1616 EVP_PKEY_{encrypt,decrypt}_old.
1619 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1620 key API, doesn't do much yet.
1623 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1624 public key algorithms. New option to openssl utility:
1625 "list-public-key-algorithms" to print out info.
1628 *) Implement the Supported Elliptic Curves Extension for
1629 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1632 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1633 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1636 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1637 utilities such as rsa, dsa, dsaparam etc except they process any key
1641 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1642 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1643 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1647 *) Initial support for pluggable public key ASN1.
1648 De-spaghettify the public key ASN1 handling. Move public and private
1649 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1650 algorithm specific handling to a single module within the relevant
1651 algorithm directory. Add functions to allow (near) opaque processing
1652 of public and private key structures.
1655 *) Implement the Supported Point Formats Extension for
1656 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1659 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1660 for the psk identity [hint] and the psk callback functions to the
1661 SSL_SESSION, SSL and SSL_CTX structure.
1664 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1668 SSL_CTX_use_psk_identity_hint
1669 SSL_get_psk_identity_hint
1670 SSL_get_psk_identity
1671 SSL_use_psk_identity_hint
1673 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1675 *) Add RFC 3161 compliant time stamp request creation, response generation
1676 and response verification functionality.
1677 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1679 *) Add initial support for TLS extensions, specifically for the server_name
1680 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1681 have new members for a host name. The SSL data structure has an
1682 additional member SSL_CTX *initial_ctx so that new sessions can be
1683 stored in that context to allow for session resumption, even after the
1684 SSL has been switched to a new SSL_CTX in reaction to a client's
1685 server_name extension.
1687 New functions (subject to change):
1689 SSL_get_servername()
1690 SSL_get_servername_type()
1693 New CTRL codes and macros (subject to change):
1695 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1696 - SSL_CTX_set_tlsext_servername_callback()
1697 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1698 - SSL_CTX_set_tlsext_servername_arg()
1699 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1701 openssl s_client has a new '-servername ...' option.
1703 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1704 '-key2 ...', '-servername_fatal' (subject to change). This allows
1705 testing the HostName extension for a specific single host name ('-cert'
1706 and '-key' remain fallbacks for handshakes without HostName
1707 negotiation). If the unrecogninzed_name alert has to be sent, this by
1708 default is a warning; it becomes fatal with the '-servername_fatal'
1711 [Peter Sylvester, Remy Allais, Christophe Renou]
1713 *) Whirlpool hash implementation is added.
1716 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1717 bn(64,32). Because of instruction set limitations it doesn't have
1718 any negative impact on performance. This was done mostly in order
1719 to make it possible to share assembler modules, such as bn_mul_mont
1720 implementations, between 32- and 64-bit builds without hassle.
1723 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1724 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1728 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1729 dedicated Montgomery multiplication procedure, is introduced.
1730 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1731 "64-bit" performance on certain 32-bit targets.
1734 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1735 in SSL structures. New SSL ctrl to set maximum send fragment size.
1736 Save memory by seeting the I/O buffer sizes dynamically instead of
1737 using the maximum available value.
1740 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1741 in addition to the text details.
1744 *) Very, very preliminary EXPERIMENTAL support for printing of general
1745 ASN1 structures. This currently produces rather ugly output and doesn't
1746 handle several customised structures at all.
1749 *) Integrated support for PVK file format and some related formats such
1750 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1751 these in the 'rsa' and 'dsa' utilities.
1754 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1757 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1758 place for the (very old) "NETSCAPE" format certificates which are now
1759 handled using new ASN1 code equivalents.
1762 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1763 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1764 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1767 *) Modify CRL distribution points extension code to print out previously
1768 unsupported fields. Enhance extension setting code to allow setting of
1772 *) Add print and set support for Issuing Distribution Point CRL extension.
1775 *) Change 'Configure' script to enable Camellia by default.
1778 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
1780 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1781 Thanks to Antonio Martin, Enterprise Secure Access Research and
1782 Development, Cisco Systems, Inc. for discovering this bug and
1783 preparing a fix. (CVE-2012-0050)
1786 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
1788 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1789 of the Vaudenay padding oracle attack on CBC mode encryption
1790 which enables an efficient plaintext recovery attack against
1791 the OpenSSL implementation of DTLS. Their attack exploits timing
1792 differences arising during decryption processing. A research
1793 paper describing this attack can be found at:
1794 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1795 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1796 Security Group at Royal Holloway, University of London
1797 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1798 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1799 for preparing the fix. (CVE-2011-4108)
1800 [Robin Seggelmann, Michael Tuexen]
1802 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
1803 [Ben Laurie, Kasper <ekasper@google.com>]
1805 *) Clear bytes used for block padding of SSL 3.0 records.
1807 [Adam Langley (Google)]
1809 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1810 Kadianakis <desnacked@gmail.com> for discovering this issue and
1811 Adam Langley for preparing the fix. (CVE-2011-4619)
1812 [Adam Langley (Google)]
1814 *) Prevent malformed RFC3779 data triggering an assertion failure.
1815 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1816 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1817 [Rob Austein <sra@hactrn.net>]
1819 *) Fix ssl_ciph.c set-up race.
1820 [Adam Langley (Google)]
1822 *) Fix spurious failures in ecdsatest.c.
1823 [Emilia Käsper (Google)]
1825 *) Fix the BIO_f_buffer() implementation (which was mixing different
1826 interpretations of the '..._len' fields).
1827 [Adam Langley (Google)]
1829 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1830 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1831 threads won't reuse the same blinding coefficients.
1833 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1834 lock to call BN_BLINDING_invert_ex, and avoids one use of
1835 BN_BLINDING_update for each BN_BLINDING structure (previously,
1836 the last update always remained unused).
1837 [Emilia Käsper (Google)]
1839 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1840 for multi-threaded use of ECDH.
1841 [Adam Langley (Google)]
1843 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1846 *) Add protection against ECDSA timing attacks as mentioned in the paper
1847 by Billy Bob Brumley and Nicola Tuveri, see:
1849 http://eprint.iacr.org/2011/232.pdf
1851 [Billy Bob Brumley and Nicola Tuveri]
1853 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
1855 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1856 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1858 *) Fix bug in string printing code: if *any* escaping is enabled we must
1859 escape the escape character (backslash) or the resulting string is
1863 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
1865 *) Disable code workaround for ancient and obsolete Netscape browsers
1866 and servers: an attacker can use it in a ciphersuite downgrade attack.
1867 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1870 *) Fixed J-PAKE implementation error, originally discovered by
1871 Sebastien Martini, further info and confirmation from Stefan
1872 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1875 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
1877 *) Fix extension code to avoid race conditions which can result in a buffer
1878 overrun vulnerability: resumed sessions must not be modified as they can
1879 be shared by multiple threads. CVE-2010-3864
1882 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
1885 *) Don't reencode certificate when calculating signature: cache and use
1886 the original encoding instead. This makes signature verification of
1887 some broken encodings work correctly.
1890 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
1891 is also one of the inputs.
1892 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1894 *) Don't repeatedly append PBE algorithms to table if they already exist.
1895 Sort table on each new add. This effectively makes the table read only
1896 after all algorithms are added and subsequent calls to PKCS12_pbe_add
1900 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
1902 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
1905 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
1906 access or freeing data twice (CVE-2010-0742)
1907 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
1909 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
1910 common in certificates and some applications which only call
1911 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
1915 Reduce copying into .apps and .test in makevms.com
1916 Don't try to use blank CA certificate in CA.com
1917 Allow use of C files from original directories in maketests.com
1918 [Steven M. Schweda" <sms@antinode.info>]
1920 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1922 *) When rejecting SSL/TLS records due to an incorrect version number, never
1923 update s->server with a new major version number. As of
1924 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1925 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1926 the previous behavior could result in a read attempt at NULL when
1927 receiving specific incorrect SSL/TLS records once record payload
1928 protection is active. (CVE-2010-0740)
1929 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1931 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1932 could be crashed if the relevant tables were not present (e.g. chrooted).
1933 [Tomas Hoger <thoger@redhat.com>]
1935 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1937 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1938 [Martin Olsson, Neel Mehta]
1940 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1941 accommodate for stack sorting, always a write lock!).
1944 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1945 excessive delays in the RAND_poll(): over a minute. As a workaround
1946 include a time check in the inner Heap32Next loop too.
1949 *) The code that handled flushing of data in SSL/TLS originally used the
1950 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1951 the problem outlined in PR#1949. The fix suggested there however can
1952 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1953 of Apache). So instead simplify the code to flush unconditionally.
1954 This should be fine since flushing with no data to flush is a no op.
1957 *) Handle TLS versions 2.0 and later properly and correctly use the
1958 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1959 off ancient servers have a habit of sticking around for a while...
1962 *) Modify compression code so it frees up structures without using the
1963 ex_data callbacks. This works around a problem where some applications
1964 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1965 restarting) then use compression (e.g. SSL with compression) later.
1966 This results in significant per-connection memory leaks and
1967 has caused some security issues including CVE-2008-1678 and
1971 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1972 change when encrypting or decrypting.
1975 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1976 connect and renegotiate with servers which do not support RI.
1977 Until RI is more widely deployed this option is enabled by default.
1980 *) Add "missing" ssl ctrls to clear options and mode.
1983 *) If client attempts to renegotiate and doesn't support RI respond with
1984 a no_renegotiation alert as required by RFC5746. Some renegotiating
1985 TLS clients will continue a connection gracefully when they receive
1986 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1987 waiting for a server hello which it will never receive. Now we treat a
1988 received no_renegotiation alert as a fatal error. This is because
1989 applications requesting a renegotiation might well expect it to succeed
1990 and would have no code in place to handle the server denying it so the
1991 only safe thing to do is to terminate the connection.
1994 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1995 peer supports secure renegotiation and 0 otherwise. Print out peer
1996 renegotiation support in s_client/s_server.
1999 *) Replace the highly broken and deprecated SPKAC certification method with
2000 the updated NID creation version. This should correctly handle UTF8.
2003 *) Implement RFC5746. Re-enable renegotiation but require the extension
2004 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2005 turns out to be a bad idea. It has been replaced by
2006 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2007 SSL_CTX_set_options(). This is really not recommended unless you
2008 know what you are doing.
2009 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2011 *) Fixes to stateless session resumption handling. Use initial_ctx when
2012 issuing and attempting to decrypt tickets in case it has changed during
2013 servername handling. Use a non-zero length session ID when attempting
2014 stateless session resumption: this makes it possible to determine if
2015 a resumption has occurred immediately after receiving server hello
2016 (several places in OpenSSL subtly assume this) instead of later in
2020 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2021 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
2022 fixes for a few places where the return code is not checked
2024 [Julia Lawall <julia@diku.dk>]
2026 *) Add --strict-warnings option to Configure script to include devteam
2027 warnings in other configurations.
2030 *) Add support for --libdir option and LIBDIR variable in makefiles. This
2031 makes it possible to install openssl libraries in locations which
2032 have names other than "lib", for example "/usr/lib64" which some
2034 [Steve Henson, based on patch from Jeremy Utley]
2036 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
2037 X690 8.9.12 and can produce some misleading textual output of OIDs.
2038 [Steve Henson, reported by Dan Kaminsky]
2040 *) Delete MD2 from algorithm tables. This follows the recommendation in
2041 several standards that it is not used in new applications due to
2042 several cryptographic weaknesses. For binary compatibility reasons
2043 the MD2 API is still compiled in by default.
2046 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
2050 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
2051 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
2053 [Guenter <lists@gknw.net>]
2055 *) Fix the server certificate chain building code to use X509_verify_cert(),
2056 it used to have an ad-hoc builder which was unable to cope with anything
2057 other than a simple chain.
2058 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
2060 *) Don't check self signed certificate signatures in X509_verify_cert()
2061 by default (a flag can override this): it just wastes time without
2062 adding any security. As a useful side effect self signed root CAs
2063 with non-FIPS digests are now usable in FIPS mode.
2066 *) In dtls1_process_out_of_seq_message() the check if the current message
2067 is already buffered was missing. For every new message was memory
2068 allocated, allowing an attacker to perform an denial of service attack
2069 with sending out of seq handshake messages until there is no memory
2070 left. Additionally every future messege was buffered, even if the
2071 sequence number made no sense and would be part of another handshake.
2072 So only messages with sequence numbers less than 10 in advance will be
2073 buffered. (CVE-2009-1378)
2074 [Robin Seggelmann, discovered by Daniel Mentz]
2076 *) Records are buffered if they arrive with a future epoch to be
2077 processed after finishing the corresponding handshake. There is
2078 currently no limitation to this buffer allowing an attacker to perform
2079 a DOS attack with sending records with future epochs until there is no
2080 memory left. This patch adds the pqueue_size() function to detemine
2081 the size of a buffer and limits the record buffer to 100 entries.
2083 [Robin Seggelmann, discovered by Daniel Mentz]
2085 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2086 parent structure is freed. (CVE-2009-1379)
2089 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2090 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2093 [Ilya O. <vrghost@gmail.com>]
2095 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2097 *) Disable renegotiation completely - this fixes a severe security
2098 problem (CVE-2009-3555) at the cost of breaking all
2099 renegotiation. Renegotiation can be re-enabled by setting
2100 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2101 run-time. This is really not recommended unless you know what
2105 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2107 *) Don't set val to NULL when freeing up structures, it is freed up by
2108 underlying code. If sizeof(void *) > sizeof(long) this can result in
2109 zeroing past the valid field. (CVE-2009-0789)
2110 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2112 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2113 checked correctly. This would allow some invalid signed attributes to
2114 appear to verify correctly. (CVE-2009-0591)
2115 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2117 *) Reject UniversalString and BMPString types with invalid lengths. This
2118 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2119 a legal length. (CVE-2009-0590)
2122 *) Set S/MIME signing as the default purpose rather than setting it
2123 unconditionally. This allows applications to override it at the store
2127 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2128 to handle some structures.
2131 *) Improve efficiency of mem_gets: don't search whole buffer each time
2133 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2135 *) New -hex option for openssl rand.
2138 *) Print out UTF8String and NumericString when parsing ASN1.
2141 *) Support NumericString type for name components.
2144 *) Allow CC in the environment to override the automatically chosen
2145 compiler. Note that nothing is done to ensure flags work with the
2149 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2151 *) Properly check EVP_VerifyFinal() and similar return values
2153 [Ben Laurie, Bodo Moeller, Google Security Team]
2155 *) Enable TLS extensions by default.
2158 *) Allow the CHIL engine to be loaded, whether the application is
2159 multithreaded or not. (This does not release the developer from the
2160 obligation to set up the dynamic locking callbacks.)
2161 [Sander Temme <sander@temme.net>]
2163 *) Use correct exit code if there is an error in dgst command.
2164 [Steve Henson; problem pointed out by Roland Dirlewanger]
2166 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2167 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2170 *) Add experimental JPAKE support, including demo authentication in
2171 s_client and s_server.
2174 *) Set the comparison function in v3_addr_canonize().
2175 [Rob Austein <sra@hactrn.net>]
2177 *) Add support for XMPP STARTTLS in s_client.
2178 [Philip Paeps <philip@freebsd.org>]
2180 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2181 to ensure that even with this option, only ciphersuites in the
2182 server's preference list will be accepted. (Note that the option
2183 applies only when resuming a session, so the earlier behavior was
2184 just about the algorithm choice for symmetric cryptography.)
2187 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2189 *) Fix NULL pointer dereference if a DTLS server received
2190 ChangeCipherSpec as first record (CVE-2009-1386).
2193 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
2194 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2197 *) The fix in 0.9.8c that supposedly got rid of unsafe
2198 double-checked locking was incomplete for RSA blinding,
2199 addressing just one layer of what turns out to have been
2200 doubly unsafe triple-checked locking.
2202 So now fix this for real by retiring the MONT_HELPER macro
2203 in crypto/rsa/rsa_eay.c.
2205 [Bodo Moeller; problem pointed out by Marius Schilder]
2207 *) Various precautionary measures:
2209 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2211 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2212 (NB: This would require knowledge of the secret session ticket key
2213 to exploit, in which case you'd be SOL either way.)
2215 - Change bn_nist.c so that it will properly handle input BIGNUMs
2216 outside the expected range.
2218 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2221 [Neel Mehta, Bodo Moeller]
2223 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2224 the load fails. Useful for distros.
2225 [Ben Laurie and the FreeBSD team]
2227 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2230 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2233 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2235 This work was sponsored by Logica.
2238 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2239 keystores. Support for SSL/TLS client authentication too.
2240 Not compiled unless enable-capieng specified to Configure.
2242 This work was sponsored by Logica.
2245 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2246 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2247 attribute creation routines such as certifcate requests and PKCS#12
2251 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2253 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2254 handshake which could lead to a cilent crash as found using the
2255 Codenomicon TLS test suite (CVE-2008-1672)
2256 [Steve Henson, Mark Cox]
2258 *) Fix double free in TLS server name extensions which could lead to
2259 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
2262 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
2264 Clear the error queue to ensure that error entries left from
2265 older function calls do not interfere with the correct operation.
2266 [Lutz Jaenicke, Erik de Castro Lopo]
2268 *) Remove root CA certificates of commercial CAs:
2270 The OpenSSL project does not recommend any specific CA and does not
2271 have any policy with respect to including or excluding any CA.
2272 Therefore it does not make any sense to ship an arbitrary selection
2273 of root CA certificates with the OpenSSL software.
2276 *) RSA OAEP patches to fix two separate invalid memory reads.
2277 The first one involves inputs when 'lzero' is greater than
2278 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
2279 before the beginning of from). The second one involves inputs where
2280 the 'db' section contains nothing but zeroes (there is a one-byte
2281 invalid read after the end of 'db').
2282 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2284 *) Partial backport from 0.9.9-dev:
2286 Introduce bn_mul_mont (dedicated Montgomery multiplication
2287 procedure) as a candidate for BIGNUM assembler implementation.
2288 While 0.9.9-dev uses assembler for various architectures, only
2289 x86_64 is available by default here in the 0.9.8 branch, and
2290 32-bit x86 is available through a compile-time setting.
2292 To try the 32-bit x86 assembler implementation, use Configure
2293 option "enable-montasm" (which exists only for this backport).
2295 As "enable-montasm" for 32-bit x86 disclaims code stability
2296 anyway, in this constellation we activate additional code
2297 backported from 0.9.9-dev for further performance improvements,
2298 namely BN_from_montgomery_word. (To enable this otherwise,
2299 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
2301 [Andy Polyakov (backport partially by Bodo Moeller)]
2303 *) Add TLS session ticket callback. This allows an application to set
2304 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
2305 values. This is useful for key rollover for example where several key
2306 sets may exist with different names.
2309 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
2310 This was broken until now in 0.9.8 releases, such that the only way
2311 a registered ENGINE could be used (assuming it initialises
2312 successfully on the host) was to explicitly set it as the default
2313 for the relevant algorithms. This is in contradiction with 0.9.7
2314 behaviour and the documentation. With this fix, when an ENGINE is
2315 registered into a given algorithm's table of implementations, the
2316 'uptodate' flag is reset so that auto-discovery will be used next
2317 time a new context for that algorithm attempts to select an
2319 [Ian Lister (tweaked by Geoff Thorpe)]
2321 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
2322 implemention in the following ways:
2324 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
2327 Lack of BER streaming support means one pass streaming processing is
2328 only supported if data is detached: setting the streaming flag is
2329 ignored for embedded content.
2331 CMS support is disabled by default and must be explicitly enabled
2332 with the enable-cms configuration option.
2335 *) Update the GMP engine glue to do direct copies between BIGNUM and
2336 mpz_t when openssl and GMP use the same limb size. Otherwise the
2337 existing "conversion via a text string export" trick is still used.
2338 [Paul Sheer <paulsheer@gmail.com>]
2340 *) Zlib compression BIO. This is a filter BIO which compressed and
2341 uncompresses any data passed through it.
2344 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
2345 RFC3394 compatible AES key wrapping.
2348 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
2349 sets string data without copying. X509_ALGOR_set0() and
2350 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
2351 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
2352 from an X509_ATTRIBUTE structure optionally checking it occurs only
2353 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
2357 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
2358 to get the expected BN_FLG_CONSTTIME behavior.
2359 [Bodo Moeller (Google)]
2363 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
2364 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
2365 - added some more tests to do_tests.pl
2366 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
2367 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
2368 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
2369 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
2370 - various changes to netware.pl to enable gcc-cross builds on Win32
2372 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
2373 - various changes to fix missing prototype warnings
2374 - fixed x86nasm.pl to create correct asm files for NASM COFF output
2375 - added AES, WHIRLPOOL and CPUID assembler code to build files
2376 - added missing AES assembler make rules to mk1mf.pl
2377 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
2378 [Guenter Knauf <eflash@gmx.net>]
2380 *) Implement certificate status request TLS extension defined in RFC3546.
2381 A client can set the appropriate parameters and receive the encoded
2382 OCSP response via a callback. A server can query the supplied parameters
2383 and set the encoded OCSP response in the callback. Add simplified examples
2384 to s_client and s_server.
2387 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
2389 *) Fix various bugs:
2390 + Binary incompatibility of ssl_ctx_st structure
2391 + DTLS interoperation with non-compliant servers
2392 + Don't call get_session_cb() without proposed session
2393 + Fix ia64 assembler code
2394 [Andy Polyakov, Steve Henson]
2396 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
2398 *) DTLS Handshake overhaul. There were longstanding issues with
2399 OpenSSL DTLS implementation, which were making it impossible for
2400 RFC 4347 compliant client to communicate with OpenSSL server.
2401 Unfortunately just fixing these incompatibilities would "cut off"
2402 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
2403 server keeps tolerating non RFC compliant syntax. The opposite is
2404 not true, 0.9.8f client can not communicate with earlier server.
2405 This update even addresses CVE-2007-4995.
2408 *) Changes to avoid need for function casts in OpenSSL: some compilers
2409 (gcc 4.2 and later) reject their use.
2410 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
2413 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2414 RFC4507bis. The encrypted ticket format is an encrypted encoded
2415 SSL_SESSION structure, that way new session features are automatically
2418 If a client application caches session in an SSL_SESSION structure
2419 support is transparent because tickets are now stored in the encoded
2422 The SSL_CTX structure automatically generates keys for ticket
2423 protection in servers so again support should be possible
2424 with no application modification.
2426 If a client or server wishes to disable RFC4507 support then the option
2427 SSL_OP_NO_TICKET can be set.
2429 Add a TLS extension debugging callback to allow the contents of any client
2430 or server extensions to be examined.
2432 This work was sponsored by Google.
2435 *) Add initial support for TLS extensions, specifically for the server_name
2436 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2437 have new members for a host name. The SSL data structure has an
2438 additional member SSL_CTX *initial_ctx so that new sessions can be
2439 stored in that context to allow for session resumption, even after the
2440 SSL has been switched to a new SSL_CTX in reaction to a client's
2441 server_name extension.
2443 New functions (subject to change):
2445 SSL_get_servername()
2446 SSL_get_servername_type()
2449 New CTRL codes and macros (subject to change):
2451 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2452 - SSL_CTX_set_tlsext_servername_callback()
2453 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2454 - SSL_CTX_set_tlsext_servername_arg()
2455 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2457 openssl s_client has a new '-servername ...' option.
2459 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2460 '-key2 ...', '-servername_fatal' (subject to change). This allows
2461 testing the HostName extension for a specific single host name ('-cert'
2462 and '-key' remain fallbacks for handshakes without HostName
2463 negotiation). If the unrecogninzed_name alert has to be sent, this by
2464 default is a warning; it becomes fatal with the '-servername_fatal'
2467 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2469 *) Add AES and SSE2 assembly language support to VC++ build.
2472 *) Mitigate attack on final subtraction in Montgomery reduction.
2475 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2476 (which previously caused an internal error).
2479 *) Squeeze another 10% out of IGE mode when in != out.
2482 *) AES IGE mode speedup.
2483 [Dean Gaudet (Google)]
2485 *) Add the Korean symmetric 128-bit cipher SEED (see
2486 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2487 add SEED ciphersuites from RFC 4162:
2489 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2490 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2491 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2492 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2494 To minimize changes between patchlevels in the OpenSSL 0.9.8
2495 series, SEED remains excluded from compilation unless OpenSSL
2496 is configured with 'enable-seed'.
2497 [KISA, Bodo Moeller]
2499 *) Mitigate branch prediction attacks, which can be practical if a
2500 single processor is shared, allowing a spy process to extract
2501 information. For detailed background information, see
2502 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2503 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2504 and Necessary Software Countermeasures"). The core of the change
2505 are new versions BN_div_no_branch() and
2506 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2507 respectively, which are slower, but avoid the security-relevant
2508 conditional branches. These are automatically called by BN_div()
2509 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2510 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2511 remove a conditional branch.
2513 BN_FLG_CONSTTIME is the new name for the previous
2514 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2515 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2516 in the exponent causes BN_mod_exp_mont() to use the alternative
2517 implementation in BN_mod_exp_mont_consttime().) The old name
2518 remains as a deprecated alias.
2520 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2521 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2522 constant-time implementations for more than just exponentiation.
2523 Here too the old name is kept as a deprecated alias.
2525 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2526 the BN_BLINDING structure gets an independent copy of the
2527 modulus. This means that the previous "BIGNUM *m" argument to
2528 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2529 essentially becomes "const BIGNUM *m", although we can't actually
2530 change this in the header file before 0.9.9. It allows
2531 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2532 enable BN_FLG_CONSTTIME.
2534 [Matthew D Wood (Intel Corp)]
2536 *) In the SSL/TLS server implementation, be strict about session ID
2537 context matching (which matters if an application uses a single
2538 external cache for different purposes). Previously,
2539 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2540 set. This did ensure strict client verification, but meant that,
2541 with applications using a single external cache for quite
2542 different requirements, clients could circumvent ciphersuite
2543 restrictions for a given session ID context by starting a session
2544 in a different context.
2547 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2548 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2549 authentication-only ciphersuites.
2552 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2553 not complete and could lead to a possible single byte overflow
2554 (CVE-2007-5135) [Ben Laurie]
2556 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2558 *) Since AES128 and AES256 (and similarly Camellia128 and
2559 Camellia256) share a single mask bit in the logic of
2560 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2561 kludge to work properly if AES128 is available and AES256 isn't
2562 (or if Camellia128 is available and Camellia256 isn't).
2565 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2566 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2567 When a point or a seed is encoded in a BIT STRING, we need to
2568 prevent the removal of trailing zero bits to get the proper DER
2569 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2570 of a NamedBitList, for which trailing 0 bits need to be removed.)
2573 *) Have SSL/TLS server implementation tolerate "mismatched" record
2574 protocol version while receiving ClientHello even if the
2575 ClientHello is fragmented. (The server can't insist on the
2576 particular protocol version it has chosen before the ServerHello
2577 message has informed the client about his choice.)
2580 *) Add RFC 3779 support.
2581 [Rob Austein for ARIN, Ben Laurie]
2583 *) Load error codes if they are not already present instead of using a
2584 static variable. This allows them to be cleanly unloaded and reloaded.
2585 Improve header file function name parsing.
2588 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2589 or CAPABILITY handshake as required by RFCs.
2592 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2594 *) Introduce limits to prevent malicious keys being able to
2595 cause a denial of service. (CVE-2006-2940)
2596 [Steve Henson, Bodo Moeller]
2598 *) Fix ASN.1 parsing of certain invalid structures that can result
2599 in a denial of service. (CVE-2006-2937) [Steve Henson]
2601 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2602 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2604 *) Fix SSL client code which could crash if connecting to a
2605 malicious SSLv2 server. (CVE-2006-4343)
2606 [Tavis Ormandy and Will Drewry, Google Security Team]
2608 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2609 match only those. Before that, "AES256-SHA" would be interpreted
2610 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2611 the same strength classification in 0.9.7h) as we currently only
2612 have a single AES bit in the ciphersuite description bitmap.
2613 That change, however, also applied to ciphersuite strings such as
2614 "RC4-MD5" that intentionally matched multiple ciphersuites --
2615 namely, SSL 2.0 ciphersuites in addition to the more common ones
2616 from SSL 3.0/TLS 1.0.
2618 So we change the selection algorithm again: Naming an explicit
2619 ciphersuite selects this one ciphersuite, and any other similar
2620 ciphersuite (same bitmap) from *other* protocol versions.
2621 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2622 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2624 Since SSL 2.0 does not have any ciphersuites for which the
2625 128/256 bit distinction would be relevant, this works for now.
2626 The proper fix will be to use different bits for AES128 and
2627 AES256, which would have avoided the problems from the beginning;
2628 however, bits are scarce, so we can only do this in a new release
2629 (not just a patchlevel) when we can change the SSL_CIPHER
2630 definition to split the single 'unsigned long mask' bitmap into
2631 multiple values to extend the available space.
2635 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2637 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2638 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2640 *) Add AES IGE and biIGE modes.
2643 *) Change the Unix randomness entropy gathering to use poll() when
2644 possible instead of select(), since the latter has some
2645 undesirable limitations.
2646 [Darryl Miles via Richard Levitte and Bodo Moeller]
2648 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2649 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2650 cannot be implicitly activated as part of, e.g., the "AES" alias.
2651 However, please upgrade to OpenSSL 0.9.9[-dev] for
2652 non-experimental use of the ECC ciphersuites to get TLS extension
2653 support, which is required for curve and point format negotiation
2654 to avoid potential handshake problems.
2657 *) Disable rogue ciphersuites:
2659 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2660 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2661 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2663 The latter two were purportedly from
2664 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2667 Also deactivate the remaining ciphersuites from
2668 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2669 unofficial, and the ID has long expired.
2672 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2673 dual-core machines) and other potential thread-safety issues.
2676 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2677 versions), which is now available for royalty-free use
2678 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2679 Also, add Camellia TLS ciphersuites from RFC 4132.
2681 To minimize changes between patchlevels in the OpenSSL 0.9.8
2682 series, Camellia remains excluded from compilation unless OpenSSL
2683 is configured with 'enable-camellia'.
2686 *) Disable the padding bug check when compression is in use. The padding
2687 bug check assumes the first packet is of even length, this is not
2688 necessarily true if compresssion is enabled and can result in false
2689 positives causing handshake failure. The actual bug test is ancient
2690 code so it is hoped that implementations will either have fixed it by
2691 now or any which still have the bug do not support compression.
2694 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2696 *) When applying a cipher rule check to see if string match is an explicit
2697 cipher suite and only match that one cipher suite if it is.
2700 *) Link in manifests for VC++ if needed.
2701 [Austin Ziegler <halostatue@gmail.com>]
2703 *) Update support for ECC-based TLS ciphersuites according to
2704 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2705 TLS extensions, which are supported starting with the 0.9.9
2706 branch, not in the OpenSSL 0.9.8 branch).
2709 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2710 opaque EVP_CIPHER_CTX handling.
2713 *) Fixes and enhancements to zlib compression code. We now only use
2714 "zlib1.dll" and use the default __cdecl calling convention on Win32
2715 to conform with the standards mentioned here:
2716 http://www.zlib.net/DLL_FAQ.txt
2717 Static zlib linking now works on Windows and the new --with-zlib-include
2718 --with-zlib-lib options to Configure can be used to supply the location
2719 of the headers and library. Gracefully handle case where zlib library
2723 *) Several fixes and enhancements to the OID generation code. The old code
2724 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2725 handle numbers larger than ULONG_MAX, truncated printing and had a
2726 non standard OBJ_obj2txt() behaviour.
2729 *) Add support for building of engines under engine/ as shared libraries
2730 under VC++ build system.
2733 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2734 Hopefully, we will not see any false combination of paths any more.
2737 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2739 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2740 (part of SSL_OP_ALL). This option used to disable the
2741 countermeasure against man-in-the-middle protocol-version
2742 rollback in the SSL 2.0 server implementation, which is a bad
2743 idea. (CVE-2005-2969)
2745 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2746 for Information Security, National Institute of Advanced Industrial
2747 Science and Technology [AIST], Japan)]
2749 *) Add two function to clear and return the verify parameter flags.
2752 *) Keep cipherlists sorted in the source instead of sorting them at
2753 runtime, thus removing the need for a lock.
2756 *) Avoid some small subgroup attacks in Diffie-Hellman.
2757 [Nick Mathewson and Ben Laurie]
2759 *) Add functions for well-known primes.
2762 *) Extended Windows CE support.
2763 [Satoshi Nakamura and Andy Polyakov]
2765 *) Initialize SSL_METHOD structures at compile time instead of during
2766 runtime, thus removing the need for a lock.
2769 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2770 attempting to decrypt each encrypted key in turn. Add support to
2774 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2776 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2779 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2782 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2783 key into the same file any more.
2786 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2789 *) Add -utf8 command line and config file option to 'ca'.
2790 [Stefan <stf@udoma.org]
2792 *) Removed the macro des_crypt(), as it seems to conflict with some
2793 libraries. Use DES_crypt().
2796 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2797 involves renaming the source and generated shared-libs for
2798 both. The engines will accept the corrected or legacy ids
2799 ('ncipher' and '4758_cca' respectively) when binding. NB,
2800 this only applies when building 'shared'.
2801 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2803 *) Add attribute functions to EVP_PKEY structure. Modify
2804 PKCS12_create() to recognize a CSP name attribute and
2805 use it. Make -CSP option work again in pkcs12 utility.
2808 *) Add new functionality to the bn blinding code:
2809 - automatic re-creation of the BN_BLINDING parameters after
2810 a fixed number of uses (currently 32)
2811 - add new function for parameter creation
2812 - introduce flags to control the update behaviour of the
2813 BN_BLINDING parameters
2814 - hide BN_BLINDING structure
2815 Add a second BN_BLINDING slot to the RSA structure to improve
2816 performance when a single RSA object is shared among several
2820 *) Add support for DTLS.
2821 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2823 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2824 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2827 *) Remove buggy and incompletet DH cert support from
2828 ssl/ssl_rsa.c and ssl/s3_both.c
2831 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2832 the apps/openssl applications.
2835 *) Compile clean with "-Wall -Wmissing-prototypes
2836 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2837 DEBUG_SAFESTACK must also be set.
2840 *) Change ./Configure so that certain algorithms can be disabled by default.
2841 The new counterpiece to "no-xxx" is "enable-xxx".
2843 The patented RC5 and MDC2 algorithms will now be disabled unless
2844 "enable-rc5" and "enable-mdc2", respectively, are specified.
2846 (IDEA remains enabled despite being patented. This is because IDEA
2847 is frequently required for interoperability, and there is no license
2848 fee for non-commercial use. As before, "no-idea" can be used to
2849 avoid this algorithm.)
2853 *) Add processing of proxy certificates (see RFC 3820). This work was
2854 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2855 EGEE (Enabling Grids for E-science in Europe).
2858 *) RC4 performance overhaul on modern architectures/implementations, such
2859 as Intel P4, IA-64 and AMD64.
2862 *) New utility extract-section.pl. This can be used specify an alternative
2863 section number in a pod file instead of having to treat each file as
2864 a separate case in Makefile. This can be done by adding two lines to the
2867 =for comment openssl_section:XXX
2869 The blank line is mandatory.
2873 *) New arguments -certform, -keyform and -pass for s_client and s_server
2874 to allow alternative format key and certificate files and passphrase
2878 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2879 update associated structures and add various utility functions.
2881 Add new policy related verify parameters, include policy checking in
2882 standard verify code. Enhance 'smime' application with extra parameters
2883 to support policy checking and print out.
2886 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2887 Nehemiah processors. These extensions support AES encryption in hardware
2888 as well as RNG (though RNG support is currently disabled).
2889 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2891 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2894 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2895 [Andy Polyakov and a number of other people]
2897 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2898 implementation contributed by IBM.
2899 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2901 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2902 exponent rather than 'unsigned long'. There is a corresponding change to
2903 the new 'rsa_keygen' element of the RSA_METHOD structure.
2904 [Jelte Jansen, Geoff Thorpe]
2906 *) Functionality for creating the initial serial number file is now
2907 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2909 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2910 number file to 1, which is bound to cause problems. To avoid
2911 the problems while respecting compatibility between different 0.9.7
2912 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2913 CA.pl for serial number initialization. With the new release 0.9.8,
2914 we can fix the problem directly in the 'ca' utility.)
2917 *) Reduced header interdepencies by declaring more opaque objects in
2918 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2919 give fewer recursive includes, which could break lazy source code - so
2920 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2921 developers should define this symbol when building and using openssl to
2922 ensure they track the recommended behaviour, interfaces, [etc], but
2923 backwards-compatible behaviour prevails when this isn't defined.
2926 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2929 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2930 This will generate a random key of the appropriate length based on the
2931 cipher context. The EVP_CIPHER can provide its own random key generation
2932 routine to support keys of a specific form. This is used in the des and
2933 3des routines to generate a key of the correct parity. Update S/MIME
2934 code to use new functions and hence generate correct parity DES keys.
2935 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2936 valid (weak or incorrect parity).
2939 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2940 as looking them up. This is useful when the verified structure may contain
2941 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2942 present unless the new PKCS7_NO_CRL flag is asserted.
2945 *) Extend ASN1 oid configuration module. It now additionally accepts the
2948 shortName = some long name, 1.2.3.4
2951 *) Reimplemented the BN_CTX implementation. There is now no more static
2952 limitation on the number of variables it can handle nor the depth of the
2953 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2954 information can now expand as required, and rather than having a single
2955 static array of bignums, BN_CTX now uses a linked-list of such arrays
2956 allowing it to expand on demand whilst maintaining the usefulness of
2957 BN_CTX's "bundling".
2960 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2961 to allow all RSA operations to function using a single BN_CTX.
2964 *) Preliminary support for certificate policy evaluation and checking. This
2965 is initially intended to pass the tests outlined in "Conformance Testing
2966 of Relying Party Client Certificate Path Processing Logic" v1.07.
2969 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2970 remained unused and not that useful. A variety of other little bignum
2971 tweaks and fixes have also been made continuing on from the audit (see
2975 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2976 associated ASN1, EVP and SSL functions and old ASN1 macros.
2979 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2980 and this should never fail. So the return value from the use of
2981 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2982 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2985 *) BN_CTX_get() should return zero-valued bignums, providing the same
2986 initialised value as BN_new().
2987 [Geoff Thorpe, suggested by Ulf Möller]
2989 *) Support for inhibitAnyPolicy certificate extension.
2992 *) An audit of the BIGNUM code is underway, for which debugging code is
2993 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2994 is considered valid when processing BIGNUMs, and causes execution to
2995 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2996 further steps are taken to deliberately pollute unused data in BIGNUM
2997 structures to try and expose faulty code further on. For now, openssl will
2998 (in its default mode of operation) continue to tolerate the inconsistent
2999 forms that it has tolerated in the past, but authors and packagers should
3000 consider trying openssl and their own applications when compiled with
3001 these debugging symbols defined. It will help highlight potential bugs in
3002 their own code, and will improve the test coverage for OpenSSL itself. At
3003 some point, these tighter rules will become openssl's default to improve
3004 maintainability, though the assert()s and other overheads will remain only
3005 in debugging configurations. See bn.h for more details.
3006 [Geoff Thorpe, Nils Larsch, Ulf Möller]
3008 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
3009 that can only be obtained through BN_CTX_new() (which implicitly
3010 initialises it). The presence of this function only made it possible
3011 to overwrite an existing structure (and cause memory leaks).
3014 *) Because of the callback-based approach for implementing LHASH as a
3015 template type, lh_insert() adds opaque objects to hash-tables and
3016 lh_doall() or lh_doall_arg() are typically used with a destructor callback
3017 to clean up those corresponding objects before destroying the hash table
3018 (and losing the object pointers). So some over-zealous constifications in
3019 LHASH have been relaxed so that lh_insert() does not take (nor store) the
3020 objects as "const" and the lh_doall[_arg] callback wrappers are not
3021 prototyped to have "const" restrictions on the object pointers they are
3022 given (and so aren't required to cast them away any more).
3025 *) The tmdiff.h API was so ugly and minimal that our own timing utility
3026 (speed) prefers to use its own implementation. The two implementations
3027 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
3028 its object type properly exposed (MS_TM) instead of casting to/from "char
3029 *". This may still change yet if someone realises MS_TM and "ms_time_***"
3030 aren't necessarily the greatest nomenclatures - but this is what was used
3031 internally to the implementation so I've used that for now.
3034 *) Ensure that deprecated functions do not get compiled when
3035 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
3036 the self-tests were still using deprecated key-generation functions so
3037 these have been updated also.
3040 *) Reorganise PKCS#7 code to separate the digest location functionality
3041 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
3042 New function PKCS7_set_digest() to set the digest type for PKCS#7
3043 digestedData type. Add additional code to correctly generate the
3044 digestedData type and add support for this type in PKCS7 initialization
3048 *) New function PKCS7_set0_type_other() this initializes a PKCS7
3049 structure of type "other".
3052 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
3053 sure the loop does correctly stop and breaking ("division by zero")
3054 modulus operations are not performed. The (pre-generated) prime
3055 table crypto/bn/bn_prime.h was already correct, but it could not be
3056 re-generated on some platforms because of the "division by zero"
3057 situation in the script.
3058 [Ralf S. Engelschall]
3060 *) Update support for ECC-based TLS ciphersuites according to
3061 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
3062 SHA-1 now is only used for "small" curves (where the
3063 representation of a field element takes up to 24 bytes); for
3064 larger curves, the field element resulting from ECDH is directly
3065 used as premaster secret.
3066 [Douglas Stebila (Sun Microsystems Laboratories)]
3068 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
3069 curve secp160r1 to the tests.
3070 [Douglas Stebila (Sun Microsystems Laboratories)]
3072 *) Add the possibility to load symbols globally with DSO.
3073 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
3075 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
3076 control of the error stack.
3079 *) Add support for STORE in ENGINE.
3082 *) Add the STORE type. The intention is to provide a common interface
3083 to certificate and key stores, be they simple file-based stores, or
3084 HSM-type store, or LDAP stores, or...
3085 NOTE: The code is currently UNTESTED and isn't really used anywhere.
3088 *) Add a generic structure called OPENSSL_ITEM. This can be used to
3089 pass a list of arguments to any function as well as provide a way
3090 for a function to pass data back to the caller.
3093 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
3094 works like BUF_strdup() but can be used to duplicate a portion of
3095 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
3099 *) Add the function sk_find_ex() which works like sk_find(), but will
3100 return an index to an element even if an exact match couldn't be
3101 found. The index is guaranteed to point at the element where the
3102 searched-for key would be inserted to preserve sorting order.
3105 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
3106 takes an extra flags argument for optional functionality. Currently,
3107 the following flags are defined:
3109 OBJ_BSEARCH_VALUE_ON_NOMATCH
3110 This one gets OBJ_bsearch_ex() to return a pointer to the first
3111 element where the comparing function returns a negative or zero
3114 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
3115 This one gets OBJ_bsearch_ex() to return a pointer to the first
3116 element where the comparing function returns zero. This is useful
3117 if there are more than one element where the comparing function
3121 *) Make it possible to create self-signed certificates with 'openssl ca'
3122 in such a way that the self-signed certificate becomes part of the
3123 CA database and uses the same mechanisms for serial number generation
3124 as all other certificate signing. The new flag '-selfsign' enables
3125 this functionality. Adapt CA.sh and CA.pl.in.
3128 *) Add functionality to check the public key of a certificate request
3129 against a given private. This is useful to check that a certificate
3130 request can be signed by that key (self-signing).
3133 *) Make it possible to have multiple active certificates with the same
3134 subject in the CA index file. This is done only if the keyword
3135 'unique_subject' is set to 'no' in the main CA section (default
3136 if 'CA_default') of the configuration file. The value is saved
3137 with the database itself in a separate index attribute file,
3138 named like the index file with '.attr' appended to the name.
3141 *) Generate muti valued AVAs using '+' notation in config files for
3145 *) Support for nameConstraints certificate extension.
3148 *) Support for policyConstraints certificate extension.
3151 *) Support for policyMappings certificate extension.
3154 *) Make sure the default DSA_METHOD implementation only uses its
3155 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3156 and change its own handlers to be NULL so as to remove unnecessary
3157 indirection. This lets alternative implementations fallback to the
3158 default implementation more easily.
3161 *) Support for directoryName in GeneralName related extensions
3165 *) Make it possible to link applications using Makefile.shared.
3166 Make that possible even when linking against static libraries!
3169 *) Support for single pass processing for S/MIME signing. This now
3170 means that S/MIME signing can be done from a pipe, in addition
3171 cleartext signing (multipart/signed type) is effectively streaming
3172 and the signed data does not need to be all held in memory.
3174 This is done with a new flag PKCS7_STREAM. When this flag is set
3175 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3176 is done after the data is output (and digests calculated) in
3177 SMIME_write_PKCS7().
3180 *) Add full support for -rpath/-R, both in shared libraries and
3181 applications, at least on the platforms where it's known how
3185 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3186 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3187 will now compute a table of multiples of the generator that
3188 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3189 faster (notably in the case of a single point multiplication,
3190 scalar * generator).
3191 [Nils Larsch, Bodo Moeller]
3193 *) IPv6 support for certificate extensions. The various extensions
3194 which use the IP:a.b.c.d can now take IPv6 addresses using the
3195 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3199 *) Added an ENGINE that implements RSA by performing private key
3200 exponentiations with the GMP library. The conversions to and from
3201 GMP's mpz_t format aren't optimised nor are any montgomery forms
3202 cached, and on x86 it appears OpenSSL's own performance has caught up.
3203 However there are likely to be other architectures where GMP could
3204 provide a boost. This ENGINE is not built in by default, but it can be
3205 specified at Configure time and should be accompanied by the necessary
3206 linker additions, eg;
3207 ./config -DOPENSSL_USE_GMP -lgmp
3210 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3211 testing availability of engines with "-t" - the old behaviour is
3212 produced by increasing the feature's verbosity with "-tt".
3215 *) ECDSA routines: under certain error conditions uninitialized BN objects
3216 could be freed. Solution: make sure initialization is performed early
3217 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3221 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3222 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3223 software implementations. For DSA and DH, parameter generation can
3224 also be overriden by providing the appropriate method callbacks.
3227 *) Change the "progress" mechanism used in key-generation and
3228 primality testing to functions that take a new BN_GENCB pointer in
3229 place of callback/argument pairs. The new API functions have "_ex"
3230 postfixes and the older functions are reimplemented as wrappers for
3231 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3232 declarations of the old functions to help (graceful) attempts to
3233 migrate to the new functions. Also, the new key-generation API
3234 functions operate on a caller-supplied key-structure and return
3235 success/failure rather than returning a key or NULL - this is to
3236 help make "keygen" another member function of RSA_METHOD etc.
3238 Example for using the new callback interface:
3240 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3244 BN_GENCB_set(&my_cb, my_callback, my_arg);
3246 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3247 /* For the meaning of a, b in calls to my_callback(), see the
3248 * documentation of the function that calls the callback.
3249 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3250 * my_callback should return 1 if it wants BN_is_prime_ex()
3251 * to continue, or 0 to stop.
3256 *) Change the ZLIB compression method to be stateful, and make it
3257 available to TLS with the number defined in
3258 draft-ietf-tls-compression-04.txt.
3261 *) Add the ASN.1 structures and functions for CertificatePair, which
3262 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
3264 CertificatePair ::= SEQUENCE {
3265 forward [0] Certificate OPTIONAL,
3266 reverse [1] Certificate OPTIONAL,
3267 -- at least one of the pair shall be present -- }
3269 Also implement the PEM functions to read and write certificate
3270 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
3272 This needed to be defined, mostly for the sake of the LDAP
3273 attribute crossCertificatePair, but may prove useful elsewhere as
3277 *) Make it possible to inhibit symlinking of shared libraries in
3278 Makefile.shared, for Cygwin's sake.
3281 *) Extend the BIGNUM API by creating a function
3282 void BN_set_negative(BIGNUM *a, int neg);
3283 and a macro that behave like
3284 int BN_is_negative(const BIGNUM *a);
3286 to avoid the need to access 'a->neg' directly in applications.
3289 *) Implement fast modular reduction for pseudo-Mersenne primes
3290 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
3291 EC_GROUP_new_curve_GFp() will now automatically use this
3293 [Nils Larsch <nla@trustcenter.de>]
3295 *) Add new lock type (CRYPTO_LOCK_BN).
3298 *) Change the ENGINE framework to automatically load engines
3299 dynamically from specific directories unless they could be
3300 found to already be built in or loaded. Move all the
3301 current engines except for the cryptodev one to a new
3303 The engines in engines/ are built as shared libraries if
3304 the "shared" options was given to ./Configure or ./config.
3305 Otherwise, they are inserted in libcrypto.a.
3306 /usr/local/ssl/engines is the default directory for dynamic
3307 engines, but that can be overriden at configure time through
3308 the usual use of --prefix and/or --openssldir, and at run
3309 time with the environment variable OPENSSL_ENGINES.
3310 [Geoff Thorpe and Richard Levitte]
3312 *) Add Makefile.shared, a helper makefile to build shared
3313 libraries. Addapt Makefile.org.
3316 *) Add version info to Win32 DLLs.
3317 [Peter 'Luna' Runestig" <peter@runestig.com>]
3319 *) Add new 'medium level' PKCS#12 API. Certificates and keys
3320 can be added using this API to created arbitrary PKCS#12
3321 files while avoiding the low level API.
3323 New options to PKCS12_create(), key or cert can be NULL and
3324 will then be omitted from the output file. The encryption
3325 algorithm NIDs can be set to -1 for no encryption, the mac
3326 iteration count can be set to 0 to omit the mac.
3328 Enhance pkcs12 utility by making the -nokeys and -nocerts
3329 options work when creating a PKCS#12 file. New option -nomac
3330 to omit the mac, NONE can be set for an encryption algorithm.
3331 New code is modified to use the enhanced PKCS12_create()
3332 instead of the low level API.
3335 *) Extend ASN1 encoder to support indefinite length constructed
3336 encoding. This can output sequences tags and octet strings in
3337 this form. Modify pk7_asn1.c to support indefinite length
3338 encoding. This is experimental and needs additional code to
3339 be useful, such as an ASN1 bio and some enhanced streaming
3342 Extend template encode functionality so that tagging is passed
3343 down to the template encoder.
3346 *) Let 'openssl req' fail if an argument to '-newkey' is not
3347 recognized instead of using RSA as a default.
3350 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
3351 As these are not official, they are not included in "ALL";
3352 the "ECCdraft" ciphersuite group alias can be used to select them.
3353 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
3355 *) Add ECDH engine support.
3356 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
3358 *) Add ECDH in new directory crypto/ecdh/.
3359 [Douglas Stebila (Sun Microsystems Laboratories)]
3361 *) Let BN_rand_range() abort with an error after 100 iterations
3362 without success (which indicates a broken PRNG).
3365 *) Change BN_mod_sqrt() so that it verifies that the input value
3366 is really the square of the return value. (Previously,
3367 BN_mod_sqrt would show GIGO behaviour.)
3370 *) Add named elliptic curves over binary fields from X9.62, SECG,
3371 and WAP/WTLS; add OIDs that were still missing.
3373 [Sheueling Chang Shantz and Douglas Stebila
3374 (Sun Microsystems Laboratories)]
3376 *) Extend the EC library for elliptic curves over binary fields
3377 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
3380 EC_GF2m_simple_method
3384 EC_GROUP_new_curve_GF2m
3385 EC_GROUP_set_curve_GF2m
3386 EC_GROUP_get_curve_GF2m
3387 EC_POINT_set_affine_coordinates_GF2m
3388 EC_POINT_get_affine_coordinates_GF2m
3389 EC_POINT_set_compressed_coordinates_GF2m
3391 Point compression for binary fields is disabled by default for
3392 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
3395 As binary polynomials are represented as BIGNUMs, various members
3396 of the EC_GROUP and EC_POINT data structures can be shared
3397 between the implementations for prime fields and binary fields;
3398 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
3399 are essentially identical to their ..._GFp counterparts.
3400 (For simplicity, the '..._GFp' prefix has been dropped from
3401 various internal method names.)
3403 An internal 'field_div' method (similar to 'field_mul' and
3404 'field_sqr') has been added; this is used only for binary fields.
3406 [Sheueling Chang Shantz and Douglas Stebila
3407 (Sun Microsystems Laboratories)]
3409 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
3410 through methods ('mul', 'precompute_mult').
3412 The generic implementations (now internally called 'ec_wNAF_mul'
3413 and 'ec_wNAF_precomputed_mult') remain the default if these
3414 methods are undefined.
3416 [Sheueling Chang Shantz and Douglas Stebila
3417 (Sun Microsystems Laboratories)]
3419 *) New function EC_GROUP_get_degree, which is defined through
3420 EC_METHOD. For curves over prime fields, this returns the bit
3421 length of the modulus.
3423 [Sheueling Chang Shantz and Douglas Stebila
3424 (Sun Microsystems Laboratories)]
3426 *) New functions EC_GROUP_dup, EC_POINT_dup.
3427 (These simply call ..._new and ..._copy).
3429 [Sheueling Chang Shantz and Douglas Stebila
3430 (Sun Microsystems Laboratories)]
3432 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3433 Polynomials are represented as BIGNUMs (where the sign bit is not
3434 used) in the following functions [macros]:
3437 BN_GF2m_sub [= BN_GF2m_add]
3438 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3439 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3440 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3442 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3443 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3444 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3445 BN_GF2m_cmp [= BN_ucmp]
3447 (Note that only the 'mod' functions are actually for fields GF(2^m).
3448 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3450 For some functions, an the irreducible polynomial defining a
3451 field can be given as an 'unsigned int[]' with strictly
3452 decreasing elements giving the indices of those bits that are set;
3453 i.e., p[] represents the polynomial
3454 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3456 p[0] > p[1] > ... > p[k] = 0.
3457 This applies to the following functions:
3462 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3463 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3465 BN_GF2m_mod_sqrt_arr
3466 BN_GF2m_mod_solve_quad_arr
3470 Conversion can be performed by the following functions:
3475 bntest.c has additional tests for binary polynomial arithmetic.
3477 Two implementations for BN_GF2m_mod_div() are available.
3478 The default algorithm simply uses BN_GF2m_mod_inv() and
3479 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3480 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3481 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3483 [Sheueling Chang Shantz and Douglas Stebila
3484 (Sun Microsystems Laboratories)]
3486 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3487 functionality is disabled at compile-time.
3488 [Douglas Stebila <douglas.stebila@sun.com>]
3490 *) Change default behaviour of 'openssl asn1parse' so that more
3491 information is visible when viewing, e.g., a certificate:
3493 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3494 mode the content of non-printable OCTET STRINGs is output in a
3495 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3496 avoid the appearance of a printable string.
3497 [Nils Larsch <nla@trustcenter.de>]
3499 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3501 EC_GROUP_set_asn1_flag()
3502 EC_GROUP_get_asn1_flag()
3503 EC_GROUP_set_point_conversion_form()
3504 EC_GROUP_get_point_conversion_form()
3505 These control ASN1 encoding details:
3506 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3507 has been set to OPENSSL_EC_NAMED_CURVE.
3508 - Points are encoded in uncompressed form by default; options for
3509 asn1_for are as for point2oct, namely
3510 POINT_CONVERSION_COMPRESSED
3511 POINT_CONVERSION_UNCOMPRESSED
3512 POINT_CONVERSION_HYBRID
3514 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3517 EC_GROUP_get0_seed()
3518 EC_GROUP_get_seed_len()
3519 This is used only for ASN1 purposes (so far).
3520 [Nils Larsch <nla@trustcenter.de>]
3522 *) Add 'field_type' member to EC_METHOD, which holds the NID
3523 of the appropriate field type OID. The new function
3524 EC_METHOD_get_field_type() returns this value.
3525 [Nils Larsch <nla@trustcenter.de>]
3530 EC_POINT_point2hex()
3531 EC_POINT_hex2point()
3532 providing useful interfaces to EC_POINT_point2oct() and
3533 EC_POINT_oct2point().
3534 [Nils Larsch <nla@trustcenter.de>]
3536 *) Change internals of the EC library so that the functions
3537 EC_GROUP_set_generator()
3538 EC_GROUP_get_generator()
3539 EC_GROUP_get_order()
3540 EC_GROUP_get_cofactor()
3541 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3542 to methods, which would lead to unnecessary code duplication when
3543 adding different types of curves.
3544 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3546 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3547 arithmetic, and such that modified wNAFs are generated
3548 (which avoid length expansion in many cases).
3551 *) Add a function EC_GROUP_check_discriminant() (defined via
3552 EC_METHOD) that verifies that the curve discriminant is non-zero.
3554 Add a function EC_GROUP_check() that makes some sanity tests
3555 on a EC_GROUP, its generator and order. This includes
3556 EC_GROUP_check_discriminant().
3557 [Nils Larsch <nla@trustcenter.de>]
3559 *) Add ECDSA in new directory crypto/ecdsa/.
3561 Add applications 'openssl ecparam' and 'openssl ecdsa'
3562 (these are based on 'openssl dsaparam' and 'openssl dsa').
3564 ECDSA support is also included in various other files across the
3565 library. Most notably,
3566 - 'openssl req' now has a '-newkey ecdsa:file' option;
3567 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3568 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3569 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3570 them suitable for ECDSA where domain parameters must be
3571 extracted before the specific public key;
3572 - ECDSA engine support has been added.
3573 [Nils Larsch <nla@trustcenter.de>]
3575 *) Include some named elliptic curves, and add OIDs from X9.62,
3576 SECG, and WAP/WTLS. Each curve can be obtained from the new
3578 EC_GROUP_new_by_curve_name(),
3579 and the list of available named curves can be obtained with
3580 EC_get_builtin_curves().
3581 Also add a 'curve_name' member to EC_GROUP objects, which can be
3583 EC_GROUP_set_curve_name()
3584 EC_GROUP_get_curve_name()
3585 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3587 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3588 was actually never needed) and in BN_mul(). The removal in BN_mul()
3589 required a small change in bn_mul_part_recursive() and the addition
3590 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3591 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3592 bn_sub_words() and bn_add_words() except they take arrays with
3596 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3598 *) Cleanse PEM buffers before freeing them since they may contain
3600 [Benjamin Bennett <ben@psc.edu>]
3602 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3603 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3604 authentication-only ciphersuites.
3607 *) Since AES128 and AES256 share a single mask bit in the logic of
3608 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3609 kludge to work properly if AES128 is available and AES256 isn't.
3612 *) Expand security boundary to match 1.1.1 module.
3615 *) Remove redundant features: hash file source, editing of test vectors
3616 modify fipsld to use external fips_premain.c signature.
3619 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3620 run algorithm test programs.
3623 *) Make algorithm test programs more tolerant of whitespace.
3626 *) Have SSL/TLS server implementation tolerate "mismatched" record
3627 protocol version while receiving ClientHello even if the
3628 ClientHello is fragmented. (The server can't insist on the
3629 particular protocol version it has chosen before the ServerHello
3630 message has informed the client about his choice.)
3633 *) Load error codes if they are not already present instead of using a
3634 static variable. This allows them to be cleanly unloaded and reloaded.
3637 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3639 *) Introduce limits to prevent malicious keys being able to
3640 cause a denial of service. (CVE-2006-2940)
3641 [Steve Henson, Bodo Moeller]
3643 *) Fix ASN.1 parsing of certain invalid structures that can result
3644 in a denial of service. (CVE-2006-2937) [Steve Henson]
3646 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3647 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3649 *) Fix SSL client code which could crash if connecting to a
3650 malicious SSLv2 server. (CVE-2006-4343)
3651 [Tavis Ormandy and Will Drewry, Google Security Team]
3653 *) Change ciphersuite string processing so that an explicit
3654 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3655 will no longer include "AES128-SHA"), and any other similar
3656 ciphersuite (same bitmap) from *other* protocol versions (so that
3657 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3658 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3659 changes from 0.9.8b and 0.9.8d.
3662 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3664 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3665 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3667 *) Change the Unix randomness entropy gathering to use poll() when
3668 possible instead of select(), since the latter has some
3669 undesirable limitations.
3670 [Darryl Miles via Richard Levitte and Bodo Moeller]
3672 *) Disable rogue ciphersuites:
3674 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3675 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3676 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3678 The latter two were purportedly from
3679 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3682 Also deactive the remaining ciphersuites from
3683 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3684 unofficial, and the ID has long expired.
3687 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3688 dual-core machines) and other potential thread-safety issues.
3691 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3693 *) Adapt fipsld and the build system to link against the validated FIPS
3694 module in FIPS mode.
3697 *) Fixes for VC++ 2005 build under Windows.
3700 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3701 from a Windows bash shell such as MSYS. It is autodetected from the
3702 "config" script when run from a VC++ environment. Modify standard VC++
3703 build to use fipscanister.o from the GNU make build.
3706 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3708 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3709 The value now differs depending on if you build for FIPS or not.
3710 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3711 safely run with a non-FIPSed libcrypto, as it may crash because of
3712 the difference induced by this change.
3715 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3717 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3718 (part of SSL_OP_ALL). This option used to disable the
3719 countermeasure against man-in-the-middle protocol-version
3720 rollback in the SSL 2.0 server implementation, which is a bad
3721 idea. (CVE-2005-2969)
3723 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3724 for Information Security, National Institute of Advanced Industrial
3725 Science and Technology [AIST], Japan)]
3727 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3728 mainly for FIPS compliance and not fully integrated at this stage.
3731 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3732 the exponentiation using a fixed-length exponent. (Otherwise,
3733 the information leaked through timing could expose the secret key
3734 after many signatures; cf. Bleichenbacher's attack on DSA with
3738 *) Make a new fixed-window mod_exp implementation the default for
3739 RSA, DSA, and DH private-key operations so that the sequence of
3740 squares and multiplies and the memory access pattern are
3741 independent of the particular secret key. This will mitigate
3742 cache-timing and potential related attacks.
3744 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3745 and this is automatically used by BN_mod_exp_mont() if the new flag
3746 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3747 will use this BN flag for private exponents unless the flag
3748 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3749 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3751 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3753 *) Change the client implementation for SSLv23_method() and
3754 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3755 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3756 (Previously, the SSL 2.0 backwards compatible Client Hello
3757 message format would be used even with SSL_OP_NO_SSLv2.)
3760 *) Add support for smime-type MIME parameter in S/MIME messages which some
3764 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3765 a threadsafe manner. Modify rsa code to use new function and add calls
3766 to dsa and dh code (which had race conditions before).
3769 *) Include the fixed error library code in the C error file definitions
3770 instead of fixing them up at runtime. This keeps the error code
3771 structures constant.
3774 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3776 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3779 *) Fixes for newer kerberos headers. NB: the casts are needed because
3780 the 'length' field is signed on one version and unsigned on another
3781 with no (?) obvious way to tell the difference, without these VC++
3782 complains. Also the "definition" of FAR (blank) is no longer included
3783 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3784 some needed definitions.
3787 *) Undo Cygwin change.
3790 *) Added support for proxy certificates according to RFC 3820.
3791 Because they may be a security thread to unaware applications,
3792 they must be explicitely allowed in run-time. See
3793 docs/HOWTO/proxy_certificates.txt for further information.
3796 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3798 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3799 server and client random values. Previously
3800 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3801 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3803 This change has negligible security impact because:
3805 1. Server and client random values still have 24 bytes of pseudo random
3808 2. Server and client random values are sent in the clear in the initial
3811 3. The master secret is derived using the premaster secret (48 bytes in
3812 size for static RSA ciphersuites) as well as client server and random
3815 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3818 [Stephen Henson, reported by UK NISCC]
3820 *) Use Windows randomness collection on Cygwin.
3823 *) Fix hang in EGD/PRNGD query when communication socket is closed
3824 prematurely by EGD/PRNGD.
3825 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3827 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3830 *) Back-port of selected performance improvements from development
3831 branch, as well as improved support for PowerPC platforms.
3834 *) Add lots of checks for memory allocation failure, error codes to indicate
3835 failure and freeing up memory if a failure occurs.
3836 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3838 *) Add new -passin argument to dgst.
3841 *) Perform some character comparisons of different types in X509_NAME_cmp:
3842 this is needed for some certificates that reencode DNs into UTF8Strings
3843 (in violation of RFC3280) and can't or wont issue name rollover
3847 *) Make an explicit check during certificate validation to see that
3848 the CA setting in each certificate on the chain is correct. As a
3849 side effect always do the following basic checks on extensions,
3850 not just when there's an associated purpose to the check:
3852 - if there is an unhandled critical extension (unless the user
3853 has chosen to ignore this fault)
3854 - if the path length has been exceeded (if one is set at all)
3855 - that certain extensions fit the associated purpose (if one has
3859 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3861 *) Avoid a race condition when CRLs are checked in a multi threaded
3862 environment. This would happen due to the reordering of the revoked
3863 entries during signature checking and serial number lookup. Now the
3864 encoding is cached and the serial number sort performed under a lock.
3865 Add new STACK function sk_is_sorted().
3868 *) Add Delta CRL to the extension code.
3871 *) Various fixes to s3_pkt.c so alerts are sent properly.
3872 [David Holmes <d.holmes@f5.com>]
3874 *) Reduce the chances of duplicate issuer name and serial numbers (in
3875 violation of RFC3280) using the OpenSSL certificate creation utilities.
3876 This is done by creating a random 64 bit value for the initial serial
3877 number when a serial number file is created or when a self signed
3878 certificate is created using 'openssl req -x509'. The initial serial
3879 number file is created using 'openssl x509 -next_serial' in CA.pl
3880 rather than being initialized to 1.
3883 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3885 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3886 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3887 [Joe Orton, Steve Henson]
3889 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3891 [Joe Orton, Steve Henson]
3893 *) Make it possible to have multiple active certificates with the same
3894 subject in the CA index file. This is done only if the keyword
3895 'unique_subject' is set to 'no' in the main CA section (default
3896 if 'CA_default') of the configuration file. The value is saved
3897 with the database itself in a separate index attribute file,
3898 named like the index file with '.attr' appended to the name.
3901 *) X509 verify fixes. Disable broken certificate workarounds when
3902 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3903 keyUsage extension present. Don't accept CRLs with unhandled critical
3904 extensions: since verify currently doesn't process CRL extensions this
3905 rejects a CRL with *any* critical extensions. Add new verify error codes
3909 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3910 A clarification of RFC2560 will require the use of OCTET STRINGs and
3911 some implementations cannot handle the current raw format. Since OpenSSL
3912 copies and compares OCSP nonces as opaque blobs without any attempt at
3913 parsing them this should not create any compatibility issues.
3916 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3917 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3918 this HMAC (and other) operations are several times slower than OpenSSL
3922 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3923 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3925 *) Use the correct content when signing type "other".
3928 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3930 *) Fix various bugs revealed by running the NISCC test suite:
3932 Stop out of bounds reads in the ASN1 code when presented with
3933 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3935 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3937 If verify callback ignores invalid public key errors don't try to check
3938 certificate signature with the NULL public key.
3942 *) New -ignore_err option in ocsp application to stop the server
3943 exiting on the first error in a request.
3946 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3947 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3951 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3952 extra data after the compression methods not only for TLS 1.0
3953 but also for SSL 3.0 (as required by the specification).
3954 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3956 *) Change X509_certificate_type() to mark the key as exported/exportable
3957 when it's 512 *bits* long, not 512 bytes.
3960 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3961 blocks during encryption.
3964 *) Various fixes to base64 BIO and non blocking I/O. On write
3965 flushes were not handled properly if the BIO retried. On read
3966 data was not being buffered properly and had various logic bugs.
3967 This also affects blocking I/O when the data being decoded is a
3971 *) Various S/MIME bugfixes and compatibility changes:
3972 output correct application/pkcs7 MIME type if
3973 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3974 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3975 of files as .eml work). Correctly handle very long lines in MIME
3979 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3981 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3982 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3983 a protocol version number mismatch like a decryption error
3984 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3987 *) Turn on RSA blinding by default in the default implementation
3988 to avoid a timing attack. Applications that don't want it can call
3989 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3990 They would be ill-advised to do so in most cases.
3991 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3993 *) Change RSA blinding code so that it works when the PRNG is not
3994 seeded (in this case, the secret RSA exponent is abused as
3995 an unpredictable seed -- if it is not unpredictable, there
3996 is no point in blinding anyway). Make RSA blinding thread-safe
3997 by remembering the creator's thread ID in rsa->blinding and
3998 having all other threads use local one-time blinding factors
3999 (this requires more computation than sharing rsa->blinding, but
4000 avoids excessive locking; and if an RSA object is not shared
4001 between threads, blinding will still be very fast).
4004 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
4005 ENGINE as defaults for all supported algorithms irrespective of
4006 the 'flags' parameter. 'flags' is now honoured, so applications
4007 should make sure they are passing it correctly.
4010 *) Target "mingw" now allows native Windows code to be generated in
4011 the Cygwin environment as well as with the MinGW compiler.
4014 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
4016 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4017 via timing by performing a MAC computation even if incorrrect
4018 block cipher padding has been found. This is a countermeasure
4019 against active attacks where the attacker has to distinguish
4020 between bad padding and a MAC verification error. (CVE-2003-0078)
4022 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4023 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4024 Martin Vuagnoux (EPFL, Ilion)]
4026 *) Make the no-err option work as intended. The intention with no-err
4027 is not to have the whole error stack handling routines removed from
4028 libcrypto, it's only intended to remove all the function name and
4029 reason texts, thereby removing some of the footprint that may not
4030 be interesting if those errors aren't displayed anyway.
4032 NOTE: it's still possible for any application or module to have it's
4033 own set of error texts inserted. The routines are there, just not
4034 used by default when no-err is given.
4037 *) Add support for FreeBSD on IA64.
4038 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
4040 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
4041 Kerberos function mit_des_cbc_cksum(). Before this change,
4042 the value returned by DES_cbc_cksum() was like the one from
4043 mit_des_cbc_cksum(), except the bytes were swapped.
4044 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
4046 *) Allow an application to disable the automatic SSL chain building.
4047 Before this a rather primitive chain build was always performed in
4048 ssl3_output_cert_chain(): an application had no way to send the
4049 correct chain if the automatic operation produced an incorrect result.
4051 Now the chain builder is disabled if either:
4053 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
4055 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
4057 The reasoning behind this is that an application would not want the
4058 auto chain building to take place if extra chain certificates are
4059 present and it might also want a means of sending no additional
4060 certificates (for example the chain has two certificates and the
4064 *) Add the possibility to build without the ENGINE framework.
4065 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4067 *) Under Win32 gmtime() can return NULL: check return value in
4068 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
4071 *) DSA routines: under certain error conditions uninitialized BN objects
4072 could be freed. Solution: make sure initialization is performed early
4073 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
4074 Nils Larsch <nla@trustcenter.de> via PR#459)
4077 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
4078 checked on reconnect on the client side, therefore session resumption
4079 could still fail with a "ssl session id is different" error. This
4080 behaviour is masked when SSL_OP_ALL is used due to
4081 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
4082 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4083 followup to PR #377.
4086 *) IA-32 assembler support enhancements: unified ELF targets, support
4087 for SCO/Caldera platforms, fix for Cygwin shared build.
4090 *) Add support for FreeBSD on sparc64. As a consequence, support for
4091 FreeBSD on non-x86 processors is separate from x86 processors on
4092 the config script, much like the NetBSD support.
4093 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
4095 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
4097 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
4100 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
4101 code (06) was taken as the first octet of the session ID and the last
4102 octet was ignored consequently. As a result SSLv2 client side session
4103 caching could not have worked due to the session ID mismatch between
4105 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4109 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
4110 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
4114 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
4115 seems that in spite of existing for more than a year, many application
4116 author have done nothing to provide the necessary callbacks, which
4117 means that this particular engine will not work properly anywhere.
4118 This is a very unfortunate situation which forces us, in the name
4119 of usability, to give the hw_ncipher.c a static lock, which is part
4121 NOTE: This is for the 0.9.7 series ONLY. This hack will never
4122 appear in 0.9.8 or later. We EXPECT application authors to have
4123 dealt properly with this when 0.9.8 is released (unless we actually
4124 make such changes in the libcrypto locking code that changes will
4125 have to be made anyway).
4128 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
4129 octets have been read, EOF or an error occurs. Without this change
4130 some truncated ASN1 structures will not produce an error.
4133 *) Disable Heimdal support, since it hasn't been fully implemented.
4134 Still give the possibility to force the use of Heimdal, but with
4135 warnings and a request that patches get sent to openssl-dev.
4138 *) Add the VC-CE target, introduce the WINCE sysname, and add
4139 INSTALL.WCE and appropriate conditionals to make it build.
4140 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4142 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4143 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4144 edit numbers of the version.
4145 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4147 *) Introduce safe string copy and catenation functions
4148 (BUF_strlcpy() and BUF_strlcat()).
4149 [Ben Laurie (CHATS) and Richard Levitte]
4151 *) Avoid using fixed-size buffers for one-line DNs.
4152 [Ben Laurie (CHATS)]
4154 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4155 resizing buffers containing secrets, and use where appropriate.
4156 [Ben Laurie (CHATS)]
4158 *) Avoid using fixed size buffers for configuration file location.
4159 [Ben Laurie (CHATS)]
4161 *) Avoid filename truncation for various CA files.
4162 [Ben Laurie (CHATS)]
4164 *) Use sizeof in preference to magic numbers.
4165 [Ben Laurie (CHATS)]
4167 *) Avoid filename truncation in cert requests.
4168 [Ben Laurie (CHATS)]
4170 *) Add assertions to check for (supposedly impossible) buffer
4172 [Ben Laurie (CHATS)]
4174 *) Don't cache truncated DNS entries in the local cache (this could
4175 potentially lead to a spoofing attack).
4176 [Ben Laurie (CHATS)]
4178 *) Fix various buffers to be large enough for hex/decimal
4179 representations in a platform independent manner.
4180 [Ben Laurie (CHATS)]
4182 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4183 resizing buffers containing secrets, and use where appropriate.
4184 [Ben Laurie (CHATS)]
4186 *) Add BIO_indent() to avoid much slightly worrying code to do
4188 [Ben Laurie (CHATS)]
4190 *) Convert sprintf()/BIO_puts() to BIO_printf().
4191 [Ben Laurie (CHATS)]
4193 *) buffer_gets() could terminate with the buffer only half
4195 [Ben Laurie (CHATS)]
4197 *) Add assertions to prevent user-supplied crypto functions from
4198 overflowing internal buffers by having large block sizes, etc.
4199 [Ben Laurie (CHATS)]
4201 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4203 [Ben Laurie (CHATS)]
4205 *) Eliminate unused copy of key in RC4.
4206 [Ben Laurie (CHATS)]
4208 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4209 [Ben Laurie (CHATS)]
4211 *) Fix off-by-one error in EGD path.
4212 [Ben Laurie (CHATS)]
4214 *) If RANDFILE path is too long, ignore instead of truncating.
4215 [Ben Laurie (CHATS)]
4217 *) Eliminate unused and incorrectly sized X.509 structure
4219 [Ben Laurie (CHATS)]
4221 *) Eliminate unused and dangerous function knumber().
4222 [Ben Laurie (CHATS)]
4224 *) Eliminate unused and dangerous structure, KSSL_ERR.
4225 [Ben Laurie (CHATS)]
4227 *) Protect against overlong session ID context length in an encoded
4228 session object. Since these are local, this does not appear to be
4230 [Ben Laurie (CHATS)]
4232 *) Change from security patch (see 0.9.6e below) that did not affect
4233 the 0.9.6 release series:
4235 Remote buffer overflow in SSL3 protocol - an attacker could
4236 supply an oversized master key in Kerberos-enabled versions.
4238 [Ben Laurie (CHATS)]
4240 *) Change the SSL kerb5 codes to match RFC 2712.
4243 *) Make -nameopt work fully for req and add -reqopt switch.
4244 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4246 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4247 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4249 *) Make sure tests can be performed even if the corresponding algorithms
4250 have been removed entirely. This was also the last step to make
4251 OpenSSL compilable with DJGPP under all reasonable conditions.
4252 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
4254 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
4255 to allow version independent disabling of normally unselected ciphers,
4256 which may be activated as a side-effect of selecting a single cipher.
4258 (E.g., cipher list string "RSA" enables ciphersuites that are left
4259 out of "ALL" because they do not provide symmetric encryption.
4260 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
4261 [Lutz Jaenicke, Bodo Moeller]
4263 *) Add appropriate support for separate platform-dependent build
4264 directories. The recommended way to make a platform-dependent
4265 build directory is the following (tested on Linux), maybe with
4268 # Place yourself outside of the OpenSSL source tree. In
4269 # this example, the environment variable OPENSSL_SOURCE
4270 # is assumed to contain the absolute OpenSSL source directory.
4271 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
4272 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
4273 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
4274 mkdir -p `dirname $F`
4275 ln -s $OPENSSL_SOURCE/$F $F
4278 To be absolutely sure not to disturb the source tree, a "make clean"
4279 is a good thing. If it isn't successfull, don't worry about it,
4280 it probably means the source directory is very clean.
4283 *) Make sure any ENGINE control commands make local copies of string
4284 pointers passed to them whenever necessary. Otherwise it is possible
4285 the caller may have overwritten (or deallocated) the original string
4286 data when a later ENGINE operation tries to use the stored values.
4287 [Götz Babin-Ebell <babinebell@trustcenter.de>]
4289 *) Improve diagnostics in file reading and command-line digests.
4290 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
4292 *) Add AES modes CFB and OFB to the object database. Correct an
4293 error in AES-CFB decryption.
4296 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
4297 allows existing EVP_CIPHER_CTX structures to be reused after
4298 calling EVP_*Final(). This behaviour is used by encryption
4299 BIOs and some applications. This has the side effect that
4300 applications must explicitly clean up cipher contexts with
4301 EVP_CIPHER_CTX_cleanup() or they will leak memory.
4304 *) Check the values of dna and dnb in bn_mul_recursive before calling
4305 bn_mul_comba (a non zero value means the a or b arrays do not contain
4306 n2 elements) and fallback to bn_mul_normal if either is not zero.
4309 *) Fix escaping of non-ASCII characters when using the -subj option
4310 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
4313 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
4314 form for "surname", serialNumber has no short form.
4315 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
4316 therefore remove "mail" short name for "internet 7".
4317 The OID for unique identifiers in X509 certificates is
4318 x500UniqueIdentifier, not uniqueIdentifier.
4319 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
4322 *) Add an "init" command to the ENGINE config module and auto initialize
4323 ENGINEs. Without any "init" command the ENGINE will be initialized
4324 after all ctrl commands have been executed on it. If init=1 the
4325 ENGINE is initailized at that point (ctrls before that point are run
4326 on the uninitialized ENGINE and after on the initialized one). If
4327 init=0 then the ENGINE will not be iniatialized at all.
4330 *) Fix the 'app_verify_callback' interface so that the user-defined
4331 argument is actually passed to the callback: In the
4332 SSL_CTX_set_cert_verify_callback() prototype, the callback
4333 declaration has been changed from
4336 int (*cb)(X509_STORE_CTX *,void *);
4337 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
4338 i=s->ctx->app_verify_callback(&ctx)
4339 has been changed into
4340 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
4342 To update applications using SSL_CTX_set_cert_verify_callback(),
4343 a dummy argument can be added to their callback functions.
4344 [D. K. Smetters <smetters@parc.xerox.com>]
4346 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
4347 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
4349 *) Add and OPENSSL_LOAD_CONF define which will cause
4350 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
4351 This allows older applications to transparently support certain
4352 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
4353 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
4354 load the config file and OPENSSL_add_all_algorithms_conf() which will
4355 always load it have also been added.
4358 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
4359 Adjust NIDs and EVP layer.
4360 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4362 *) Config modules support in openssl utility.
4364 Most commands now load modules from the config file,
4365 though in a few (such as version) this isn't done
4366 because it couldn't be used for anything.
4368 In the case of ca and req the config file used is
4369 the same as the utility itself: that is the -config
4370 command line option can be used to specify an
4374 *) Move default behaviour from OPENSSL_config(). If appname is NULL
4375 use "openssl_conf" if filename is NULL use default openssl config file.
4378 *) Add an argument to OPENSSL_config() to allow the use of an alternative
4379 config section name. Add a new flag to tolerate a missing config file
4380 and move code to CONF_modules_load_file().
4383 *) Support for crypto accelerator cards from Accelerated Encryption
4384 Processing, www.aep.ie. (Use engine 'aep')
4385 The support was copied from 0.9.6c [engine] and adapted/corrected
4386 to work with the new engine framework.
4387 [AEP Inc. and Richard Levitte]
4389 *) Support for SureWare crypto accelerator cards from Baltimore
4390 Technologies. (Use engine 'sureware')
4391 The support was copied from 0.9.6c [engine] and adapted
4392 to work with the new engine framework.
4395 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
4396 make the newer ENGINE framework commands for the CHIL engine work.
4397 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
4399 *) Make it possible to produce shared libraries on ReliantUNIX.
4400 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
4402 *) Add the configuration target debug-linux-ppro.
4403 Make 'openssl rsa' use the general key loading routines
4404 implemented in apps.c, and make those routines able to
4405 handle the key format FORMAT_NETSCAPE and the variant
4407 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4409 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4410 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4412 *) Add -keyform to rsautl, and document -engine.
4413 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
4415 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
4416 BIO_R_NO_SUCH_FILE error code rather than the generic
4417 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
4420 *) Add new functions
4422 ERR_peek_last_error_line
4423 ERR_peek_last_error_line_data.
4424 These are similar to
4427 ERR_peek_error_line_data,
4428 but report on the latest error recorded rather than the first one
4429 still in the error queue.
4430 [Ben Laurie, Bodo Moeller]
4432 *) default_algorithms option in ENGINE config module. This allows things
4434 default_algorithms = ALL
4435 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4438 *) Prelminary ENGINE config module.
4441 *) New experimental application configuration code.
4444 *) Change the AES code to follow the same name structure as all other
4445 symmetric ciphers, and behave the same way. Move everything to
4446 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4447 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4449 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4450 [Ben Laurie and Theo de Raadt]
4452 *) Add option to output public keys in req command.
4453 [Massimiliano Pala madwolf@openca.org]
4455 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4456 (up to about 10% better than before for P-192 and P-224).
4459 *) New functions/macros
4461 SSL_CTX_set_msg_callback(ctx, cb)
4462 SSL_CTX_set_msg_callback_arg(ctx, arg)
4463 SSL_set_msg_callback(ssl, cb)
4464 SSL_set_msg_callback_arg(ssl, arg)
4466 to request calling a callback function
4468 void cb(int write_p, int version, int content_type,
4469 const void *buf, size_t len, SSL *ssl, void *arg)
4471 whenever a protocol message has been completely received
4472 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4473 protocol version according to which the SSL library interprets
4474 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4475 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4476 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4477 specification (change_cipher_spec(20), alert(21), handshake(22)).
4478 'buf' and 'len' point to the actual message, 'ssl' to the
4479 SSL object, and 'arg' is the application-defined value set by
4480 SSL[_CTX]_set_msg_callback_arg().
4482 'openssl s_client' and 'openssl s_server' have new '-msg' options
4483 to enable a callback that displays all protocol messages.
4486 *) Change the shared library support so shared libraries are built as
4487 soon as the corresponding static library is finished, and thereby get
4488 openssl and the test programs linked against the shared library.
4489 This still only happens when the keyword "shard" has been given to
4490 the configuration scripts.
4492 NOTE: shared library support is still an experimental thing, and
4493 backward binary compatibility is still not guaranteed.
4494 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4496 *) Add support for Subject Information Access extension.
4497 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4499 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4500 additional bytes when new memory had to be allocated, not just
4501 when reusing an existing buffer.
4504 *) New command line and configuration option 'utf8' for the req command.
4505 This allows field values to be specified as UTF8 strings.
4508 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4509 runs for the former and machine-readable output for the latter.
4512 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4513 of the e-mail address in the DN (i.e., it will go into a certificate
4514 extension only). The new configuration file option 'email_in_dn = no'
4515 has the same effect.
4516 [Massimiliano Pala madwolf@openca.org]
4518 *) Change all functions with names starting with des_ to be starting
4519 with DES_ instead. Add wrappers that are compatible with libdes,
4520 but are named _ossl_old_des_*. Finally, add macros that map the
4521 des_* symbols to the corresponding _ossl_old_des_* if libdes
4522 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4523 desired, the des_* symbols will be mapped to DES_*, with one
4526 Since we provide two compatibility mappings, the user needs to
4527 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4528 compatibility is desired. The default (i.e., when that macro
4529 isn't defined) is OpenSSL 0.9.6c compatibility.
4531 There are also macros that enable and disable the support of old
4532 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4533 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4534 are defined, the default will apply: to support the old des routines.
4536 In either case, one must include openssl/des.h to get the correct
4537 definitions. Do not try to just include openssl/des_old.h, that
4540 NOTE: This is a major break of an old API into a new one. Software
4541 authors are encouraged to switch to the DES_ style functions. Some
4542 time in the future, des_old.h and the libdes compatibility functions
4543 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4544 default), and then completely removed.
4547 *) Test for certificates which contain unsupported critical extensions.
4548 If such a certificate is found during a verify operation it is
4549 rejected by default: this behaviour can be overridden by either
4550 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4551 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4552 X509_supported_extension() has also been added which returns 1 if a
4553 particular extension is supported.
4556 *) Modify the behaviour of EVP cipher functions in similar way to digests
4557 to retain compatibility with existing code.
4560 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4561 compatibility with existing code. In particular the 'ctx' parameter does
4562 not have to be to be initialized before the call to EVP_DigestInit() and
4563 it is tidied up after a call to EVP_DigestFinal(). New function
4564 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4565 EVP_MD_CTX_copy() changed to not require the destination to be
4566 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4567 requires the destination to be valid.
4569 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4570 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4573 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4574 so that complete 'Handshake' protocol structures are kept in memory
4575 instead of overwriting 'msg_type' and 'length' with 'body' data.
4578 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4579 [Massimo Santin via Richard Levitte]
4581 *) Major restructuring to the underlying ENGINE code. This includes
4582 reduction of linker bloat, separation of pure "ENGINE" manipulation
4583 (initialisation, etc) from functionality dealing with implementations
4584 of specific crypto iterfaces. This change also introduces integrated
4585 support for symmetric ciphers and digest implementations - so ENGINEs
4586 can now accelerate these by providing EVP_CIPHER and EVP_MD
4587 implementations of their own. This is detailed in crypto/engine/README
4588 as it couldn't be adequately described here. However, there are a few
4589 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4590 were changed in the original introduction of ENGINE code have now
4591 reverted back - the hooking from this code to ENGINE is now a good
4592 deal more passive and at run-time, operations deal directly with
4593 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4594 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4595 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4596 they were not being used by the framework as there is no concept of a
4597 BIGNUM_METHOD and they could not be generalised to the new
4598 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4599 ENGINE_cpy() has been removed as it cannot be consistently defined in
4603 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4606 *) Change mkdef.pl to sort symbols that get the same entry number,
4607 and make sure the automatically generated functions ERR_load_*
4608 become part of libeay.num as well.
4611 *) New function SSL_renegotiate_pending(). This returns true once
4612 renegotiation has been requested (either SSL_renegotiate() call
4613 or HelloRequest/ClientHello receveived from the peer) and becomes
4614 false once a handshake has been completed.
4615 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4616 sends a HelloRequest, but does not ensure that a handshake takes
4617 place. SSL_renegotiate_pending() is useful for checking if the
4618 client has followed the request.)
4621 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4622 By default, clients may request session resumption even during
4623 renegotiation (if session ID contexts permit); with this option,
4624 session resumption is possible only in the first handshake.
4626 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4627 more bits available for options that should not be part of
4628 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4631 *) Add some demos for certificate and certificate request creation.
4634 *) Make maximum certificate chain size accepted from the peer application
4635 settable (SSL*_get/set_max_cert_list()), as proposed by
4636 "Douglas E. Engert" <deengert@anl.gov>.
4639 *) Add support for shared libraries for Unixware-7
4640 (Boyd Lynn Gerber <gerberb@zenez.com>).
4643 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4644 be done prior to destruction. Use this to unload error strings from
4645 ENGINEs that load their own error strings. NB: This adds two new API
4646 functions to "get" and "set" this destroy handler in an ENGINE.
4649 *) Alter all existing ENGINE implementations (except "openssl" and
4650 "openbsd") to dynamically instantiate their own error strings. This
4651 makes them more flexible to be built both as statically-linked ENGINEs
4652 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4653 Also, add stub code to each that makes building them as self-contained
4654 shared-libraries easier (see README.ENGINE).
4657 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4658 implementations into applications that are completely implemented in
4659 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4660 commands that can be used to configure what shared-library to load and
4661 to control aspects of the way it is handled. Also, made an update to
4662 the README.ENGINE file that brings its information up-to-date and
4663 provides some information and instructions on the "dynamic" ENGINE
4664 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4667 *) Make it possible to unload ranges of ERR strings with a new
4668 "ERR_unload_strings" function.
4671 *) Add a copy() function to EVP_MD.
4674 *) Make EVP_MD routines take a context pointer instead of just the
4675 md_data void pointer.
4678 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4679 that the digest can only process a single chunk of data
4680 (typically because it is provided by a piece of
4681 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4682 is only going to provide a single chunk of data, and hence the
4683 framework needn't accumulate the data for oneshot drivers.
4686 *) As with "ERR", make it possible to replace the underlying "ex_data"
4687 functions. This change also alters the storage and management of global
4688 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4689 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4690 index counters. The API functions that use this state have been changed
4691 to take a "class_index" rather than pointers to the class's local STACK
4692 and counter, and there is now an API function to dynamically create new
4693 classes. This centralisation allows us to (a) plug a lot of the
4694 thread-safety problems that existed, and (b) makes it possible to clean
4695 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4696 such data would previously have always leaked in application code and
4697 workarounds were in place to make the memory debugging turn a blind eye
4698 to it. Application code that doesn't use this new function will still
4699 leak as before, but their memory debugging output will announce it now
4700 rather than letting it slide.
4702 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4703 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4704 has a return value to indicate success or failure.
4707 *) Make it possible to replace the underlying "ERR" functions such that the
4708 global state (2 LHASH tables and 2 locks) is only used by the "default"
4709 implementation. This change also adds two functions to "get" and "set"
4710 the implementation prior to it being automatically set the first time
4711 any other ERR function takes place. Ie. an application can call "get",
4712 pass the return value to a module it has just loaded, and that module
4713 can call its own "set" function using that value. This means the
4714 module's "ERR" operations will use (and modify) the error state in the
4715 application and not in its own statically linked copy of OpenSSL code.
4718 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4719 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4720 the operation, and provides a more encapsulated way for external code
4721 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4722 to use these functions rather than manually incrementing the counts.
4724 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4727 *) Add EVP test program.
4730 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4733 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4734 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4735 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4736 These allow a CRL to be built without having to access X509_CRL fields
4737 directly. Modify 'ca' application to use new functions.
4740 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4741 bug workarounds. Rollback attack detection is a security feature.
4742 The problem will only arise on OpenSSL servers when TLSv1 is not
4743 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4744 Software authors not wanting to support TLSv1 will have special reasons
4745 for their choice and can explicitly enable this option.
4746 [Bodo Moeller, Lutz Jaenicke]
4748 *) Rationalise EVP so it can be extended: don't include a union of
4749 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4750 (similar to those existing for EVP_CIPHER_CTX).
4755 EVP_MD_CTX_init(&md); /* new function call */
4756 EVP_DigestInit(&md, EVP_sha1());
4757 EVP_DigestUpdate(&md, in, len);
4758 EVP_DigestFinal(&md, out, NULL);
4759 EVP_MD_CTX_cleanup(&md); /* new function call */
4763 *) Make DES key schedule conform to the usual scheme, as well as
4764 correcting its structure. This means that calls to DES functions
4765 now have to pass a pointer to a des_key_schedule instead of a
4766 plain des_key_schedule (which was actually always a pointer
4769 des_key_schedule ks;
4771 des_set_key_checked(..., &ks);
4772 des_ncbc_encrypt(..., &ks, ...);
4774 (Note that a later change renames 'des_...' into 'DES_...'.)
4777 *) Initial reduction of linker bloat: the use of some functions, such as
4778 PEM causes large amounts of unused functions to be linked in due to
4779 poor organisation. For example pem_all.c contains every PEM function
4780 which has a knock on effect of linking in large amounts of (unused)
4781 ASN1 code. Grouping together similar functions and splitting unrelated
4782 functions prevents this.
4785 *) Cleanup of EVP macros.
4788 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4789 correct _ecb suffix.
4792 *) Add initial OCSP responder support to ocsp application. The
4793 revocation information is handled using the text based index
4794 use by the ca application. The responder can either handle
4795 requests generated internally, supplied in files (for example
4796 via a CGI script) or using an internal minimal server.
4799 *) Add configuration choices to get zlib compression for TLS.
4802 *) Changes to Kerberos SSL for RFC 2712 compliance:
4803 1. Implemented real KerberosWrapper, instead of just using
4804 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4805 2. Implemented optional authenticator field of KerberosWrapper.
4807 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4808 and authenticator structs; see crypto/krb5/.
4810 Generalized Kerberos calls to support multiple Kerberos libraries.
4811 [Vern Staats <staatsvr@asc.hpc.mil>,
4812 Jeffrey Altman <jaltman@columbia.edu>
4813 via Richard Levitte]
4815 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4816 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4817 values for each of the key sizes rather than having just
4818 parameters (and 'speed' generating keys each time).
4821 *) Speed up EVP routines.
4824 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4825 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4826 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4827 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4829 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4830 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4831 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4834 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4836 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4839 *) Added the OS2-EMX target.
4840 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4842 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4843 to support NCONF routines in extension code. New function CONF_set_nconf()
4844 to allow functions which take an NCONF to also handle the old LHASH
4845 structure: this means that the old CONF compatible routines can be
4846 retained (in particular wrt extensions) without having to duplicate the
4847 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4850 *) Enhance the general user interface with mechanisms for inner control
4851 and with possibilities to have yes/no kind of prompts.
4854 *) Change all calls to low level digest routines in the library and
4855 applications to use EVP. Add missing calls to HMAC_cleanup() and
4856 don't assume HMAC_CTX can be copied using memcpy().
4857 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4859 *) Add the possibility to control engines through control names but with
4860 arbitrary arguments instead of just a string.
4861 Change the key loaders to take a UI_METHOD instead of a callback
4862 function pointer. NOTE: this breaks binary compatibility with earlier
4863 versions of OpenSSL [engine].
4864 Adapt the nCipher code for these new conditions and add a card insertion
4868 *) Enhance the general user interface with mechanisms to better support
4869 dialog box interfaces, application-defined prompts, the possibility
4870 to use defaults (for example default passwords from somewhere else)
4871 and interrupts/cancellations.
4874 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4875 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4878 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4879 tidy up some unnecessarily weird code in 'sk_new()').
4880 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4882 *) Change the key loading routines for ENGINEs to use the same kind
4883 callback (pem_password_cb) as all other routines that need this
4887 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4888 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4889 than this minimum value is recommended.
4892 *) New random seeder for OpenVMS, using the system process statistics
4893 that are easily reachable.
4896 *) Windows apparently can't transparently handle global
4897 variables defined in DLLs. Initialisations such as:
4899 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4901 wont compile. This is used by the any applications that need to
4902 declare their own ASN1 modules. This was fixed by adding the option
4903 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4904 needed for static libraries under Win32.
4907 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4908 setting of purpose and trust fields. New X509_STORE trust and
4909 purpose functions and tidy up setting in other SSL functions.
4912 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4913 structure. These are inherited by X509_STORE_CTX when it is
4914 initialised. This allows various defaults to be set in the
4915 X509_STORE structure (such as flags for CRL checking and custom
4916 purpose or trust settings) for functions which only use X509_STORE_CTX
4917 internally such as S/MIME.
4919 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4920 trust settings if they are not set in X509_STORE. This allows X509_STORE
4921 purposes and trust (in S/MIME for example) to override any set by default.
4923 Add command line options for CRL checking to smime, s_client and s_server
4927 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4928 are set then the CRL is looked up in the X509_STORE structure and
4929 its validity and signature checked, then if the certificate is found
4930 in the CRL the verify fails with a revoked error.
4932 Various new CRL related callbacks added to X509_STORE_CTX structure.
4934 Command line options added to 'verify' application to support this.
4936 This needs some additional work, such as being able to handle multiple
4937 CRLs with different times, extension based lookup (rather than just
4938 by subject name) and ultimately more complete V2 CRL extension
4942 *) Add a general user interface API (crypto/ui/). This is designed
4943 to replace things like des_read_password and friends (backward
4944 compatibility functions using this new API are provided).
4945 The purpose is to remove prompting functions from the DES code
4946 section as well as provide for prompting through dialog boxes in
4947 a window system and the like.
4950 *) Add "ex_data" support to ENGINE so implementations can add state at a
4951 per-structure level rather than having to store it globally.
4954 *) Make it possible for ENGINE structures to be copied when retrieved by
4955 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4956 This causes the "original" ENGINE structure to act like a template,
4957 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4958 operational state can be localised to each ENGINE structure, despite the
4959 fact they all share the same "methods". New ENGINE structures returned in
4960 this case have no functional references and the return value is the single
4961 structural reference. This matches the single structural reference returned
4962 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4966 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4967 needs to match any other type at all we need to manually clear the
4971 *) Changes to the "openssl engine" utility to include;
4972 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4973 about an ENGINE's available control commands.
4974 - executing control commands from command line arguments using the
4975 '-pre' and '-post' switches. '-post' is only used if '-t' is
4976 specified and the ENGINE is successfully initialised. The syntax for
4977 the individual commands are colon-separated, for example;
4978 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4981 *) New dynamic control command support for ENGINEs. ENGINEs can now
4982 declare their own commands (numbers), names (strings), descriptions,
4983 and input types for run-time discovery by calling applications. A
4984 subset of these commands are implicitly classed as "executable"
4985 depending on their input type, and only these can be invoked through
4986 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4987 can be based on user input, config files, etc). The distinction is
4988 that "executable" commands cannot return anything other than a boolean
4989 result and can only support numeric or string input, whereas some
4990 discoverable commands may only be for direct use through
4991 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4992 pointers, or other custom uses. The "executable" commands are to
4993 support parameterisations of ENGINE behaviour that can be
4994 unambiguously defined by ENGINEs and used consistently across any
4995 OpenSSL-based application. Commands have been added to all the
4996 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4997 control over shared-library paths without source code alterations.
5000 *) Changed all ENGINE implementations to dynamically allocate their
5001 ENGINEs rather than declaring them statically. Apart from this being
5002 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
5003 this also allows the implementations to compile without using the
5004 internal engine_int.h header.
5007 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
5008 'const' value. Any code that should be able to modify a RAND_METHOD
5009 should already have non-const pointers to it (ie. they should only
5010 modify their own ones).
5013 *) Made a variety of little tweaks to the ENGINE code.
5014 - "atalla" and "ubsec" string definitions were moved from header files
5015 to C code. "nuron" string definitions were placed in variables
5016 rather than hard-coded - allowing parameterisation of these values
5017 later on via ctrl() commands.
5018 - Removed unused "#if 0"'d code.
5019 - Fixed engine list iteration code so it uses ENGINE_free() to release
5020 structural references.
5021 - Constified the RAND_METHOD element of ENGINE structures.
5022 - Constified various get/set functions as appropriate and added
5023 missing functions (including a catch-all ENGINE_cpy that duplicates
5024 all ENGINE values onto a new ENGINE except reference counts/state).
5025 - Removed NULL parameter checks in get/set functions. Setting a method
5026 or function to NULL is a way of cancelling out a previously set
5027 value. Passing a NULL ENGINE parameter is just plain stupid anyway
5028 and doesn't justify the extra error symbols and code.
5029 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
5030 flags from engine_int.h to engine.h.
5031 - Changed prototypes for ENGINE handler functions (init(), finish(),
5032 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
5035 *) Implement binary inversion algorithm for BN_mod_inverse in addition
5036 to the algorithm using long division. The binary algorithm can be
5037 used only if the modulus is odd. On 32-bit systems, it is faster
5038 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
5039 roughly 5-15% for 256-bit moduli), so we use it only for moduli
5040 up to 450 bits. In 64-bit environments, the binary algorithm
5041 appears to be advantageous for much longer moduli; here we use it
5042 for moduli up to 2048 bits.
5045 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
5046 could not support the combine flag in choice fields.
5049 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
5050 extensions from a certificate request to the certificate.
5053 *) Allow multiple 'certopt' and 'nameopt' options to be separated
5054 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
5055 file: this allows the display of the certificate about to be
5056 signed to be customised, to allow certain fields to be included
5057 or excluded and extension details. The old system didn't display
5058 multicharacter strings properly, omitted fields not in the policy
5059 and couldn't display additional details such as extensions.
5062 *) Function EC_POINTs_mul for multiple scalar multiplication
5063 of an arbitrary number of elliptic curve points
5064 \sum scalars[i]*points[i],
5065 optionally including the generator defined for the EC_GROUP:
5066 scalar*generator + \sum scalars[i]*points[i].
5068 EC_POINT_mul is a simple wrapper function for the typical case
5069 that the point list has just one item (besides the optional
5073 *) First EC_METHODs for curves over GF(p):
5075 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
5076 operations and provides various method functions that can also
5077 operate with faster implementations of modular arithmetic.
5079 EC_GFp_mont_method() reuses most functions that are part of
5080 EC_GFp_simple_method, but uses Montgomery arithmetic.
5082 [Bodo Moeller; point addition and point doubling
5083 implementation directly derived from source code provided by
5084 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
5086 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
5087 crypto/ec/ec_lib.c):
5089 Curves are EC_GROUP objects (with an optional group generator)
5090 based on EC_METHODs that are built into the library.
5092 Points are EC_POINT objects based on EC_GROUP objects.
5094 Most of the framework would be able to handle curves over arbitrary
5095 finite fields, but as there are no obvious types for fields other
5096 than GF(p), some functions are limited to that for now.
5099 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
5100 that the file contains a complete HTTP response.
5103 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
5104 change the def and num file printf format specifier from "%-40sXXX"
5105 to "%-39s XXX". The latter will always guarantee a space after the
5106 field while the former will cause them to run together if the field
5107 is 40 of more characters long.
5110 *) Constify the cipher and digest 'method' functions and structures
5111 and modify related functions to take constant EVP_MD and EVP_CIPHER
5115 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
5116 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
5119 *) Modify EVP_Digest*() routines so they now return values. Although the
5120 internal software routines can never fail additional hardware versions
5124 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
5126 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
5127 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
5130 ERR_R_NESTED_ASN1_ERROR
5132 ERR_R_MISSING_ASN1_EOS
5133 were 4 .. 9, conflicting with
5134 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5136 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5137 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5139 Add new error code 'ERR_R_INTERNAL_ERROR'.
5142 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5146 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5147 sets the subject name for a new request or supersedes the
5148 subject name in a given request. Formats that can be parsed are
5149 'CN=Some Name, OU=myOU, C=IT'
5151 'CN=Some Name/OU=myOU/C=IT'.
5153 Add options '-batch' and '-verbose' to 'openssl req'.
5154 [Massimiliano Pala <madwolf@hackmasters.net>]
5156 *) Introduce the possibility to access global variables through
5157 functions on platform were that's the best way to handle exporting
5158 global variables in shared libraries. To enable this functionality,
5159 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5160 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5161 is normally done by Configure or something similar).
5163 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5164 in the source file (foo.c) like this:
5166 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5167 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5169 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5170 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5172 OPENSSL_DECLARE_GLOBAL(int,foo);
5173 #define foo OPENSSL_GLOBAL_REF(foo)
5174 OPENSSL_DECLARE_GLOBAL(double,bar);
5175 #define bar OPENSSL_GLOBAL_REF(bar)
5177 The #defines are very important, and therefore so is including the
5178 header file everywhere where the defined globals are used.
5180 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5181 of ASN.1 items, but that structure is a bit different.
5183 The largest change is in util/mkdef.pl which has been enhanced with
5184 better and easier to understand logic to choose which symbols should
5185 go into the Windows .def files as well as a number of fixes and code
5186 cleanup (among others, algorithm keywords are now sorted
5187 lexicographically to avoid constant rewrites).
5190 *) In BN_div() keep a copy of the sign of 'num' before writing the
5191 result to 'rm' because if rm==num the value will be overwritten
5192 and produce the wrong result if 'num' is negative: this caused
5193 problems with BN_mod() and BN_nnmod().
5196 *) Function OCSP_request_verify(). This checks the signature on an
5197 OCSP request and verifies the signer certificate. The signer
5198 certificate is just checked for a generic purpose and OCSP request
5202 *) Add OCSP_check_validity() function to check the validity of OCSP
5203 responses. OCSP responses are prepared in real time and may only
5204 be a few seconds old. Simply checking that the current time lies
5205 between thisUpdate and nextUpdate max reject otherwise valid responses
5206 caused by either OCSP responder or client clock inaccuracy. Instead
5207 we allow thisUpdate and nextUpdate to fall within a certain period of
5208 the current time. The age of the response can also optionally be
5209 checked. Two new options -validity_period and -status_age added to
5213 *) If signature or public key algorithm is unrecognized print out its
5214 OID rather that just UNKNOWN.
5217 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5218 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5219 ID to be generated from the issuer certificate alone which can then be
5220 passed to OCSP_id_issuer_cmp().
5223 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5224 ASN1 modules to export functions returning ASN1_ITEM pointers
5225 instead of the ASN1_ITEM structures themselves. This adds several
5226 new macros which allow the underlying ASN1 function/structure to
5227 be accessed transparently. As a result code should not use ASN1_ITEM
5228 references directly (such as &X509_it) but instead use the relevant
5229 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5230 use of the new ASN1 code on platforms where exporting structures
5231 is problematical (for example in shared libraries) but exporting
5232 functions returning pointers to structures is not.
5235 *) Add support for overriding the generation of SSL/TLS session IDs.
5236 These callbacks can be registered either in an SSL_CTX or per SSL.
5237 The purpose of this is to allow applications to control, if they wish,
5238 the arbitrary values chosen for use as session IDs, particularly as it
5239 can be useful for session caching in multiple-server environments. A
5240 command-line switch for testing this (and any client code that wishes
5241 to use such a feature) has been added to "s_server".
5242 [Geoff Thorpe, Lutz Jaenicke]
5244 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5245 of the form '#if defined(...) || defined(...) || ...' and
5246 '#if !defined(...) && !defined(...) && ...'. This also avoids
5247 the growing number of special cases it was previously handling.
5250 *) Make all configuration macros available for application by making
5251 sure they are available in opensslconf.h, by giving them names starting
5252 with "OPENSSL_" to avoid conflicts with other packages and by making
5253 sure e_os2.h will cover all platform-specific cases together with
5255 Additionally, it is now possible to define configuration/platform-
5256 specific names (called "system identities"). In the C code, these
5257 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
5258 macro with the name beginning with "OPENSSL_SYS_", which is determined
5259 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
5263 *) New option -set_serial to 'req' and 'x509' this allows the serial
5264 number to use to be specified on the command line. Previously self
5265 signed certificates were hard coded with serial number 0 and the
5266 CA options of 'x509' had to use a serial number in a file which was
5270 *) New options to 'ca' utility to support V2 CRL entry extensions.
5271 Currently CRL reason, invalidity date and hold instruction are
5272 supported. Add new CRL extensions to V3 code and some new objects.
5275 *) New function EVP_CIPHER_CTX_set_padding() this is used to
5276 disable standard block padding (aka PKCS#5 padding) in the EVP
5277 API, which was previously mandatory. This means that the data is
5278 not padded in any way and so the total length much be a multiple
5279 of the block size, otherwise an error occurs.
5282 *) Initial (incomplete) OCSP SSL support.
5285 *) New function OCSP_parse_url(). This splits up a URL into its host,
5286 port and path components: primarily to parse OCSP URLs. New -url
5287 option to ocsp utility.
5290 *) New nonce behavior. The return value of OCSP_check_nonce() now
5291 reflects the various checks performed. Applications can decide
5292 whether to tolerate certain situations such as an absent nonce
5293 in a response when one was present in a request: the ocsp application
5294 just prints out a warning. New function OCSP_add1_basic_nonce()
5295 this is to allow responders to include a nonce in a response even if
5296 the request is nonce-less.
5299 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
5300 skipped when using openssl x509 multiple times on a single input file,
5301 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
5304 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
5305 set string type: to handle setting ASN1_TIME structures. Fix ca
5306 utility to correctly initialize revocation date of CRLs.
5309 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
5310 the clients preferred ciphersuites and rather use its own preferences.
5311 Should help to work around M$ SGC (Server Gated Cryptography) bug in
5312 Internet Explorer by ensuring unchanged hash method during stepup.
5313 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
5316 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
5317 to aes and add a new 'exist' option to print out symbols that don't
5321 *) Additional options to ocsp utility to allow flags to be set and
5322 additional certificates supplied.
5325 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
5326 OCSP client a number of certificate to only verify the response
5330 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
5331 handle the new API. Currently only ECB, CBC modes supported. Add new
5334 Add TLS AES ciphersuites as described in RFC3268, "Advanced
5335 Encryption Standard (AES) Ciphersuites for Transport Layer
5336 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
5337 not enabled by default and were not part of the "ALL" ciphersuite
5338 alias because they were not yet official; they could be
5339 explicitly requested by specifying the "AESdraft" ciphersuite
5340 group alias. In the final release of OpenSSL 0.9.7, the group
5341 alias is called "AES" and is part of "ALL".)
5342 [Ben Laurie, Steve Henson, Bodo Moeller]
5344 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
5345 request to response.
5348 *) Functions for OCSP responders. OCSP_request_onereq_count(),
5349 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
5350 extract information from a certificate request. OCSP_response_create()
5351 creates a response and optionally adds a basic response structure.
5352 OCSP_basic_add1_status() adds a complete single response to a basic
5353 response and returns the OCSP_SINGLERESP structure just added (to allow
5354 extensions to be included for example). OCSP_basic_add1_cert() adds a
5355 certificate to a basic response and OCSP_basic_sign() signs a basic
5356 response with various flags. New helper functions ASN1_TIME_check()
5357 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
5358 (converts ASN1_TIME to GeneralizedTime).
5361 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
5362 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
5363 structure from a certificate. X509_pubkey_digest() digests the public_key
5364 contents: this is used in various key identifiers.
5367 *) Make sk_sort() tolerate a NULL argument.
5368 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
5370 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
5371 passed by the function are trusted implicitly. If any of them signed the
5372 response then it is assumed to be valid and is not verified.
5375 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
5376 to data. This was previously part of the PKCS7 ASN1 code. This
5377 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
5378 [Steve Henson, reported by Kenneth R. Robinette
5379 <support@securenetterm.com>]
5381 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
5382 routines: without these tracing memory leaks is very painful.
5383 Fix leaks in PKCS12 and PKCS7 routines.
5386 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
5387 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
5388 effectively meant GeneralizedTime would never be used. Now it
5389 is initialised to -1 but X509_time_adj() now has to check the value
5390 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
5391 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
5392 [Steve Henson, reported by Kenneth R. Robinette
5393 <support@securenetterm.com>]
5395 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
5396 result in a zero length in the ASN1_INTEGER structure which was
5397 not consistent with the structure when d2i_ASN1_INTEGER() was used
5398 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
5399 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
5400 where it did not print out a minus for negative ASN1_INTEGER.
5403 *) Add summary printout to ocsp utility. The various functions which
5404 convert status values to strings have been renamed to:
5405 OCSP_response_status_str(), OCSP_cert_status_str() and
5406 OCSP_crl_reason_str() and are no longer static. New options
5407 to verify nonce values and to disable verification. OCSP response
5408 printout format cleaned up.
5411 *) Add additional OCSP certificate checks. These are those specified
5412 in RFC2560. This consists of two separate checks: the CA of the
5413 certificate being checked must either be the OCSP signer certificate
5414 or the issuer of the OCSP signer certificate. In the latter case the
5415 OCSP signer certificate must contain the OCSP signing extended key
5416 usage. This check is performed by attempting to match the OCSP
5417 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
5418 in the OCSP_CERTID structures of the response.
5421 *) Initial OCSP certificate verification added to OCSP_basic_verify()
5422 and related routines. This uses the standard OpenSSL certificate
5423 verify routines to perform initial checks (just CA validity) and
5424 to obtain the certificate chain. Then additional checks will be
5425 performed on the chain. Currently the root CA is checked to see
5426 if it is explicitly trusted for OCSP signing. This is used to set
5427 a root CA as a global signing root: that is any certificate that
5428 chains to that CA is an acceptable OCSP signing certificate.
5431 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5432 extensions from a separate configuration file.
5433 As when reading extensions from the main configuration file,
5434 the '-extensions ...' option may be used for specifying the
5436 [Massimiliano Pala <madwolf@comune.modena.it>]
5438 *) New OCSP utility. Allows OCSP requests to be generated or
5439 read. The request can be sent to a responder and the output
5440 parsed, outputed or printed in text form. Not complete yet:
5441 still needs to check the OCSP response validity.
5444 *) New subcommands for 'openssl ca':
5445 'openssl ca -status <serial>' prints the status of the cert with
5446 the given serial number (according to the index file).
5447 'openssl ca -updatedb' updates the expiry status of certificates
5449 [Massimiliano Pala <madwolf@comune.modena.it>]
5451 *) New '-newreq-nodes' command option to CA.pl. This is like
5452 '-newreq', but calls 'openssl req' with the '-nodes' option
5453 so that the resulting key is not encrypted.
5454 [Damien Miller <djm@mindrot.org>]
5456 *) New configuration for the GNU Hurd.
5457 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5459 *) Initial code to implement OCSP basic response verify. This
5460 is currently incomplete. Currently just finds the signer's
5461 certificate and verifies the signature on the response.
5464 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5465 value of OPENSSLDIR. This is available via the new '-d' option
5466 to 'openssl version', and is also included in 'openssl version -a'.
5469 *) Allowing defining memory allocation callbacks that will be given
5470 file name and line number information in additional arguments
5471 (a const char* and an int). The basic functionality remains, as
5472 well as the original possibility to just replace malloc(),
5473 realloc() and free() by functions that do not know about these
5474 additional arguments. To register and find out the current
5475 settings for extended allocation functions, the following
5476 functions are provided:
5478 CRYPTO_set_mem_ex_functions
5479 CRYPTO_set_locked_mem_ex_functions
5480 CRYPTO_get_mem_ex_functions
5481 CRYPTO_get_locked_mem_ex_functions
5483 These work the same way as CRYPTO_set_mem_functions and friends.
5484 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5485 extended allocation function is enabled.
5486 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5487 a conventional allocation function is enabled.
5488 [Richard Levitte, Bodo Moeller]
5490 *) Finish off removing the remaining LHASH function pointer casts.
5491 There should no longer be any prototype-casting required when using
5492 the LHASH abstraction, and any casts that remain are "bugs". See
5493 the callback types and macros at the head of lhash.h for details
5494 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5497 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5498 If /dev/[u]random devices are not available or do not return enough
5499 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5501 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5502 /etc/entropy will be queried once each in this sequence, quering stops
5503 when enough entropy was collected without querying more sockets.
5506 *) Change the Unix RAND_poll() variant to be able to poll several
5507 random devices, as specified by DEVRANDOM, until a sufficient amount
5508 of data has been collected. We spend at most 10 ms on each file
5509 (select timeout) and read in non-blocking mode. DEVRANDOM now
5510 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5511 (previously it was just the string "/dev/urandom"), so on typical
5512 platforms the 10 ms delay will never occur.
5513 Also separate out the Unix variant to its own file, rand_unix.c.
5514 For VMS, there's a currently-empty rand_vms.c.
5517 *) Move OCSP client related routines to ocsp_cl.c. These
5518 provide utility functions which an application needing
5519 to issue a request to an OCSP responder and analyse the
5520 response will typically need: as opposed to those which an
5521 OCSP responder itself would need which will be added later.
5523 OCSP_request_sign() signs an OCSP request with an API similar
5524 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5525 response. OCSP_response_get1_basic() extracts basic response
5526 from response. OCSP_resp_find_status(): finds and extracts status
5527 information from an OCSP_CERTID structure (which will be created
5528 when the request structure is built). These are built from lower
5529 level functions which work on OCSP_SINGLERESP structures but
5530 wont normally be used unless the application wishes to examine
5531 extensions in the OCSP response for example.
5533 Replace nonce routines with a pair of functions.
5534 OCSP_request_add1_nonce() adds a nonce value and optionally
5535 generates a random value. OCSP_check_nonce() checks the
5536 validity of the nonce in an OCSP response.
5539 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5540 This doesn't copy the supplied OCSP_CERTID and avoids the
5541 need to free up the newly created id. Change return type
5542 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5543 This can then be used to add extensions to the request.
5544 Deleted OCSP_request_new(), since most of its functionality
5545 is now in OCSP_REQUEST_new() (and the case insensitive name
5546 clash) apart from the ability to set the request name which
5547 will be added elsewhere.
5550 *) Update OCSP API. Remove obsolete extensions argument from
5551 various functions. Extensions are now handled using the new
5552 OCSP extension code. New simple OCSP HTTP function which
5553 can be used to send requests and parse the response.
5556 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5557 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5558 uses the special reorder version of SET OF to sort the attributes
5559 and reorder them to match the encoded order. This resolves a long
5560 standing problem: a verify on a PKCS7 structure just after signing
5561 it used to fail because the attribute order did not match the
5562 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5563 it uses the received order. This is necessary to tolerate some broken
5564 software that does not order SET OF. This is handled by encoding
5565 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5566 to produce the required SET OF.
5569 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5570 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5571 files to get correct declarations of the ASN.1 item variables.
5574 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5575 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5576 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5577 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5578 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5579 ASN1_ITEM and no wrapper functions.
5582 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5583 replace the old function pointer based I/O routines. Change most of
5584 the *_d2i_bio() and *_d2i_fp() functions to use these.
5587 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5588 lines, recognice more "algorithms" that can be deselected, and make
5589 it complain about algorithm deselection that isn't recognised.
5592 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5593 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5594 to use new functions. Add NO_ASN1_OLD which can be set to remove
5595 some old style ASN1 functions: this can be used to determine if old
5596 code will still work when these eventually go away.
5599 *) New extension functions for OCSP structures, these follow the
5600 same conventions as certificates and CRLs.
5603 *) New function X509V3_add1_i2d(). This automatically encodes and
5604 adds an extension. Its behaviour can be customised with various
5605 flags to append, replace or delete. Various wrappers added for
5606 certifcates and CRLs.
5609 *) Fix to avoid calling the underlying ASN1 print routine when
5610 an extension cannot be parsed. Correct a typo in the
5611 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5614 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5615 entries for variables.
5618 *) Add functionality to apps/openssl.c for detecting locking
5619 problems: As the program is single-threaded, all we have
5620 to do is register a locking callback using an array for
5621 storing which locks are currently held by the program.
5624 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5625 SSL_get_ex_data_X509_STORE_idx(), which is used in
5626 ssl_verify_cert_chain() and thus can be called at any time
5627 during TLS/SSL handshakes so that thread-safety is essential.
5628 Unfortunately, the ex_data design is not at all suited
5629 for multi-threaded use, so it probably should be abolished.
5632 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5633 [Broadcom, tweaked and integrated by Geoff Thorpe]
5635 *) Move common extension printing code to new function
5636 X509V3_print_extensions(). Reorganise OCSP print routines and
5637 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5640 *) New function X509_signature_print() to remove duplication in some
5644 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5645 set (this was treated exactly the same as SET OF previously). This
5646 is used to reorder the STACK representing the structure to match the
5647 encoding. This will be used to get round a problem where a PKCS7
5648 structure which was signed could not be verified because the STACK
5649 order did not reflect the encoded order.
5652 *) Reimplement the OCSP ASN1 module using the new code.
5655 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5656 for its ASN1 operations. The old style function pointers still exist
5657 for now but they will eventually go away.
5660 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5661 completely replaces the old ASN1 functionality with a table driven
5662 encoder and decoder which interprets an ASN1_ITEM structure describing
5663 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5664 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5665 has also been converted to the new form.
5668 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5669 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5670 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5671 for negative moduli.
5674 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5675 of not touching the result's sign bit.
5678 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5682 *) Changed the LHASH code to use prototypes for callbacks, and created
5683 macros to declare and implement thin (optionally static) functions
5684 that provide type-safety and avoid function pointer casting for the
5685 type-specific callbacks.
5688 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5690 [Veers Staats <staatsvr@asc.hpc.mil>,
5691 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5693 *) Reformat the FAQ so the different questions and answers can be divided
5694 in sections depending on the subject.
5697 *) Have the zlib compression code load ZLIB.DLL dynamically under
5701 *) New function BN_mod_sqrt for computing square roots modulo a prime
5702 (using the probabilistic Tonelli-Shanks algorithm unless
5703 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5704 be handled deterministically).
5705 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5707 *) Make BN_mod_inverse faster by explicitly handling small quotients
5708 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5709 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5712 *) New function BN_kronecker.
5715 *) Fix BN_gcd so that it works on negative inputs; the result is
5716 positive unless both parameters are zero.
5717 Previously something reasonably close to an infinite loop was
5718 possible because numbers could be growing instead of shrinking
5719 in the implementation of Euclid's algorithm.
5722 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5723 sign of the number in question.
5725 Fix BN_is_word(a,w) to work correctly for w == 0.
5727 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5728 because its test if the absolute value of 'a' equals 'w'.
5729 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5730 it exists mostly for use in the implementations of BN_is_zero(),
5731 BN_is_one(), and BN_is_word().
5734 *) New function BN_swap.
5737 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5738 the exponentiation functions are more likely to produce reasonable
5739 results on negative inputs.
5742 *) Change BN_mod_mul so that the result is always non-negative.
5743 Previously, it could be negative if one of the factors was negative;
5744 I don't think anyone really wanted that behaviour.
5747 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5748 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5749 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5750 and add new functions:
5759 BN_mod_lshift1_quick
5763 These functions always generate non-negative results.
5765 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5766 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5768 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5769 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5770 be reduced modulo m.
5771 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5774 The following entry accidentily appeared in the CHANGES file
5775 distributed with OpenSSL 0.9.7. The modifications described in
5776 it do *not* apply to OpenSSL 0.9.7.
5778 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5779 was actually never needed) and in BN_mul(). The removal in BN_mul()
5780 required a small change in bn_mul_part_recursive() and the addition
5781 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5782 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5783 bn_sub_words() and bn_add_words() except they take arrays with
5788 *) In 'openssl passwd', verify passwords read from the terminal
5789 unless the '-salt' option is used (which usually means that
5790 verification would just waste user's time since the resulting
5791 hash is going to be compared with some given password hash)
5792 or the new '-noverify' option is used.
5794 This is an incompatible change, but it does not affect
5795 non-interactive use of 'openssl passwd' (passwords on the command
5796 line, '-stdin' option, '-in ...' option) and thus should not
5800 *) Remove all references to RSAref, since there's no more need for it.
5803 *) Make DSO load along a path given through an environment variable
5804 (SHLIB_PATH) with shl_load().
5807 *) Constify the ENGINE code as a result of BIGNUM constification.
5808 Also constify the RSA code and most things related to it. In a
5809 few places, most notable in the depth of the ASN.1 code, ugly
5810 casts back to non-const were required (to be solved at a later
5814 *) Make it so the openssl application has all engines loaded by default.
5817 *) Constify the BIGNUM routines a little more.
5820 *) Add the following functions:
5822 ENGINE_load_cswift()
5824 ENGINE_load_atalla()
5826 ENGINE_load_builtin_engines()
5828 That way, an application can itself choose if external engines that
5829 are built-in in OpenSSL shall ever be used or not. The benefit is
5830 that applications won't have to be linked with libdl or other dso
5831 libraries unless it's really needed.
5833 Changed 'openssl engine' to load all engines on demand.
5834 Changed the engine header files to avoid the duplication of some
5835 declarations (they differed!).
5838 *) 'openssl engine' can now list capabilities.
5841 *) Better error reporting in 'openssl engine'.
5844 *) Never call load_dh_param(NULL) in s_server.
5847 *) Add engine application. It can currently list engines by name and
5848 identity, and test if they are actually available.
5851 *) Improve RPM specification file by forcing symbolic linking and making
5852 sure the installed documentation is also owned by root.root.
5853 [Damien Miller <djm@mindrot.org>]
5855 *) Give the OpenSSL applications more possibilities to make use of
5856 keys (public as well as private) handled by engines.
5859 *) Add OCSP code that comes from CertCo.
5862 *) Add VMS support for the Rijndael code.
5865 *) Added untested support for Nuron crypto accelerator.
5868 *) Add support for external cryptographic devices. This code was
5869 previously distributed separately as the "engine" branch.
5870 [Geoff Thorpe, Richard Levitte]
5872 *) Rework the filename-translation in the DSO code. It is now possible to
5873 have far greater control over how a "name" is turned into a filename
5874 depending on the operating environment and any oddities about the
5875 different shared library filenames on each system.
5878 *) Support threads on FreeBSD-elf in Configure.
5881 *) Fix for SHA1 assembly problem with MASM: it produces
5882 warnings about corrupt line number information when assembling
5883 with debugging information. This is caused by the overlapping
5885 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5888 NCONF_get_number() has no error checking at all. As a replacement,
5889 NCONF_get_number_e() is defined (_e for "error checking") and is
5890 promoted strongly. The old NCONF_get_number is kept around for
5891 binary backward compatibility.
5892 Make it possible for methods to load from something other than a BIO,
5893 by providing a function pointer that is given a name instead of a BIO.
5894 For example, this could be used to load configuration data from an
5898 *) Fix for non blocking accept BIOs. Added new I/O special reason
5899 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5900 with non blocking I/O was not possible because no retry code was
5901 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5905 *) Added the beginnings of Rijndael support.
5908 *) Fix for bug in DirectoryString mask setting. Add support for
5909 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5910 to allow certificate printing to more controllable, additional
5911 'certopt' option to 'x509' to allow new printing options to be
5915 *) Clean old EAY MD5 hack from e_os.h.
5918 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5920 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5921 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5922 [Joe Orton, Steve Henson]
5924 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5926 *) Fix additional bug revealed by the NISCC test suite:
5928 Stop bug triggering large recursion when presented with
5929 certain ASN.1 tags (CVE-2003-0851)
5932 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5934 *) Fix various bugs revealed by running the NISCC test suite:
5936 Stop out of bounds reads in the ASN1 code when presented with
5937 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5939 If verify callback ignores invalid public key errors don't try to check
5940 certificate signature with the NULL public key.
5944 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5945 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5949 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5950 extra data after the compression methods not only for TLS 1.0
5951 but also for SSL 3.0 (as required by the specification).
5952 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5954 *) Change X509_certificate_type() to mark the key as exported/exportable
5955 when it's 512 *bits* long, not 512 bytes.
5958 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5960 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5961 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5962 a protocol version number mismatch like a decryption error
5963 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5966 *) Turn on RSA blinding by default in the default implementation
5967 to avoid a timing attack. Applications that don't want it can call
5968 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5969 They would be ill-advised to do so in most cases.
5970 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5972 *) Change RSA blinding code so that it works when the PRNG is not
5973 seeded (in this case, the secret RSA exponent is abused as
5974 an unpredictable seed -- if it is not unpredictable, there
5975 is no point in blinding anyway). Make RSA blinding thread-safe
5976 by remembering the creator's thread ID in rsa->blinding and
5977 having all other threads use local one-time blinding factors
5978 (this requires more computation than sharing rsa->blinding, but
5979 avoids excessive locking; and if an RSA object is not shared
5980 between threads, blinding will still be very fast).
5983 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5985 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5986 via timing by performing a MAC computation even if incorrrect
5987 block cipher padding has been found. This is a countermeasure
5988 against active attacks where the attacker has to distinguish
5989 between bad padding and a MAC verification error. (CVE-2003-0078)
5991 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5992 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5993 Martin Vuagnoux (EPFL, Ilion)]
5995 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5997 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5998 memory from it's contents. This is done with a counter that will
5999 place alternating values in each byte. This can be used to solve
6000 two issues: 1) the removal of calls to memset() by highly optimizing
6001 compilers, and 2) cleansing with other values than 0, since those can
6002 be read through on certain media, for example a swap space on disk.
6005 *) Bugfix: client side session caching did not work with external caching,
6006 because the session->cipher setting was not restored when reloading
6007 from the external cache. This problem was masked, when
6008 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
6009 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
6012 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
6013 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
6014 [Zeev Lieber <zeev-l@yahoo.com>]
6016 *) Undo an undocumented change introduced in 0.9.6e which caused
6017 repeated calls to OpenSSL_add_all_ciphers() and
6018 OpenSSL_add_all_digests() to be ignored, even after calling
6022 *) Change the default configuration reader to deal with last line not
6023 being properly terminated.
6026 *) Change X509_NAME_cmp() so it applies the special rules on handling
6027 DN values that are of type PrintableString, as well as RDNs of type
6028 emailAddress where the value has the type ia5String.
6029 [stefank@valicert.com via Richard Levitte]
6031 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
6032 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
6033 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
6034 the bitwise-OR of the two for use by the majority of applications
6035 wanting this behaviour, and update the docs. The documented
6036 behaviour and actual behaviour were inconsistent and had been
6037 changing anyway, so this is more a bug-fix than a behavioural
6039 [Geoff Thorpe, diagnosed by Nadav Har'El]
6041 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
6042 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
6045 *) Fix initialization code race conditions in
6046 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
6047 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
6048 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
6049 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
6050 ssl2_get_cipher_by_char(),
6051 ssl3_get_cipher_by_char().
6052 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
6054 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
6055 the cached sessions are flushed, as the remove_cb() might use ex_data
6056 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
6057 (see [openssl.org #212]).
6058 [Geoff Thorpe, Lutz Jaenicke]
6060 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
6061 length, instead of the encoding length to d2i_ASN1_OBJECT.
6064 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
6066 *) [In 0.9.6g-engine release:]
6067 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
6068 [Lynn Gazis <lgazis@rainbow.com>]
6070 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
6072 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
6073 and get fix the header length calculation.
6074 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
6075 Alon Kantor <alonk@checkpoint.com> (and others),
6078 *) Use proper error handling instead of 'assertions' in buffer
6079 overflow checks added in 0.9.6e. This prevents DoS (the
6080 assertions could call abort()).
6081 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
6083 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
6085 *) Add various sanity checks to asn1_get_length() to reject
6086 the ASN1 length bytes if they exceed sizeof(long), will appear
6087 negative or the content length exceeds the length of the
6089 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6091 *) Fix cipher selection routines: ciphers without encryption had no flags
6092 for the cipher strength set and where therefore not handled correctly
6093 by the selection routines (PR #130).
6096 *) Fix EVP_dsa_sha macro.
6100 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
6101 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
6102 that was added in OpenSSL 0.9.6d.
6104 As the countermeasure turned out to be incompatible with some
6105 broken SSL implementations, the new option is part of SSL_OP_ALL.
6106 SSL_OP_ALL is usually employed when compatibility with weird SSL
6107 implementations is desired (e.g. '-bugs' option to 's_client' and
6108 's_server'), so the new option is automatically set in many
6112 *) Changes in security patch:
6114 Changes marked "(CHATS)" were sponsored by the Defense Advanced
6115 Research Projects Agency (DARPA) and Air Force Research Laboratory,
6116 Air Force Materiel Command, USAF, under agreement number
6119 *) Add various sanity checks to asn1_get_length() to reject
6120 the ASN1 length bytes if they exceed sizeof(long), will appear
6121 negative or the content length exceeds the length of the
6122 supplied buffer. (CVE-2002-0659)
6123 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6125 *) Assertions for various potential buffer overflows, not known to
6127 [Ben Laurie (CHATS)]
6129 *) Various temporary buffers to hold ASCII versions of integers were
6130 too small for 64 bit platforms. (CVE-2002-0655)
6131 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6133 *) Remote buffer overflow in SSL3 protocol - an attacker could
6134 supply an oversized session ID to a client. (CVE-2002-0656)
6135 [Ben Laurie (CHATS)]
6137 *) Remote buffer overflow in SSL2 protocol - an attacker could
6138 supply an oversized client master key. (CVE-2002-0656)
6139 [Ben Laurie (CHATS)]
6141 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6143 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6144 encoded as NULL) with id-dsa-with-sha1.
6145 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6147 *) Check various X509_...() return values in apps/req.c.
6148 [Nils Larsch <nla@trustcenter.de>]
6150 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6151 an end-of-file condition would erronously be flagged, when the CRLF
6152 was just at the end of a processed block. The bug was discovered when
6153 processing data through a buffering memory BIO handing the data to a
6154 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6155 <ptsekov@syntrex.com> and Nedelcho Stanev.
6158 *) Implement a countermeasure against a vulnerability recently found
6159 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6160 before application data chunks to avoid the use of known IVs
6161 with data potentially chosen by the attacker.
6164 *) Fix length checks in ssl3_get_client_hello().
6167 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6168 to prevent ssl3_read_internal() from incorrectly assuming that
6169 ssl3_read_bytes() found application data while handshake
6170 processing was enabled when in fact s->s3->in_read_app_data was
6171 merely automatically cleared during the initial handshake.
6172 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6174 *) Fix object definitions for Private and Enterprise: they were not
6175 recognized in their shortname (=lowercase) representation. Extend
6176 obj_dat.pl to issue an error when using undefined keywords instead
6177 of silently ignoring the problem (Svenning Sorensen
6178 <sss@sss.dnsalias.net>).
6181 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6182 generators, i.e. generators other than 2 and 5. (Previously, the
6183 code did not properly initialise the 'add' and 'rem' values to
6184 BN_generate_prime().)
6186 In the new general case, we do not insist that 'generator' is
6187 actually a primitive root: This requirement is rather pointless;
6188 a generator of the order-q subgroup is just as good, if not
6192 *) Map new X509 verification errors to alerts. Discovered and submitted by
6193 Tom Wu <tom@arcot.com>.
6196 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6197 returning non-zero before the data has been completely received
6198 when using non-blocking I/O.
6199 [Bodo Moeller; problem pointed out by John Hughes]
6201 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6202 [Ben Laurie, Lutz Jaenicke]
6204 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6205 Yoram Zahavi <YoramZ@gilian.com>).
6208 *) Add information about CygWin 1.3 and on, and preserve proper
6209 configuration for the versions before that.
6210 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6212 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6213 check whether we deal with a copy of a session and do not delete from
6214 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6215 <izhar@checkpoint.com>.
6218 *) Do not store session data into the internal session cache, if it
6219 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6220 flag is set). Proposed by Aslam <aslam@funk.com>.
6223 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6227 *) [In 0.9.6d-engine release:]
6228 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6229 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6231 *) Add the configuration target linux-s390x.
6232 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6234 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6235 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6236 variable as an indication that a ClientHello message has been
6237 received. As the flag value will be lost between multiple
6238 invocations of ssl3_accept when using non-blocking I/O, the
6239 function may not be aware that a handshake has actually taken
6240 place, thus preventing a new session from being added to the
6243 To avoid this problem, we now set s->new_session to 2 instead of
6244 using a local variable.
6245 [Lutz Jaenicke, Bodo Moeller]
6247 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6248 if the SSL_R_LENGTH_MISMATCH error is detected.
6249 [Geoff Thorpe, Bodo Moeller]
6251 *) New 'shared_ldflag' column in Configure platform table.
6254 *) Fix EVP_CIPHER_mode macro.
6255 ["Dan S. Camper" <dan@bti.net>]
6257 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
6258 type, we must throw them away by setting rr->length to 0.
6259 [D P Chang <dpc@qualys.com>]
6261 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
6263 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
6264 <Dominikus.Scherkl@biodata.com>. (The previous implementation
6265 worked incorrectly for those cases where range = 10..._2 and
6266 3*range is two bits longer than range.)
6269 *) Only add signing time to PKCS7 structures if it is not already
6273 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
6274 OBJ_ld_ce should be OBJ_id_ce.
6275 Also some ip-pda OIDs in crypto/objects/objects.txt were
6276 incorrect (cf. RFC 3039).
6277 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
6279 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
6280 returns early because it has nothing to do.
6281 [Andy Schneider <andy.schneider@bjss.co.uk>]
6283 *) [In 0.9.6c-engine release:]
6284 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
6285 [Andy Schneider <andy.schneider@bjss.co.uk>]
6287 *) [In 0.9.6c-engine release:]
6288 Add support for Cryptographic Appliance's keyserver technology.
6289 (Use engine 'keyclient')
6290 [Cryptographic Appliances and Geoff Thorpe]
6292 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
6293 is called via tools/c89.sh because arguments have to be
6294 rearranged (all '-L' options must appear before the first object
6296 [Richard Shapiro <rshapiro@abinitio.com>]
6298 *) [In 0.9.6c-engine release:]
6299 Add support for Broadcom crypto accelerator cards, backported
6301 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
6303 *) [In 0.9.6c-engine release:]
6304 Add support for SureWare crypto accelerator cards from
6305 Baltimore Technologies. (Use engine 'sureware')
6306 [Baltimore Technologies and Mark Cox]
6308 *) [In 0.9.6c-engine release:]
6309 Add support for crypto accelerator cards from Accelerated
6310 Encryption Processing, www.aep.ie. (Use engine 'aep')
6311 [AEP Inc. and Mark Cox]
6313 *) Add a configuration entry for gcc on UnixWare.
6314 [Gary Benson <gbenson@redhat.com>]
6316 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
6317 messages are stored in a single piece (fixed-length part and
6318 variable-length part combined) and fix various bugs found on the way.
6321 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
6322 instead. BIO_gethostbyname() does not know what timeouts are
6323 appropriate, so entries would stay in cache even when they have
6325 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
6327 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
6328 faced with a pathologically small ClientHello fragment that does
6329 not contain client_version: Instead of aborting with an error,
6330 simply choose the highest available protocol version (i.e.,
6331 TLS 1.0 unless it is disabled). In practice, ClientHello
6332 messages are never sent like this, but this change gives us
6333 strictly correct behaviour at least for TLS.
6336 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
6337 never resets s->method to s->ctx->method when called from within
6338 one of the SSL handshake functions.
6339 [Bodo Moeller; problem pointed out by Niko Baric]
6341 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
6342 (sent using the client's version number) if client_version is
6343 smaller than the protocol version in use. Also change
6344 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
6345 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
6346 the client will at least see that alert.
6349 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
6353 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
6354 client receives HelloRequest while in a handshake.
6355 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
6357 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
6358 should end in 'break', not 'goto end' which circuments various
6359 cleanups done in state SSL_ST_OK. But session related stuff
6360 must be disabled for SSL_ST_OK in the case that we just sent a
6363 Also avoid some overhead by not calling ssl_init_wbio_buffer()
6364 before just sending a HelloRequest.
6365 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
6367 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
6368 reveal whether illegal block cipher padding was found or a MAC
6369 verification error occured. (Neither SSLerr() codes nor alerts
6370 are directly visible to potential attackers, but the information
6371 may leak via logfiles.)
6373 Similar changes are not required for the SSL 2.0 implementation
6374 because the number of padding bytes is sent in clear for SSL 2.0,
6375 and the extra bytes are just ignored. However ssl/s2_pkt.c
6376 failed to verify that the purported number of padding bytes is in
6380 *) Add OpenUNIX-8 support including shared libraries
6381 (Boyd Lynn Gerber <gerberb@zenez.com>).
6384 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
6385 'wristwatch attack' using huge encoding parameters (cf.
6386 James H. Manger's CRYPTO 2001 paper). Note that the
6387 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
6388 encoding parameters and hence was not vulnerable.
6391 *) BN_sqr() bug fix.
6392 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
6394 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
6395 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
6396 followed by modular reduction.
6397 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
6399 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
6400 equivalent based on BN_pseudo_rand() instead of BN_rand().
6403 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
6404 This function was broken, as the check for a new client hello message
6405 to handle SGC did not allow these large messages.
6406 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
6409 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
6412 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
6413 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
6416 *) Rework the configuration and shared library support for Tru64 Unix.
6417 The configuration part makes use of modern compiler features and
6418 still retains old compiler behavior for those that run older versions
6419 of the OS. The shared library support part includes a variant that
6420 uses the RPATH feature, and is available through the special
6421 configuration target "alpha-cc-rpath", which will never be selected
6423 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
6425 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
6426 with the same message size as in ssl3_get_certificate_request().
6427 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
6428 messages might inadvertently be reject as too long.
6429 [Petr Lampa <lampa@fee.vutbr.cz>]
6431 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6434 *) Modified SSL library such that the verify_callback that has been set
6435 specificly for an SSL object with SSL_set_verify() is actually being
6436 used. Before the change, a verify_callback set with this function was
6437 ignored and the verify_callback() set in the SSL_CTX at the time of
6438 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6439 to allow the necessary settings.
6442 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6443 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6444 done automatically (in contradiction to the requirements of the C
6445 standard). This made problems when used from OpenSSH.
6448 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6449 dh->length and always used
6451 BN_rand_range(priv_key, dh->p).
6453 BN_rand_range() is not necessary for Diffie-Hellman, and this
6454 specific range makes Diffie-Hellman unnecessarily inefficient if
6455 dh->length (recommended exponent length) is much smaller than the
6456 length of dh->p. We could use BN_rand_range() if the order of
6457 the subgroup was stored in the DH structure, but we only have
6462 BN_rand(priv_key, l, ...)
6464 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6470 RSA_eay_public_encrypt
6471 RSA_eay_private_decrypt
6472 RSA_eay_private_encrypt (signing)
6473 RSA_eay_public_decrypt (signature verification)
6475 (default implementations for RSA_public_encrypt,
6476 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6477 always reject numbers >= n.
6480 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6481 to synchronize access to 'locking_thread'. This is necessary on
6482 systems where access to 'locking_thread' (an 'unsigned long'
6483 variable) is not atomic.
6486 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6487 *before* setting the 'crypto_lock_rand' flag. The previous code had
6488 a race condition if 0 is a valid thread ID.
6489 [Travis Vitek <vitek@roguewave.com>]
6491 *) Add support for shared libraries under Irix.
6492 [Albert Chin-A-Young <china@thewrittenword.com>]
6494 *) Add configuration option to build on Linux on both big-endian and
6496 [Ralf Baechle <ralf@uni-koblenz.de>]
6498 *) Add the possibility to create shared libraries on HP-UX.
6501 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6503 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6504 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6505 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6506 PRNG state recovery was possible based on the output of
6507 one PRNG request appropriately sized to gain knowledge on
6508 'md' followed by enough consecutive 1-byte PRNG requests
6509 to traverse all of 'state'.
6511 1. When updating 'md_local' (the current thread's copy of 'md')
6512 during PRNG output generation, hash all of the previous
6513 'md_local' value, not just the half used for PRNG output.
6515 2. Make the number of bytes from 'state' included into the hash
6516 independent from the number of PRNG bytes requested.
6518 The first measure alone would be sufficient to avoid
6519 Markku-Juhani's attack. (Actually it had never occurred
6520 to me that the half of 'md_local' used for chaining was the
6521 half from which PRNG output bytes were taken -- I had always
6522 assumed that the secret half would be used.) The second
6523 measure makes sure that additional data from 'state' is never
6524 mixed into 'md_local' in small portions; this heuristically
6525 further strengthens the PRNG.
6528 *) Fix crypto/bn/asm/mips3.s.
6531 *) When only the key is given to "enc", the IV is undefined. Print out
6532 an error message in this case.
6535 *) Handle special case when X509_NAME is empty in X509 printing routines.
6538 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6539 positive and less than q.
6542 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6543 used: it isn't thread safe and the add_lock_callback should handle
6545 [Paul Rose <Paul.Rose@bridge.com>]
6547 *) Verify that incoming data obeys the block size in
6548 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6552 [Ulf Möller, Bodo Möller]
6554 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6555 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6556 when fixing the server behaviour for backwards-compatible 'client
6557 hello' messages. (Note that the attack is impractical against
6558 SSL 3.0 and TLS 1.0 anyway because length and version checking
6559 means that the probability of guessing a valid ciphertext is
6560 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6563 Before 0.9.5, the countermeasure (hide the error by generating a
6564 random 'decryption result') did not work properly because
6565 ERR_clear_error() was missing, meaning that SSL_get_error() would
6566 detect the supposedly ignored error.
6568 Both problems are now fixed.
6571 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6572 (previously it was 1024).
6575 *) Fix for compatibility mode trust settings: ignore trust settings
6576 unless some valid trust or reject settings are present.
6579 *) Fix for blowfish EVP: its a variable length cipher.
6582 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6583 parameters in DSA public key structures and return an error in the
6584 DSA routines if parameters are absent.
6587 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6588 in the current directory if neither $RANDFILE nor $HOME was set.
6589 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6590 caused some confusion to Windows users who haven't defined $HOME.
6591 Thus RAND_file_name() is changed again: e_os.h can define a
6592 DEFAULT_HOME, which will be used if $HOME is not set.
6593 For Windows, we use "C:"; on other platforms, we still require
6594 environment variables.
6596 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6597 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6598 having multiple threads call RAND_poll() concurrently.
6601 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6602 combination of a flag and a thread ID variable.
6603 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6604 flag), *other* threads can enter ssleay_add_bytes without obeying
6605 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6606 that they do not hold after the first thread unsets add_do_not_lock).
6609 *) Change bctest again: '-x' expressions are not available in all
6613 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6615 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6616 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6618 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6619 the default extension for executables, if any. Also, make the perl
6620 scripts that use symlink() to test if it really exists and use "cp"
6621 if it doesn't. All this made OpenSSL compilable and installable in
6625 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6626 If SEQUENCE is length is indefinite just set c->slen to the total
6627 amount of data available.
6628 [Steve Henson, reported by shige@FreeBSD.org]
6629 [This change does not apply to 0.9.7.]
6631 *) Change bctest to avoid here-documents inside command substitution
6632 (workaround for FreeBSD /bin/sh bug).
6633 For compatibility with Ultrix, avoid shell functions (introduced
6634 in the bctest version that searches along $PATH).
6637 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6638 with des_encrypt() defined on some operating systems, like Solaris
6642 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6643 On the Importance of Eliminating Errors in Cryptographic
6644 Computations, J. Cryptology 14 (2001) 2, 101-119,
6645 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6648 *) MIPS assembler BIGNUM division bug fix.
6651 *) Disabled incorrect Alpha assembler code.
6654 *) Fix PKCS#7 decode routines so they correctly update the length
6655 after reading an EOC for the EXPLICIT tag.
6657 [This change does not apply to 0.9.7.]
6659 *) Fix bug in PKCS#12 key generation routines. This was triggered
6660 if a 3DES key was generated with a 0 initial byte. Include
6661 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6662 (but broken) behaviour.
6665 *) Enhance bctest to search for a working bc along $PATH and print
6667 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6669 *) Fix memory leaks in err.c: free err_data string if necessary;
6670 don't write to the wrong index in ERR_set_error_data.
6673 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6677 *) Replace rdtsc with _emit statements for VC++ version 5.
6678 [Jeremy Cooper <jeremy@baymoo.org>]
6680 *) Make it possible to reuse SSLv2 sessions.
6683 *) In copy_email() check for >= 0 as a return value for
6684 X509_NAME_get_index_by_NID() since 0 is a valid index.
6685 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6687 *) Avoid coredump with unsupported or invalid public keys by checking if
6688 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6689 PKCS7_verify() fails with non detached data.
6692 *) Don't use getenv in library functions when run as setuid/setgid.
6693 New function OPENSSL_issetugid().
6696 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6697 due to incorrect handling of multi-threading:
6699 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6701 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6703 3. Count how many times MemCheck_off() has been called so that
6704 nested use can be treated correctly. This also avoids
6705 inband-signalling in the previous code (which relied on the
6706 assumption that thread ID 0 is impossible).
6709 *) Add "-rand" option also to s_client and s_server.
6712 *) Fix CPU detection on Irix 6.x.
6713 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6714 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6716 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6719 [This change does not apply to 0.9.7.]
6721 *) Use the cached encoding of an X509_NAME structure rather than
6722 copying it. This is apparently the reason for the libsafe "errors"
6723 but the code is actually correct.
6726 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6727 Bleichenbacher's DSA attack.
6728 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6729 to be set and top=0 forces the highest bit to be set; top=-1 is new
6730 and leaves the highest bit random.
6731 [Ulf Moeller, Bodo Moeller]
6733 *) In the NCONF_...-based implementations for CONF_... queries
6734 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6735 a temporary CONF structure with the data component set to NULL
6736 (which gives segmentation faults in lh_retrieve).
6737 Instead, use NULL for the CONF pointer in CONF_get_string and
6738 CONF_get_number (which may use environment variables) and directly
6739 return NULL from CONF_get_section.
6742 *) Fix potential buffer overrun for EBCDIC.
6745 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6746 keyUsage if basicConstraints absent for a CA.
6749 *) Make SMIME_write_PKCS7() write mail header values with a format that
6750 is more generally accepted (no spaces before the semicolon), since
6751 some programs can't parse those values properly otherwise. Also make
6752 sure BIO's that break lines after each write do not create invalid
6756 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6757 macros previously used would not encode an empty SEQUENCE OF
6758 and break the signature.
6760 [This change does not apply to 0.9.7.]
6762 *) Zero the premaster secret after deriving the master secret in
6766 *) Add some EVP_add_digest_alias registrations (as found in
6767 OpenSSL_add_all_digests()) to SSL_library_init()
6768 aka OpenSSL_add_ssl_algorithms(). This provides improved
6769 compatibility with peers using X.509 certificates
6770 with unconventional AlgorithmIdentifier OIDs.
6773 *) Fix for Irix with NO_ASM.
6774 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6776 *) ./config script fixes.
6777 [Ulf Moeller, Richard Levitte]
6779 *) Fix 'openssl passwd -1'.
6782 *) Change PKCS12_key_gen_asc() so it can cope with non null
6783 terminated strings whose length is passed in the passlen
6784 parameter, for example from PEM callbacks. This was done
6785 by adding an extra length parameter to asc2uni().
6786 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6788 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6789 call failed, free the DSA structure.
6792 *) Fix to uni2asc() to cope with zero length Unicode strings.
6793 These are present in some PKCS#12 files.
6796 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6797 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6798 when writing a 32767 byte record.
6799 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6801 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6802 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6804 (RSA objects have a reference count access to which is protected
6805 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6806 so they are meant to be shared between threads.)
6807 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6808 "Reddie, Steven" <Steven.Reddie@ca.com>]
6810 *) Fix a deadlock in CRYPTO_mem_leaks().
6813 *) Use better test patterns in bntest.
6816 *) rand_win.c fix for Borland C.
6819 *) BN_rshift bugfix for n == 0.
6822 *) Add a 'bctest' script that checks for some known 'bc' bugs
6823 so that 'make test' does not abort just because 'bc' is broken.
6826 *) Store verify_result within SSL_SESSION also for client side to
6827 avoid potential security hole. (Re-used sessions on the client side
6828 always resulted in verify_result==X509_V_OK, not using the original
6829 result of the server certificate verification.)
6832 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6833 SSL3_RT_APPLICATION_DATA, return 0.
6834 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6838 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6839 releases, have been re-implemented by renaming the previous
6840 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6841 and ssl3_read_internal, respectively, and adding 'peek' parameters
6842 to them. The new ssl[23]_{read,peek} functions are calls to
6843 ssl[23]_read_internal with the 'peek' flag set appropriately.
6844 A 'peek' parameter has also been added to ssl3_read_bytes, which
6845 does the actual work for ssl3_read_internal.
6848 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6849 the method-specific "init()" handler. Also clean up ex_data after
6850 calling the method-specific "finish()" handler. Previously, this was
6851 happening the other way round.
6854 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6855 The previous value, 12, was not always sufficient for BN_mod_exp().
6858 *) Make sure that shared libraries get the internal name engine with
6859 the full version number and not just 0. This should mark the
6860 shared libraries as not backward compatible. Of course, this should
6861 be changed again when we can guarantee backward binary compatibility.
6864 *) Fix typo in get_cert_by_subject() in by_dir.c
6865 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6867 *) Rework the system to generate shared libraries:
6869 - Make note of the expected extension for the shared libraries and
6870 if there is a need for symbolic links from for example libcrypto.so.0
6871 to libcrypto.so.0.9.7. There is extended info in Configure for
6874 - Make as few rebuilds of the shared libraries as possible.
6876 - Still avoid linking the OpenSSL programs with the shared libraries.
6878 - When installing, install the shared libraries separately from the
6882 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6884 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6885 and not in SSL_clear because the latter is also used by the
6886 accept/connect functions; previously, the settings made by
6887 SSL_set_read_ahead would be lost during the handshake.
6888 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6890 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6891 Previously, it would create entries for disableed algorithms no
6895 *) Added several new manual pages for SSL_* function.
6898 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6900 *) In ssl23_get_client_hello, generate an error message when faced
6901 with an initial SSL 3.0/TLS record that is too small to contain the
6902 first two bytes of the ClientHello message, i.e. client_version.
6903 (Note that this is a pathologic case that probably has never happened
6904 in real life.) The previous approach was to use the version number
6905 from the record header as a substitute; but our protocol choice
6906 should not depend on that one because it is not authenticated
6907 by the Finished messages.
6910 *) More robust randomness gathering functions for Windows.
6911 [Jeffrey Altman <jaltman@columbia.edu>]
6913 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6914 not set then we don't setup the error code for issuer check errors
6915 to avoid possibly overwriting other errors which the callback does
6916 handle. If an application does set the flag then we assume it knows
6917 what it is doing and can handle the new informational codes
6921 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6922 a general "ANY" type, as such it should be able to decode anything
6923 including tagged types. However it didn't check the class so it would
6924 wrongly interpret tagged types in the same way as their universal
6925 counterpart and unknown types were just rejected. Changed so that the
6926 tagged and unknown types are handled in the same way as a SEQUENCE:
6927 that is the encoding is stored intact. There is also a new type
6928 "V_ASN1_OTHER" which is used when the class is not universal, in this
6929 case we have no idea what the actual type is so we just lump them all
6933 *) On VMS, stdout may very well lead to a file that is written to
6934 in a record-oriented fashion. That means that every write() will
6935 write a separate record, which will be read separately by the
6936 programs trying to read from it. This can be very confusing.
6938 The solution is to put a BIO filter in the way that will buffer
6939 text until a linefeed is reached, and then write everything a
6940 line at a time, so every record written will be an actual line,
6941 not chunks of lines and not (usually doesn't happen, but I've
6942 seen it once) several lines in one record. BIO_f_linebuffer() is
6945 Currently, it's a VMS-only method, because that's where it has
6946 been tested well enough.
6949 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6950 it can return incorrect results.
6951 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6952 but it was in 0.9.6-beta[12].)
6955 *) Disable the check for content being present when verifying detached
6956 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6957 include zero length content when signing messages.
6960 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6961 BIO_ctrl (for BIO pairs).
6964 *) Add DSO method for VMS.
6967 *) Bug fix: Montgomery multiplication could produce results with the
6971 *) Add RPM specification openssl.spec and modify it to build three
6972 packages. The default package contains applications, application
6973 documentation and run-time libraries. The devel package contains
6974 include files, static libraries and function documentation. The
6975 doc package contains the contents of the doc directory. The original
6976 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6979 *) Add a large number of documentation files for many SSL routines.
6980 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6982 *) Add a configuration entry for Sony News 4.
6983 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6985 *) Don't set the two most significant bits to one when generating a
6986 random number < q in the DSA library.
6989 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6990 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6991 the underlying transport is blocking) if a handshake took place.
6992 (The default behaviour is needed by applications such as s_client
6993 and s_server that use select() to determine when to use SSL_read;
6994 but for applications that know in advance when to expect data, it
6995 just makes things more complicated.)
6998 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
7002 *) Add a few more EBCDIC conditionals that make `req' and `x509'
7003 work better on such systems.
7004 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7006 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
7007 Update PKCS12_parse() so it copies the friendlyName and the
7008 keyid to the certificates aux info.
7011 *) Fix bug in PKCS7_verify() which caused an infinite loop
7012 if there was more than one signature.
7013 [Sven Uszpelkat <su@celocom.de>]
7015 *) Major change in util/mkdef.pl to include extra information
7016 about each symbol, as well as presentig variables as well
7017 as functions. This change means that there's n more need
7018 to rebuild the .num files when some algorithms are excluded.
7021 *) Allow the verify time to be set by an application,
7022 rather than always using the current time.
7025 *) Phase 2 verify code reorganisation. The certificate
7026 verify code now looks up an issuer certificate by a
7027 number of criteria: subject name, authority key id
7028 and key usage. It also verifies self signed certificates
7029 by the same criteria. The main comparison function is
7030 X509_check_issued() which performs these checks.
7032 Lot of changes were necessary in order to support this
7033 without completely rewriting the lookup code.
7035 Authority and subject key identifier are now cached.
7037 The LHASH 'certs' is X509_STORE has now been replaced
7038 by a STACK_OF(X509_OBJECT). This is mainly because an
7039 LHASH can't store or retrieve multiple objects with
7040 the same hash value.
7042 As a result various functions (which were all internal
7043 use only) have changed to handle the new X509_STORE
7044 structure. This will break anything that messed round
7045 with X509_STORE internally.
7047 The functions X509_STORE_add_cert() now checks for an
7048 exact match, rather than just subject name.
7050 The X509_STORE API doesn't directly support the retrieval
7051 of multiple certificates matching a given criteria, however
7052 this can be worked round by performing a lookup first
7053 (which will fill the cache with candidate certificates)
7054 and then examining the cache for matches. This is probably
7055 the best we can do without throwing out X509_LOOKUP
7056 entirely (maybe later...).
7058 The X509_VERIFY_CTX structure has been enhanced considerably.
7060 All certificate lookup operations now go via a get_issuer()
7061 callback. Although this currently uses an X509_STORE it
7062 can be replaced by custom lookups. This is a simple way
7063 to bypass the X509_STORE hackery necessary to make this
7064 work and makes it possible to use more efficient techniques
7065 in future. A very simple version which uses a simple
7066 STACK for its trusted certificate store is also provided
7067 using X509_STORE_CTX_trusted_stack().
7069 The verify_cb() and verify() callbacks now have equivalents
7070 in the X509_STORE_CTX structure.
7072 X509_STORE_CTX also has a 'flags' field which can be used
7073 to customise the verify behaviour.
7076 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
7077 excludes S/MIME capabilities.
7080 *) When a certificate request is read in keep a copy of the
7081 original encoding of the signed data and use it when outputing
7082 again. Signatures then use the original encoding rather than
7083 a decoded, encoded version which may cause problems if the
7084 request is improperly encoded.
7087 *) For consistency with other BIO_puts implementations, call
7088 buffer_write(b, ...) directly in buffer_puts instead of calling
7091 In BIO_puts, increment b->num_write as in BIO_write.
7092 [Peter.Sylvester@EdelWeb.fr]
7094 *) Fix BN_mul_word for the case where the word is 0. (We have to use
7095 BN_zero, we may not return a BIGNUM with an array consisting of
7099 *) Avoid calling abort() from within the library when problems are
7100 detected, except if preprocessor symbols have been defined
7101 (such as REF_CHECK, BN_DEBUG etc.).
7104 *) New openssl application 'rsautl'. This utility can be
7105 used for low level RSA operations. DER public key
7106 BIO/fp routines also added.
7109 *) New Configure entry and patches for compiling on QNX 4.
7110 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
7112 *) A demo state-machine implementation was sponsored by
7113 Nuron (http://www.nuron.com/) and is now available in
7114 demos/state_machine.
7117 *) New options added to the 'dgst' utility for signature
7118 generation and verification.
7121 *) Unrecognized PKCS#7 content types are now handled via a
7122 catch all ASN1_TYPE structure. This allows unsupported
7123 types to be stored as a "blob" and an application can
7124 encode and decode it manually.
7127 *) Fix various signed/unsigned issues to make a_strex.c
7129 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7131 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7132 length if passed a buffer. ASN1_INTEGER_to_BN failed
7133 if passed a NULL BN and its argument was negative.
7134 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7136 *) Modification to PKCS#7 encoding routines to output definite
7137 length encoding. Since currently the whole structures are in
7138 memory there's not real point in using indefinite length
7139 constructed encoding. However if OpenSSL is compiled with
7140 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7143 *) Added BIO_vprintf() and BIO_vsnprintf().
7146 *) Added more prefixes to parse for in the the strings written
7147 through a logging bio, to cover all the levels that are available
7148 through syslog. The prefixes are now:
7150 PANIC, EMERG, EMR => LOG_EMERG
7151 ALERT, ALR => LOG_ALERT
7152 CRIT, CRI => LOG_CRIT
7153 ERROR, ERR => LOG_ERR
7154 WARNING, WARN, WAR => LOG_WARNING
7155 NOTICE, NOTE, NOT => LOG_NOTICE
7156 INFO, INF => LOG_INFO
7157 DEBUG, DBG => LOG_DEBUG
7159 and as before, if none of those prefixes are present at the
7160 beginning of the string, LOG_ERR is chosen.
7162 On Win32, the LOG_* levels are mapped according to this:
7164 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7165 LOG_WARNING => EVENTLOG_WARNING_TYPE
7166 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7170 *) Made it possible to reconfigure with just the configuration
7171 argument "reconf" or "reconfigure". The command line arguments
7172 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7173 and are retrieved from there when reconfiguring.
7177 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7179 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7182 *) The obj_dat.pl script was messing up the sorting of object
7183 names. The reason was that it compared the quoted version
7184 of strings as a result "OCSP" > "OCSP Signing" because
7185 " > SPACE. Changed script to store unquoted versions of
7186 names and add quotes on output. It was also omitting some
7187 names from the lookup table if they were given a default
7188 value (that is if SN is missing it is given the same
7189 value as LN and vice versa), these are now added on the
7190 grounds that if an object has a name we should be able to
7191 look it up. Finally added warning output when duplicate
7192 short or long names are found.
7195 *) Changes needed for Tandem NSK.
7196 [Scott Uroff <scott@xypro.com>]
7198 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7199 RSA_padding_check_SSLv23(), special padding was never detected
7200 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7201 version rollback attacks was not effective.
7203 In s23_clnt.c, don't use special rollback-attack detection padding
7204 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7205 client; similarly, in s23_srvr.c, don't do the rollback check if
7206 SSL 2.0 is the only protocol enabled in the server.
7209 *) Make it possible to get hexdumps of unprintable data with 'openssl
7210 asn1parse'. By implication, the functions ASN1_parse_dump() and
7211 BIO_dump_indent() are added.
7214 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7215 these print out strings and name structures based on various
7216 flags including RFC2253 support and proper handling of
7217 multibyte characters. Added options to the 'x509' utility
7218 to allow the various flags to be set.
7221 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7222 Also change the functions X509_cmp_current_time() and
7223 X509_gmtime_adj() work with an ASN1_TIME structure,
7224 this will enable certificates using GeneralizedTime in validity
7225 dates to be checked.
7228 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7229 negative public key encodings) on by default,
7230 NO_NEG_PUBKEY_BUG can be set to disable it.
7233 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7234 content octets. An i2c_ASN1_OBJECT is unnecessary because
7235 the encoding can be trivially obtained from the structure.
7238 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7239 not read locks (CRYPTO_r_[un]lock).
7242 *) A first attempt at creating official support for shared
7243 libraries through configuration. I've kept it so the
7244 default is static libraries only, and the OpenSSL programs
7245 are always statically linked for now, but there are
7246 preparations for dynamic linking in place.
7247 This has been tested on Linux and Tru64.
7250 *) Randomness polling function for Win9x, as described in:
7251 Peter Gutmann, Software Generation of Practically Strong
7255 *) Fix so PRNG is seeded in req if using an already existing
7259 *) New options to smime application. -inform and -outform
7260 allow alternative formats for the S/MIME message including
7261 PEM and DER. The -content option allows the content to be
7262 specified separately. This should allow things like Netscape
7263 form signing output easier to verify.
7266 *) Fix the ASN1 encoding of tags using the 'long form'.
7269 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
7270 STRING types. These convert content octets to and from the
7271 underlying type. The actual tag and length octets are
7272 already assumed to have been read in and checked. These
7273 are needed because all other string types have virtually
7274 identical handling apart from the tag. By having versions
7275 of the ASN1 functions that just operate on content octets
7276 IMPLICIT tagging can be handled properly. It also allows
7277 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
7278 and ASN1_INTEGER are identical apart from the tag.
7281 *) Change the handling of OID objects as follows:
7283 - New object identifiers are inserted in objects.txt, following
7284 the syntax given in objects.README.
7285 - objects.pl is used to process obj_mac.num and create a new
7287 - obj_dat.pl is used to create a new obj_dat.h, using the data in
7290 This is currently kind of a hack, and the perl code in objects.pl
7291 isn't very elegant, but it works as I intended. The simplest way
7292 to check that it worked correctly is to look in obj_dat.h and
7293 check the array nid_objs and make sure the objects haven't moved
7294 around (this is important!). Additions are OK, as well as
7295 consistent name changes.
7298 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
7301 *) Addition of the command line parameter '-rand file' to 'openssl req'.
7302 The given file adds to whatever has already been seeded into the
7303 random pool through the RANDFILE configuration file option or
7304 environment variable, or the default random state file.
7307 *) mkstack.pl now sorts each macro group into lexical order.
7308 Previously the output order depended on the order the files
7309 appeared in the directory, resulting in needless rewriting
7313 *) Patches to make OpenSSL compile under Win32 again. Mostly
7314 work arounds for the VC++ problem that it treats func() as
7315 func(void). Also stripped out the parts of mkdef.pl that
7316 added extra typesafe functions: these no longer exist.
7319 *) Reorganisation of the stack code. The macros are now all
7320 collected in safestack.h . Each macro is defined in terms of
7321 a "stack macro" of the form SKM_<name>(type, a, b). The
7322 DEBUG_SAFESTACK is now handled in terms of function casts,
7323 this has the advantage of retaining type safety without the
7324 use of additional functions. If DEBUG_SAFESTACK is not defined
7325 then the non typesafe macros are used instead. Also modified the
7326 mkstack.pl script to handle the new form. Needs testing to see
7327 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
7328 the default if no major problems. Similar behaviour for ASN1_SET_OF
7329 and PKCS12_STACK_OF.
7332 *) When some versions of IIS use the 'NET' form of private key the
7333 key derivation algorithm is different. Normally MD5(password) is
7334 used as a 128 bit RC4 key. In the modified case
7335 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
7336 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
7337 as the old Netscape_RSA functions except they have an additional
7338 'sgckey' parameter which uses the modified algorithm. Also added
7339 an -sgckey command line option to the rsa utility. Thanks to
7340 Adrian Peck <bertie@ncipher.com> for posting details of the modified
7341 algorithm to openssl-dev.
7344 *) The evp_local.h macros were using 'c.##kname' which resulted in
7345 invalid expansion on some systems (SCO 5.0.5 for example).
7346 Corrected to 'c.kname'.
7347 [Phillip Porch <root@theporch.com>]
7349 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
7350 a STACK of email addresses from a certificate or request, these look
7351 in the subject name and the subject alternative name extensions and
7352 omit any duplicate addresses.
7355 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
7356 This makes DSA verification about 2 % faster.
7359 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
7360 (meaning that now 2^5 values will be precomputed, which is only 4 KB
7361 plus overhead for 1024 bit moduli).
7362 This makes exponentiations about 0.5 % faster for 1024 bit
7363 exponents (as measured by "openssl speed rsa2048").
7366 *) Rename memory handling macros to avoid conflicts with other
7368 Malloc => OPENSSL_malloc
7369 Malloc_locked => OPENSSL_malloc_locked
7370 Realloc => OPENSSL_realloc
7371 Free => OPENSSL_free
7374 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
7375 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
7378 *) CygWin32 support.
7379 [John Jarvie <jjarvie@newsguy.com>]
7381 *) The type-safe stack code has been rejigged. It is now only compiled
7382 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
7383 by default all type-specific stack functions are "#define"d back to
7384 standard stack functions. This results in more streamlined output
7385 but retains the type-safety checking possibilities of the original
7389 *) The STACK code has been cleaned up, and certain type declarations
7390 that didn't make a lot of sense have been brought in line. This has
7391 also involved a cleanup of sorts in safestack.h to more correctly
7392 map type-safe stack functions onto their plain stack counterparts.
7393 This work has also resulted in a variety of "const"ifications of
7394 lots of the code, especially "_cmp" operations which should normally
7395 be prototyped with "const" parameters anyway.
7398 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
7399 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
7400 (The PRNG state consists of two parts, the large pool 'state' and 'md',
7401 where all of 'md' is used each time the PRNG is used, but 'state'
7402 is used only indexed by a cyclic counter. As entropy may not be
7403 well distributed from the beginning, 'md' is important as a
7404 chaining variable. However, the output function chains only half
7405 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
7406 all of 'md', and seeding with STATE_SIZE dummy bytes will result
7407 in all of 'state' being rewritten, with the new values depending
7408 on virtually all of 'md'. This overcomes the 80 bit limitation.)
7411 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
7412 the handshake is continued after ssl_verify_cert_chain();
7413 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
7414 can lead to 'unexplainable' connection aborts later.
7415 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
7417 *) Major EVP API cipher revision.
7418 Add hooks for extra EVP features. This allows various cipher
7419 parameters to be set in the EVP interface. Support added for variable
7420 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
7421 setting of RC2 and RC5 parameters.
7423 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
7426 Remove lots of duplicated code from the EVP library. For example *every*
7427 cipher init() function handles the 'iv' in the same way according to the
7428 cipher mode. They also all do nothing if the 'key' parameter is NULL and
7429 for CFB and OFB modes they zero ctx->num.
7431 New functionality allows removal of S/MIME code RC2 hack.
7433 Most of the routines have the same form and so can be declared in terms
7436 By shifting this to the top level EVP_CipherInit() it can be removed from
7437 all individual ciphers. If the cipher wants to handle IVs or keys
7438 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7441 Change lots of functions like EVP_EncryptUpdate() to now return a
7442 value: although software versions of the algorithms cannot fail
7443 any installed hardware versions can.
7446 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7447 this option is set, tolerate broken clients that send the negotiated
7448 protocol version number instead of the requested protocol version
7452 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7453 i.e. non-zero for export ciphersuites, zero otherwise.
7454 Previous versions had this flag inverted, inconsistent with
7455 rsa_tmp_cb (..._TMP_RSA_CB).
7456 [Bodo Moeller; problem reported by Amit Chopra]
7458 *) Add missing DSA library text string. Work around for some IIS
7459 key files with invalid SEQUENCE encoding.
7462 *) Add a document (doc/standards.txt) that list all kinds of standards
7463 and so on that are implemented in OpenSSL.
7466 *) Enhance c_rehash script. Old version would mishandle certificates
7467 with the same subject name hash and wouldn't handle CRLs at all.
7468 Added -fingerprint option to crl utility, to support new c_rehash
7472 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7475 *) Fix for SSL server purpose checking. Server checking was
7476 rejecting certificates which had extended key usage present
7477 but no ssl client purpose.
7478 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7480 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7481 is a little unclear about how a blank password is handled.
7482 Since the password in encoded as a BMPString with terminating
7483 double NULL a zero length password would end up as just the
7484 double NULL. However no password at all is different and is
7485 handled differently in the PKCS#12 key generation code. NS
7486 treats a blank password as zero length. MSIE treats it as no
7487 password on export: but it will try both on import. We now do
7488 the same: PKCS12_parse() tries zero length and no password if
7489 the password is set to "" or NULL (NULL is now a valid password:
7490 it wasn't before) as does the pkcs12 application.
7493 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7494 perror when PEM_read_bio_X509_REQ fails, the error message must
7495 be obtained from the error queue.
7498 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7499 it in ERR_remove_state if appropriate, and change ERR_get_state
7500 accordingly to avoid race conditions (this is necessary because
7501 thread_hash is no longer constant once set).
7504 *) Bugfix for linux-elf makefile.one.
7507 *) RSA_get_default_method() will now cause a default
7508 RSA_METHOD to be chosen if one doesn't exist already.
7509 Previously this was only set during a call to RSA_new()
7510 or RSA_new_method(NULL) meaning it was possible for
7511 RSA_get_default_method() to return NULL.
7514 *) Added native name translation to the existing DSO code
7515 that will convert (if the flag to do so is set) filenames
7516 that are sufficiently small and have no path information
7517 into a canonical native form. Eg. "blah" converted to
7518 "libblah.so" or "blah.dll" etc.
7521 *) New function ERR_error_string_n(e, buf, len) which is like
7522 ERR_error_string(e, buf), but writes at most 'len' bytes
7523 including the 0 terminator. For ERR_error_string_n, 'buf'
7525 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7527 *) CONF library reworked to become more general. A new CONF
7528 configuration file reader "class" is implemented as well as a
7529 new functions (NCONF_*, for "New CONF") to handle it. The now
7530 old CONF_* functions are still there, but are reimplemented to
7531 work in terms of the new functions. Also, a set of functions
7532 to handle the internal storage of the configuration data is
7533 provided to make it easier to write new configuration file
7534 reader "classes" (I can definitely see something reading a
7535 configuration file in XML format, for example), called _CONF_*,
7536 or "the configuration storage API"...
7538 The new configuration file reading functions are:
7540 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7541 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7543 NCONF_default, NCONF_WIN32
7545 NCONF_dump_fp, NCONF_dump_bio
7547 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7548 NCONF_new creates a new CONF object. This works in the same way
7549 as other interfaces in OpenSSL, like the BIO interface.
7550 NCONF_dump_* dump the internal storage of the configuration file,
7551 which is useful for debugging. All other functions take the same
7552 arguments as the old CONF_* functions wth the exception of the
7553 first that must be a `CONF *' instead of a `LHASH *'.
7555 To make it easer to use the new classes with the old CONF_* functions,
7556 the function CONF_set_default_method is provided.
7559 *) Add '-tls1' option to 'openssl ciphers', which was already
7560 mentioned in the documentation but had not been implemented.
7561 (This option is not yet really useful because even the additional
7562 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7565 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7566 OpenSSL-based applications) load shared libraries and bind to
7567 them in a portable way.
7568 [Geoff Thorpe, with contributions from Richard Levitte]
7570 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7572 *) Make sure _lrotl and _lrotr are only used with MSVC.
7574 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7575 (the default implementation of RAND_status).
7577 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7578 to '-clrext' (= clear extensions), as intended and documented.
7579 [Bodo Moeller; inconsistency pointed out by Michael Attili
7582 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7583 was larger than the MD block size.
7584 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7586 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7587 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7588 using the passed key: if the passed key was a private key the result
7589 of X509_print(), for example, would be to print out all the private key
7593 *) des_quad_cksum() byte order bug fix.
7594 [Ulf Möller, using the problem description in krb4-0.9.7, where
7595 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7597 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7599 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7601 *) For easily testing in shell scripts whether some command
7602 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7603 returns with exit code 0 iff no command of the given name is available.
7604 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7605 the output goes to stdout and nothing is printed to stderr.
7606 Additional arguments are always ignored.
7608 Since for each cipher there is a command of the same name,
7609 the 'no-cipher' compilation switches can be tested this way.
7611 ('openssl no-XXX' is not able to detect pseudo-commands such
7612 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7615 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7618 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7619 is set; it will be thrown away anyway because each handshake creates
7621 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7622 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7623 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7624 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7627 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7628 'Q' and 'R' lose their special meanings (quit/renegotiate).
7629 This is part of what -quiet does; unlike -quiet, -ign_eof
7630 does not suppress any output.
7633 *) Add compatibility options to the purpose and trust code. The
7634 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7635 accepts a certificate or CA, this was the previous behaviour,
7636 with all the associated security issues.
7638 X509_TRUST_COMPAT is the old trust behaviour: only and
7639 automatically trust self signed roots in certificate store. A
7640 new trust setting X509_TRUST_DEFAULT is used to specify that
7641 a purpose has no associated trust setting and it should instead
7642 use the value in the default purpose.
7645 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7646 and fix a memory leak.
7649 *) In util/mkerr.pl (which implements 'make errors'), preserve
7650 reason strings from the previous version of the .c file, as
7651 the default to have only downcase letters (and digits) in
7652 automatically generated reasons codes is not always appropriate.
7655 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7656 using strerror. Previously, ERR_reason_error_string() returned
7657 library names as reason strings for SYSerr; but SYSerr is a special
7658 case where small numbers are errno values, not library numbers.
7661 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7662 converts DSA parameters into DH parameters. (When creating parameters,
7663 DSA_generate_parameters is used.)
7666 *) Include 'length' (recommended exponent length) in C code generated
7667 by 'openssl dhparam -C'.
7670 *) The second argument to set_label in perlasm was already being used
7671 so couldn't be used as a "file scope" flag. Moved to third argument
7675 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7676 instead of RAND_bytes for encryption IVs and salts.
7679 *) Include RAND_status() into RAND_METHOD instead of implementing
7680 it only for md_rand.c Otherwise replacing the PRNG by calling
7681 RAND_set_rand_method would be impossible.
7684 *) Don't let DSA_generate_key() enter an infinite loop if the random
7685 number generation fails.
7688 *) New 'rand' application for creating pseudo-random output.
7691 *) Added configuration support for Linux/IA64
7692 [Rolf Haberrecker <rolf@suse.de>]
7694 *) Assembler module support for Mingw32.
7697 *) Shared library support for HPUX (in shlib/).
7698 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7700 *) Shared library support for Solaris gcc.
7701 [Lutz Behnke <behnke@trustcenter.de>]
7703 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7705 *) PKCS7_encrypt() was adding text MIME headers twice because they
7706 were added manually and by SMIME_crlf_copy().
7709 *) In bntest.c don't call BN_rand with zero bits argument.
7710 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7712 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7713 case was implemented. This caused BN_div_recp() to fail occasionally.
7716 *) Add an optional second argument to the set_label() in the perl
7717 assembly language builder. If this argument exists and is set
7718 to 1 it signals that the assembler should use a symbol whose
7719 scope is the entire file, not just the current function. This
7720 is needed with MASM which uses the format label:: for this scope.
7721 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7723 *) Change the ASN1 types so they are typedefs by default. Before
7724 almost all types were #define'd to ASN1_STRING which was causing
7725 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7729 *) Change names of new functions to the new get1/get0 naming
7730 convention: After 'get1', the caller owns a reference count
7731 and has to call ..._free; 'get0' returns a pointer to some
7732 data structure without incrementing reference counters.
7733 (Some of the existing 'get' functions increment a reference
7734 counter, some don't.)
7735 Similarly, 'set1' and 'add1' functions increase reference
7736 counters or duplicate objects.
7739 *) Allow for the possibility of temp RSA key generation failure:
7740 the code used to assume it always worked and crashed on failure.
7743 *) Fix potential buffer overrun problem in BIO_printf().
7744 [Ulf Möller, using public domain code by Patrick Powell; problem
7745 pointed out by David Sacerdote <das33@cornell.edu>]
7747 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7748 RAND_egd() and RAND_status(). In the command line application,
7749 the EGD socket can be specified like a seed file using RANDFILE
7753 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7754 Some CAs (e.g. Verisign) distribute certificates in this form.
7757 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7758 list to exclude them. This means that no special compilation option
7759 is needed to use anonymous DH: it just needs to be included in the
7763 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7764 EVP_MD_type. The old functionality is available in a new macro called
7765 EVP_MD_md(). Change code that uses it and update docs.
7768 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7769 where the 'void *' argument is replaced by a function pointer argument.
7770 Previously 'void *' was abused to point to functions, which works on
7771 many platforms, but is not correct. As these functions are usually
7772 called by macros defined in OpenSSL header files, most source code
7773 should work without changes.
7776 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7777 sections with information on -D... compiler switches used for
7778 compiling the library so that applications can see them. To enable
7779 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7780 must be defined. E.g.,
7781 #define OPENSSL_ALGORITHM_DEFINES
7782 #include <openssl/opensslconf.h>
7783 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7784 [Richard Levitte, Ulf and Bodo Möller]
7786 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7790 *) Change the 'other' type in certificate aux info to a STACK_OF
7791 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7792 the required ASN1 format: arbitrary types determined by an OID.
7795 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7796 argument to 'req'. This is not because the function is newer or
7797 better than others it just uses the work 'NEW' in the certificate
7798 request header lines. Some software needs this.
7801 *) Reorganise password command line arguments: now passwords can be
7802 obtained from various sources. Delete the PEM_cb function and make
7803 it the default behaviour: i.e. if the callback is NULL and the
7804 usrdata argument is not NULL interpret it as a null terminated pass
7805 phrase. If usrdata and the callback are NULL then the pass phrase
7806 is prompted for as usual.
7809 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7810 the support is automatically enabled. The resulting binaries will
7811 autodetect the card and use it if present.
7812 [Ben Laurie and Compaq Inc.]
7814 *) Work around for Netscape hang bug. This sends certificate request
7815 and server done in one record. Since this is perfectly legal in the
7816 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7817 the bugs/SSLv3 entry for more info.
7820 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7823 *) Add -rand argument to smime and pkcs12 applications and read/write
7827 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7830 *) Add command line password options to the remaining applications.
7833 *) Bug fix for BN_div_recp() for numerators with an even number of
7837 *) More tests in bntest.c, and changed test_bn output.
7840 *) ./config recognizes MacOS X now.
7843 *) Bug fix for BN_div() when the first words of num and divsor are
7844 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7847 *) Add support for various broken PKCS#8 formats, and command line
7848 options to produce them.
7851 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7852 get temporary BIGNUMs from a BN_CTX.
7855 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7859 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7860 include a #define from the old name to the new. The original intent
7861 was that statically linked binaries could for example just call
7862 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7863 link with digests. This never worked becayse SSLeay_add_all_digests()
7864 and SSLeay_add_all_ciphers() were in the same source file so calling
7865 one would link with the other. They are now in separate source files.
7868 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7871 *) Use a less unusual form of the Miller-Rabin primality test (it used
7872 a binary algorithm for exponentiation integrated into the Miller-Rabin
7873 loop, our standard modexp algorithms are faster).
7876 *) Support for the EBCDIC character set completed.
7877 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7879 *) Source code cleanups: use const where appropriate, eliminate casts,
7880 use void * instead of char * in lhash.
7883 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7884 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7885 this the server could overwrite ephemeral keys that the client
7889 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7890 using 50 iterations of the Rabin-Miller test.
7892 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7893 iterations of the Rabin-Miller test as required by the appendix
7894 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7895 As BN_is_prime_fasttest includes trial division, DSA parameter
7896 generation becomes much faster.
7898 This implies a change for the callback functions in DSA_is_prime
7899 and DSA_generate_parameters: The callback function is called once
7900 for each positive witness in the Rabin-Miller test, not just
7901 occasionally in the inner loop; and the parameters to the
7902 callback function now provide an iteration count for the outer
7903 loop rather than for the current invocation of the inner loop.
7904 DSA_generate_parameters additionally can call the callback
7905 function with an 'iteration count' of -1, meaning that a
7906 candidate has passed the trial division test (when q is generated
7907 from an application-provided seed, trial division is skipped).
7910 *) New function BN_is_prime_fasttest that optionally does trial
7911 division before starting the Rabin-Miller test and has
7912 an additional BN_CTX * argument (whereas BN_is_prime always
7913 has to allocate at least one BN_CTX).
7914 'callback(1, -1, cb_arg)' is called when a number has passed the
7915 trial division stage.
7918 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7922 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7925 *) New function BN_pseudo_rand().
7928 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7929 bignum version of BN_from_montgomery() with the working code from
7930 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7934 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7935 made it impossible to use the same SSL_SESSION data structure in
7936 SSL2 clients in multiple threads.
7939 *) The return value of RAND_load_file() no longer counts bytes obtained
7940 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7941 to seed the PRNG (previously an explicit byte count was required).
7942 [Ulf Möller, Bodo Möller]
7944 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7945 used (char *) instead of (void *) and had casts all over the place.
7948 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7951 *) Retain source code compatibility for BN_prime_checks macro:
7952 BN_is_prime(..., BN_prime_checks, ...) now uses
7953 BN_prime_checks_for_size to determine the appropriate number of
7954 Rabin-Miller iterations.
7957 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7958 DH_CHECK_P_NOT_SAFE_PRIME.
7959 (Check if this is true? OpenPGP calls them "strong".)
7962 *) Merge the functionality of "dh" and "gendh" programs into a new program
7963 "dhparam". The old programs are retained for now but will handle DH keys
7964 (instead of parameters) in future.
7967 *) Make the ciphers, s_server and s_client programs check the return values
7968 when a new cipher list is set.
7971 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7972 ciphers. Before when the 56bit ciphers were enabled the sorting was
7975 The syntax for the cipher sorting has been extended to support sorting by
7976 cipher-strength (using the strength_bits hard coded in the tables).
7977 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7979 Fix a bug in the cipher-command parser: when supplying a cipher command
7980 string with an "undefined" symbol (neither command nor alphanumeric
7981 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7982 an error is flagged.
7984 Due to the strength-sorting extension, the code of the
7985 ssl_create_cipher_list() function was completely rearranged. I hope that
7986 the readability was also increased :-)
7987 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7989 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7990 for the first serial number and places 2 in the serial number file. This
7991 avoids problems when the root CA is created with serial number zero and
7992 the first user certificate has the same issuer name and serial number
7996 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7997 the new code. Add documentation for this stuff.
8000 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
8001 X509_*() to X509at_*() on the grounds that they don't handle X509
8002 structures and behave in an analagous way to the X509v3 functions:
8003 they shouldn't be called directly but wrapper functions should be used
8006 So we also now have some wrapper functions that call the X509at functions
8007 when passed certificate requests. (TO DO: similar things can be done with
8008 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
8009 things. Some of these need some d2i or i2d and print functionality
8010 because they handle more complex structures.)
8013 *) Add missing #ifndefs that caused missing symbols when building libssl
8014 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
8015 NO_RSA in ssl/s2*.c.
8016 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
8018 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
8019 has a return value which indicates the quality of the random data
8020 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
8021 error queue. New function RAND_pseudo_bytes() generates output that is
8022 guaranteed to be unique but not unpredictable. RAND_add is like
8023 RAND_seed, but takes an extra argument for an entropy estimate
8024 (RAND_seed always assumes full entropy).
8027 *) Do more iterations of Rabin-Miller probable prime test (specifically,
8028 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
8029 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
8030 in crypto/bn/bn_prime.c for the complete table). This guarantees a
8031 false-positive rate of at most 2^-80 for random input.
8034 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
8037 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
8038 in the 0.9.5 release), this returns the chain
8039 from an X509_CTX structure with a dup of the stack and all
8040 the X509 reference counts upped: so the stack will exist
8041 after X509_CTX_cleanup() has been called. Modify pkcs12.c
8044 Also make SSL_SESSION_print() print out the verify return
8048 *) Add manpage for the pkcs12 command. Also change the default
8049 behaviour so MAC iteration counts are used unless the new
8050 -nomaciter option is used. This improves file security and
8051 only older versions of MSIE (4.0 for example) need it.
8054 *) Honor the no-xxx Configure options when creating .DEF files.
8057 *) Add PKCS#10 attributes to field table: challengePassword,
8058 unstructuredName and unstructuredAddress. These are taken from
8059 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
8060 international characters are used.
8062 More changes to X509_ATTRIBUTE code: allow the setting of types
8063 based on strings. Remove the 'loc' parameter when adding
8064 attributes because these will be a SET OF encoding which is sorted
8068 *) Initial changes to the 'req' utility to allow request generation
8069 automation. This will allow an application to just generate a template
8070 file containing all the field values and have req construct the
8073 Initial support for X509_ATTRIBUTE handling. Stacks of these are
8074 used all over the place including certificate requests and PKCS#7
8075 structures. They are currently handled manually where necessary with
8076 some primitive wrappers for PKCS#7. The new functions behave in a
8077 manner analogous to the X509 extension functions: they allow
8078 attributes to be looked up by NID and added.
8080 Later something similar to the X509V3 code would be desirable to
8081 automatically handle the encoding, decoding and printing of the
8082 more complex types. The string types like challengePassword can
8083 be handled by the string table functions.
8085 Also modified the multi byte string table handling. Now there is
8086 a 'global mask' which masks out certain types. The table itself
8087 can use the flag STABLE_NO_MASK to ignore the mask setting: this
8088 is useful when for example there is only one permissible type
8089 (as in countryName) and using the mask might result in no valid
8093 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
8094 SSL_get_peer_finished to allow applications to obtain the latest
8095 Finished messages sent to the peer or expected from the peer,
8096 respectively. (SSL_get_peer_finished is usually the Finished message
8097 actually received from the peer, otherwise the protocol will be aborted.)
8099 As the Finished message are message digests of the complete handshake
8100 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
8101 be used for external authentication procedures when the authentication
8102 provided by SSL/TLS is not desired or is not enough.
8105 *) Enhanced support for Alpha Linux is added. Now ./config checks if
8106 the host supports BWX extension and if Compaq C is present on the
8107 $PATH. Just exploiting of the BWX extension results in 20-30%
8108 performance kick for some algorithms, e.g. DES and RC4 to mention
8109 a couple. Compaq C in turn generates ~20% faster code for MD5 and
8113 *) Add support for MS "fast SGC". This is arguably a violation of the
8114 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
8115 weak crypto and after checking the certificate is SGC a second one
8116 with strong crypto. MS SGC stops the first handshake after receiving
8117 the server certificate message and sends a second client hello. Since
8118 a server will typically do all the time consuming operations before
8119 expecting any further messages from the client (server key exchange
8120 is the most expensive) there is little difference between the two.
8122 To get OpenSSL to support MS SGC we have to permit a second client
8123 hello message after we have sent server done. In addition we have to
8124 reset the MAC if we do get this second client hello.
8127 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
8128 if a DER encoded private key is RSA or DSA traditional format. Changed
8129 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
8130 format DER encoded private key. Newer code should use PKCS#8 format which
8131 has the key type encoded in the ASN1 structure. Added DER private key
8132 support to pkcs8 application.
8135 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8136 ciphersuites has been selected (as required by the SSL 3/TLS 1
8137 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8138 is set, we interpret this as a request to violate the specification
8139 (the worst that can happen is a handshake failure, and 'correct'
8140 behaviour would result in a handshake failure anyway).
8143 *) In SSL_CTX_add_session, take into account that there might be multiple
8144 SSL_SESSION structures with the same session ID (e.g. when two threads
8145 concurrently obtain them from an external cache).
8146 The internal cache can handle only one SSL_SESSION with a given ID,
8147 so if there's a conflict, we now throw out the old one to achieve
8151 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8152 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8153 some routines that use cipher OIDs: some ciphers do not have OIDs
8154 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8158 *) Simplify the trust setting structure and code. Now we just have
8159 two sequences of OIDs for trusted and rejected settings. These will
8160 typically have values the same as the extended key usage extension
8161 and any application specific purposes.
8163 The trust checking code now has a default behaviour: it will just
8164 check for an object with the same NID as the passed id. Functions can
8165 be provided to override either the default behaviour or the behaviour
8166 for a given id. SSL client, server and email already have functions
8167 in place for compatibility: they check the NID and also return "trusted"
8168 if the certificate is self signed.
8171 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8172 traditional format into an EVP_PKEY structure.
8175 *) Add a password callback function PEM_cb() which either prompts for
8176 a password if usr_data is NULL or otherwise assumes it is a null
8177 terminated password. Allow passwords to be passed on command line
8178 environment or config files in a few more utilities.
8181 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8182 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8183 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8184 Update documentation.
8187 *) Support for ASN1 "NULL" type. This could be handled before by using
8188 ASN1_TYPE but there wasn't any function that would try to read a NULL
8189 and produce an error if it couldn't. For compatibility we also have
8190 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8191 don't allocate anything because they don't need to.
8194 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8196 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8198 *) Rebuild of the memory allocation routines used by OpenSSL code and
8199 possibly others as well. The purpose is to make an interface that
8200 provide hooks so anyone can build a separate set of allocation and
8201 deallocation routines to be used by OpenSSL, for example memory
8202 pool implementations, or something else, which was previously hard
8203 since Malloc(), Realloc() and Free() were defined as macros having
8204 the values malloc, realloc and free, respectively (except for Win32
8205 compilations). The same is provided for memory debugging code.
8206 OpenSSL already comes with functionality to find memory leaks, but
8207 this gives people a chance to debug other memory problems.
8209 With these changes, a new set of functions and macros have appeared:
8211 CRYPTO_set_mem_debug_functions() [F]
8212 CRYPTO_get_mem_debug_functions() [F]
8213 CRYPTO_dbg_set_options() [F]
8214 CRYPTO_dbg_get_options() [F]
8215 CRYPTO_malloc_debug_init() [M]
8217 The memory debug functions are NULL by default, unless the library
8218 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8219 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8220 gives the standard debugging functions that come with OpenSSL) or
8221 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8222 provided by the library user) must be used. When the standard
8223 debugging functions are used, CRYPTO_dbg_set_options can be used to
8224 request additional information:
8225 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8226 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8228 Also, things like CRYPTO_set_mem_functions will always give the
8229 expected result (the new set of functions is used for allocation
8230 and deallocation) at all times, regardless of platform and compiler
8233 To finish it up, some functions that were never use in any other
8234 way than through macros have a new API and new semantic:
8237 CRYPTO_dbg_realloc()
8240 All macros of value have retained their old syntax.
8241 [Richard Levitte and Bodo Moeller]
8243 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8244 ordering of SMIMECapabilities wasn't in "strength order" and there
8245 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8249 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8250 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8251 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
8253 *) Merge in my S/MIME library for OpenSSL. This provides a simple
8254 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
8255 functionality to handle multipart/signed properly) and a utility
8256 called 'smime' to call all this stuff. This is based on code I
8257 originally wrote for Celo who have kindly allowed it to be
8258 included in OpenSSL.
8261 *) Add variants des_set_key_checked and des_set_key_unchecked of
8262 des_set_key (aka des_key_sched). Global variable des_check_key
8263 decides which of these is called by des_set_key; this way
8264 des_check_key behaves as it always did, but applications and
8265 the library itself, which was buggy for des_check_key == 1,
8266 have a cleaner way to pick the version they need.
8269 *) New function PKCS12_newpass() which changes the password of a
8273 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
8274 dynamic mix. In both cases the ids can be used as an index into the
8275 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
8276 functions so they accept a list of the field values and the
8277 application doesn't need to directly manipulate the X509_TRUST
8281 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
8285 *) Modify the way the V3 extension code looks up extensions. This now
8286 works in a similar way to the object code: we have some "standard"
8287 extensions in a static table which is searched with OBJ_bsearch()
8288 and the application can add dynamic ones if needed. The file
8289 crypto/x509v3/ext_dat.h now has the info: this file needs to be
8290 updated whenever a new extension is added to the core code and kept
8291 in ext_nid order. There is a simple program 'tabtest.c' which checks
8292 this. New extensions are not added too often so this file can readily
8293 be maintained manually.
8295 There are two big advantages in doing things this way. The extensions
8296 can be looked up immediately and no longer need to be "added" using
8297 X509V3_add_standard_extensions(): this function now does nothing.
8298 [Side note: I get *lots* of email saying the extension code doesn't
8299 work because people forget to call this function]
8300 Also no dynamic allocation is done unless new extensions are added:
8301 so if we don't add custom extensions there is no need to call
8302 X509V3_EXT_cleanup().
8305 *) Modify enc utility's salting as follows: make salting the default. Add a
8306 magic header, so unsalted files fail gracefully instead of just decrypting
8307 to garbage. This is because not salting is a big security hole, so people
8308 should be discouraged from doing it.
8311 *) Fixes and enhancements to the 'x509' utility. It allowed a message
8312 digest to be passed on the command line but it only used this
8313 parameter when signing a certificate. Modified so all relevant
8314 operations are affected by the digest parameter including the
8315 -fingerprint and -x509toreq options. Also -x509toreq choked if a
8316 DSA key was used because it didn't fix the digest.
8319 *) Initial certificate chain verify code. Currently tests the untrusted
8320 certificates for consistency with the verify purpose (which is set
8321 when the X509_STORE_CTX structure is set up) and checks the pathlength.
8323 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
8324 this is because it will reject chains with invalid extensions whereas
8325 every previous version of OpenSSL and SSLeay made no checks at all.
8327 Trust code: checks the root CA for the relevant trust settings. Trust
8328 settings have an initial value consistent with the verify purpose: e.g.
8329 if the verify purpose is for SSL client use it expects the CA to be
8330 trusted for SSL client use. However the default value can be changed to
8331 permit custom trust settings: one example of this would be to only trust
8332 certificates from a specific "secure" set of CAs.
8334 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
8335 which should be used for version portability: especially since the
8336 verify structure is likely to change more often now.
8338 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
8339 to set them. If not set then assume SSL clients will verify SSL servers
8342 Two new options to the verify program: -untrusted allows a set of
8343 untrusted certificates to be passed in and -purpose which sets the
8344 intended purpose of the certificate. If a purpose is set then the
8345 new chain verify code is used to check extension consistency.
8348 *) Support for the authority information access extension.
8351 *) Modify RSA and DSA PEM read routines to transparently handle
8352 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
8353 public keys in a format compatible with certificate
8354 SubjectPublicKeyInfo structures. Unfortunately there were already
8355 functions called *_PublicKey_* which used various odd formats so
8356 these are retained for compatibility: however the DSA variants were
8357 never in a public release so they have been deleted. Changed dsa/rsa
8358 utilities to handle the new format: note no releases ever handled public
8359 keys so we should be OK.
8361 The primary motivation for this change is to avoid the same fiasco
8362 that dogs private keys: there are several incompatible private key
8363 formats some of which are standard and some OpenSSL specific and
8364 require various evil hacks to allow partial transparent handling and
8365 even then it doesn't work with DER formats. Given the option anything
8366 other than PKCS#8 should be dumped: but the other formats have to
8367 stay in the name of compatibility.
8369 With public keys and the benefit of hindsight one standard format
8370 is used which works with EVP_PKEY, RSA or DSA structures: though
8371 it clearly returns an error if you try to read the wrong kind of key.
8373 Added a -pubkey option to the 'x509' utility to output the public key.
8374 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
8375 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
8376 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
8377 that do the same as the EVP_PKEY_assign_*() except they up the
8378 reference count of the added key (they don't "swallow" the
8382 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
8383 CRLs would fail if the file contained no certificates or no CRLs:
8384 added a new function to read in both types and return the number
8385 read: this means that if none are read it will be an error. The
8386 DER versions of the certificate and CRL reader would always fail
8387 because it isn't possible to mix certificates and CRLs in DER format
8388 without choking one or the other routine. Changed this to just read
8389 a certificate: this is the best we can do. Also modified the code
8390 in apps/verify.c to take notice of return codes: it was previously
8391 attempting to read in certificates from NULL pointers and ignoring
8392 any errors: this is one reason why the cert and CRL reader seemed
8393 to work. It doesn't check return codes from the default certificate
8394 routines: these may well fail if the certificates aren't installed.
8397 *) Code to support otherName option in GeneralName.
8400 *) First update to verify code. Change the verify utility
8401 so it warns if it is passed a self signed certificate:
8402 for consistency with the normal behaviour. X509_verify
8403 has been modified to it will now verify a self signed
8404 certificate if *exactly* the same certificate appears
8405 in the store: it was previously impossible to trust a
8406 single self signed certificate. This means that:
8407 openssl verify ss.pem
8408 now gives a warning about a self signed certificate but
8409 openssl verify -CAfile ss.pem ss.pem
8413 *) For servers, store verify_result in SSL_SESSION data structure
8414 (and add it to external session representation).
8415 This is needed when client certificate verifications fails,
8416 but an application-provided verification callback (set by
8417 SSL_CTX_set_cert_verify_callback) allows accepting the session
8418 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
8419 but returns 1): When the session is reused, we have to set
8420 ssl->verify_result to the appropriate error code to avoid
8422 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
8424 *) Fix a bug in the new PKCS#7 code: it didn't consider the
8425 case in PKCS7_dataInit() where the signed PKCS7 structure
8426 didn't contain any existing data because it was being created.
8427 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
8429 *) Add a salt to the key derivation routines in enc.c. This
8430 forms the first 8 bytes of the encrypted file. Also add a
8431 -S option to allow a salt to be input on the command line.
8434 *) New function X509_cmp(). Oddly enough there wasn't a function
8435 to compare two certificates. We do this by working out the SHA1
8436 hash and comparing that. X509_cmp() will be needed by the trust
8440 *) SSL_get1_session() is like SSL_get_session(), but increments
8441 the reference count in the SSL_SESSION returned.
8442 [Geoff Thorpe <geoff@eu.c2.net>]
8444 *) Fix for 'req': it was adding a null to request attributes.
8445 Also change the X509_LOOKUP and X509_INFO code to handle
8446 certificate auxiliary information.
8449 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8453 *) Add the possibility to add extra information to the memory leak
8454 detecting output, to form tracebacks, showing from where each
8455 allocation was originated: CRYPTO_push_info("constant string") adds
8456 the string plus current file name and line number to a per-thread
8457 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8458 is like calling CYRPTO_pop_info() until the stack is empty.
8459 Also updated memory leak detection code to be multi-thread-safe.
8462 *) Add options -text and -noout to pkcs7 utility and delete the
8463 encryption options which never did anything. Update docs.
8466 *) Add options to some of the utilities to allow the pass phrase
8467 to be included on either the command line (not recommended on
8468 OSes like Unix) or read from the environment. Update the
8469 manpages and fix a few bugs.
8472 *) Add a few manpages for some of the openssl commands.
8475 *) Fix the -revoke option in ca. It was freeing up memory twice,
8476 leaking and not finding already revoked certificates.
8479 *) Extensive changes to support certificate auxiliary information.
8480 This involves the use of X509_CERT_AUX structure and X509_AUX
8481 functions. An X509_AUX function such as PEM_read_X509_AUX()
8482 can still read in a certificate file in the usual way but it
8483 will also read in any additional "auxiliary information". By
8484 doing things this way a fair degree of compatibility can be
8485 retained: existing certificates can have this information added
8486 using the new 'x509' options.
8488 Current auxiliary information includes an "alias" and some trust
8489 settings. The trust settings will ultimately be used in enhanced
8490 certificate chain verification routines: currently a certificate
8491 can only be trusted if it is self signed and then it is trusted
8495 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8496 The problem was that one of the replacement routines had not been working
8497 since SSLeay releases. For now the offending routine has been replaced
8498 with non-optimised assembler. Even so, this now gives around 95%
8499 performance improvement for 1024 bit RSA signs.
8502 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8503 handling. Most clients have the effective key size in bits equal to
8504 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8505 A few however don't do this and instead use the size of the decrypted key
8506 to determine the RC2 key length and the AlgorithmIdentifier to determine
8507 the effective key length. In this case the effective key length can still
8508 be 40 bits but the key length can be 168 bits for example. This is fixed
8509 by manually forcing an RC2 key into the EVP_PKEY structure because the
8510 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8511 the key length and effective key length are equal.
8514 *) Add a bunch of functions that should simplify the creation of
8515 X509_NAME structures. Now you should be able to do:
8516 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8517 and have it automatically work out the correct field type and fill in
8518 the structures. The more adventurous can try:
8519 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8520 and it will (hopefully) work out the correct multibyte encoding.
8523 *) Change the 'req' utility to use the new field handling and multibyte
8524 copy routines. Before the DN field creation was handled in an ad hoc
8525 way in req, ca, and x509 which was rather broken and didn't support
8526 BMPStrings or UTF8Strings. Since some software doesn't implement
8527 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8528 using the dirstring_type option. See the new comment in the default
8529 openssl.cnf for more info.
8532 *) Make crypto/rand/md_rand.c more robust:
8533 - Assure unique random numbers after fork().
8534 - Make sure that concurrent threads access the global counter and
8535 md serializably so that we never lose entropy in them
8536 or use exactly the same state in multiple threads.
8537 Access to the large state is not always serializable because
8538 the additional locking could be a performance killer, and
8539 md should be large enough anyway.
8542 *) New file apps/app_rand.c with commonly needed functionality
8543 for handling the random seed file.
8545 Use the random seed file in some applications that previously did not:
8547 dsaparam -genkey (which also ignored its '-rand' option),
8550 x509 (when signing).
8551 Except on systems with /dev/urandom, it is crucial to have a random
8552 seed file at least for key creation, DSA signing, and for DH exchanges;
8553 for RSA signatures we could do without one.
8555 gendh and gendsa (unlike genrsa) used to read only the first byte
8556 of each file listed in the '-rand' option. The function as previously
8557 found in genrsa is now in app_rand.c and is used by all programs
8558 that support '-rand'.
8561 *) In RAND_write_file, use mode 0600 for creating files;
8562 don't just chmod when it may be too late.
8565 *) Report an error from X509_STORE_load_locations
8566 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8569 *) New function ASN1_mbstring_copy() this copies a string in either
8570 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8571 into an ASN1_STRING type. A mask of permissible types is passed
8572 and it chooses the "minimal" type to use or an error if not type
8576 *) Add function equivalents to the various macros in asn1.h. The old
8577 macros are retained with an M_ prefix. Code inside the library can
8578 use the M_ macros. External code (including the openssl utility)
8579 should *NOT* in order to be "shared library friendly".
8582 *) Add various functions that can check a certificate's extensions
8583 to see if it usable for various purposes such as SSL client,
8584 server or S/MIME and CAs of these types. This is currently
8585 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8586 verification. Also added a -purpose flag to x509 utility to
8587 print out all the purposes.
8590 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8594 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8595 for, obtain and decode and extension and obtain its critical flag.
8596 This allows all the necessary extension code to be handled in a
8597 single function call.
8600 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8601 platforms. See crypto/rc4/rc4_enc.c for further details.
8604 *) New -noout option to asn1parse. This causes no output to be produced
8605 its main use is when combined with -strparse and -out to extract data
8606 from a file (which may not be in ASN.1 format).
8609 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8610 when producing the local key id.
8611 [Richard Levitte <levitte@stacken.kth.se>]
8613 *) New option -dhparam in s_server. This allows a DH parameter file to be
8614 stated explicitly. If it is not stated then it tries the first server
8615 certificate file. The previous behaviour hard coded the filename
8619 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8620 a public key to be input or output. For example:
8621 openssl rsa -in key.pem -pubout -out pubkey.pem
8622 Also added necessary DSA public key functions to handle this.
8625 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8626 in the message. This was handled by allowing
8627 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8628 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8630 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8631 to the end of the strings whereas this didn't. This would cause problems
8632 if strings read with d2i_ASN1_bytes() were later modified.
8633 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8635 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8636 data and it contains EOF it will end up returning an error. This is
8637 caused by input 46 bytes long. The cause is due to the way base64
8638 BIOs find the start of base64 encoded data. They do this by trying a
8639 trial decode on each line until they find one that works. When they
8640 do a flag is set and it starts again knowing it can pass all the
8641 data directly through the decoder. Unfortunately it doesn't reset
8642 the context it uses. This means that if EOF is reached an attempt
8643 is made to pass two EOFs through the context and this causes the
8644 resulting error. This can also cause other problems as well. As is
8645 usual with these problems it takes *ages* to find and the fix is
8646 trivial: move one line.
8647 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8649 *) Ugly workaround to get s_client and s_server working under Windows. The
8650 old code wouldn't work because it needed to select() on sockets and the
8651 tty (for keypresses and to see if data could be written). Win32 only
8652 supports select() on sockets so we select() with a 1s timeout on the
8653 sockets and then see if any characters are waiting to be read, if none
8654 are present then we retry, we also assume we can always write data to
8655 the tty. This isn't nice because the code then blocks until we've
8656 received a complete line of data and it is effectively polling the
8657 keyboard at 1s intervals: however it's quite a bit better than not
8658 working at all :-) A dedicated Windows application might handle this
8659 with an event loop for example.
8662 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8663 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8664 will be called when RSA_sign() and RSA_verify() are used. This is useful
8665 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8666 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8667 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8668 This necessitated the support of an extra signature type NID_md5_sha1
8669 for SSL signatures and modifications to the SSL library to use it instead
8670 of calling RSA_public_decrypt() and RSA_private_encrypt().
8673 *) Add new -verify -CAfile and -CApath options to the crl program, these
8674 will lookup a CRL issuers certificate and verify the signature in a
8675 similar way to the verify program. Tidy up the crl program so it
8676 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8677 less strict. It will now permit CRL extensions even if it is not
8678 a V2 CRL: this will allow it to tolerate some broken CRLs.
8681 *) Initialize all non-automatic variables each time one of the openssl
8682 sub-programs is started (this is necessary as they may be started
8683 multiple times from the "OpenSSL>" prompt).
8684 [Lennart Bang, Bodo Moeller]
8686 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8687 removing all other RSA functionality (this is what NO_RSA does). This
8688 is so (for example) those in the US can disable those operations covered
8689 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8693 *) Non-copying interface to BIO pairs.
8694 (still largely untested)
8697 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8698 ASCII string. This was handled independently in various places before.
8701 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8702 UTF8 strings a character at a time.
8705 *) Use client_version from client hello to select the protocol
8706 (s23_srvr.c) and for RSA client key exchange verification
8707 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8710 *) Add various utility functions to handle SPKACs, these were previously
8711 handled by poking round in the structure internals. Added new function
8712 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8713 print, verify and generate SPKACs. Based on an original idea from
8714 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8717 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8720 *) Allow the config file extension section to be overwritten on the
8721 command line. Based on an original idea from Massimiliano Pala
8722 <madwolf@comune.modena.it>. The new option is called -extensions
8723 and can be applied to ca, req and x509. Also -reqexts to override
8724 the request extensions in req and -crlexts to override the crl extensions
8728 *) Add new feature to the SPKAC handling in ca. Now you can include
8729 the same field multiple times by preceding it by "XXXX." for example:
8732 this is the same syntax as used in the req config file.
8735 *) Allow certificate extensions to be added to certificate requests. These
8736 are specified in a 'req_extensions' option of the req section of the
8737 config file. They can be printed out with the -text option to req but
8738 are otherwise ignored at present.
8741 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8742 data read consists of only the final block it would not decrypted because
8743 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8744 A misplaced 'break' also meant the decrypted final block might not be
8745 copied until the next read.
8748 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8749 a few extra parameters to the DH structure: these will be useful if
8750 for example we want the value of 'q' or implement X9.42 DH.
8753 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8754 provides hooks that allow the default DSA functions or functions on a
8755 "per key" basis to be replaced. This allows hardware acceleration and
8756 hardware key storage to be handled without major modification to the
8757 library. Also added low level modexp hooks and CRYPTO_EX structure and
8758 associated functions.
8761 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8762 as "read only": it can't be written to and the buffer it points to will
8763 not be freed. Reading from a read only BIO is much more efficient than
8764 a normal memory BIO. This was added because there are several times when
8765 an area of memory needs to be read from a BIO. The previous method was
8766 to create a memory BIO and write the data to it, this results in two
8767 copies of the data and an O(n^2) reading algorithm. There is a new
8768 function BIO_new_mem_buf() which creates a read only memory BIO from
8769 an area of memory. Also modified the PKCS#7 routines to use read only
8773 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8774 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8775 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8776 but a retry condition occured while trying to read the rest.
8779 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8780 NID_pkcs7_encrypted by default: this was wrong since this should almost
8781 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8782 the encrypted data type: this is a more sensible place to put it and it
8783 allows the PKCS#12 code to be tidied up that duplicated this
8787 *) Changed obj_dat.pl script so it takes its input and output files on
8788 the command line. This should avoid shell escape redirection problems
8792 *) Initial support for certificate extension requests, these are included
8793 in things like Xenroll certificate requests. Included functions to allow
8794 extensions to be obtained and added.
8797 *) -crlf option to s_client and s_server for sending newlines as
8798 CRLF (as required by many protocols).
8801 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8803 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8804 [Ralf S. Engelschall]
8806 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8807 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8809 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8813 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8814 DH parameters/keys (q is lost during that conversion, but the resulting
8815 DH parameters contain its length).
8817 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8818 much faster than DH_generate_parameters (which creates parameters
8819 where p = 2*q + 1), and also the smaller q makes DH computations
8820 much more efficient (160-bit exponentiation instead of 1024-bit
8821 exponentiation); so this provides a convenient way to support DHE
8822 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8823 utter importance to use
8824 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8826 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8827 when such DH parameters are used, because otherwise small subgroup
8828 attacks may become possible!
8831 *) Avoid memory leak in i2d_DHparams.
8834 *) Allow the -k option to be used more than once in the enc program:
8835 this allows the same encrypted message to be read by multiple recipients.
8838 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8839 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8840 it will always use the numerical form of the OID, even if it has a short
8844 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8845 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8846 otherwise bn_mod_exp was called. In the case of hardware keys for example
8847 no private key components need be present and it might store extra data
8848 in the RSA structure, which cannot be accessed from bn_mod_exp.
8849 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8850 private key operations.
8853 *) Added support for SPARC Linux.
8856 *) pem_password_cb function type incompatibly changed from
8857 typedef int pem_password_cb(char *buf, int size, int rwflag);
8859 ....(char *buf, int size, int rwflag, void *userdata);
8860 so that applications can pass data to their callbacks:
8861 The PEM[_ASN1]_{read,write}... functions and macros now take an
8862 additional void * argument, which is just handed through whenever
8863 the password callback is called.
8864 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8866 New function SSL_CTX_set_default_passwd_cb_userdata.
8868 Compatibility note: As many C implementations push function arguments
8869 onto the stack in reverse order, the new library version is likely to
8870 interoperate with programs that have been compiled with the old
8871 pem_password_cb definition (PEM_whatever takes some data that
8872 happens to be on the stack as its last argument, and the callback
8873 just ignores this garbage); but there is no guarantee whatsoever that
8876 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8877 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8878 problems not only on Windows, but also on some Unix platforms.
8879 To avoid problematic command lines, these definitions are now in an
8880 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8881 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8884 *) MIPS III/IV assembler module is reimplemented.
8887 *) More DES library cleanups: remove references to srand/rand and
8888 delete an unused file.
8891 *) Add support for the the free Netwide assembler (NASM) under Win32,
8892 since not many people have MASM (ml) and it can be hard to obtain.
8893 This is currently experimental but it seems to work OK and pass all
8894 the tests. Check out INSTALL.W32 for info.
8897 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8898 without temporary keys kept an extra copy of the server key,
8899 and connections with temporary keys did not free everything in case
8903 *) New function RSA_check_key and new openssl rsa option -check
8904 for verifying the consistency of RSA keys.
8905 [Ulf Moeller, Bodo Moeller]
8907 *) Various changes to make Win32 compile work:
8908 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8909 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8910 comparison" warnings.
8911 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8914 *) Add a debugging option to PKCS#5 v2 key generation function: when
8915 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8916 derived keys are printed to stderr.
8919 *) Copy the flags in ASN1_STRING_dup().
8920 [Roman E. Pavlov <pre@mo.msk.ru>]
8922 *) The x509 application mishandled signing requests containing DSA
8923 keys when the signing key was also DSA and the parameters didn't match.
8925 It was supposed to omit the parameters when they matched the signing key:
8926 the verifying software was then supposed to automatically use the CA's
8927 parameters if they were absent from the end user certificate.
8929 Omitting parameters is no longer recommended. The test was also
8930 the wrong way round! This was probably due to unusual behaviour in
8931 EVP_cmp_parameters() which returns 1 if the parameters match.
8932 This meant that parameters were omitted when they *didn't* match and
8933 the certificate was useless. Certificates signed with 'ca' didn't have
8935 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8937 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8938 The interface is as follows:
8939 Applications can use
8940 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8941 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8942 "off" is now the default.
8943 The library internally uses
8944 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8945 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8946 to disable memory-checking temporarily.
8948 Some inconsistent states that previously were possible (and were
8949 even the default) are now avoided.
8951 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8952 with each memory chunk allocated; this is occasionally more helpful
8953 than just having a counter.
8955 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8957 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8961 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8962 which largely parallels "options", but is for changing API behaviour,
8963 whereas "options" are about protocol behaviour.
8964 Initial "mode" flags are:
8966 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8967 a single record has been written.
8968 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8969 retries use the same buffer location.
8970 (But all of the contents must be
8974 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8977 *) Fix problems with no-hmac etc.
8978 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8980 *) New functions RSA_get_default_method(), RSA_set_method() and
8981 RSA_get_method(). These allows replacement of RSA_METHODs without having
8982 to mess around with the internals of an RSA structure.
8985 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8986 Also really enable memory leak checks in openssl.c and in some
8988 [Chad C. Mulligan, Bodo Moeller]
8990 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8991 up the length of negative integers. This has now been simplified to just
8992 store the length when it is first determined and use it later, rather
8993 than trying to keep track of where data is copied and updating it to
8995 [Steve Henson, reported by Brien Wheeler
8996 <bwheeler@authentica-security.com>]
8998 *) Add a new function PKCS7_signatureVerify. This allows the verification
8999 of a PKCS#7 signature but with the signing certificate passed to the
9000 function itself. This contrasts with PKCS7_dataVerify which assumes the
9001 certificate is present in the PKCS#7 structure. This isn't always the
9002 case: certificates can be omitted from a PKCS#7 structure and be
9003 distributed by "out of band" means (such as a certificate database).
9006 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
9007 function prototypes in pem.h, also change util/mkdef.pl to add the
9008 necessary function names.
9011 *) mk1mf.pl (used by Windows builds) did not properly read the
9012 options set by Configure in the top level Makefile, and Configure
9013 was not even able to write more than one option correctly.
9014 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
9017 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
9018 file to be loaded from a BIO or FILE pointer. The BIO version will
9019 for example allow memory BIOs to contain config info.
9022 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
9023 Whoever hopes to achieve shared-library compatibility across versions
9024 must use this, not the compile-time macro.
9025 (Exercise 0.9.4: Which is the minimum library version required by
9027 Note: All this applies only to multi-threaded programs, others don't
9031 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
9032 through a BIO pair triggered the default case, i.e.
9033 SSLerr(...,SSL_R_UNKNOWN_STATE).
9036 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
9037 can use the SSL library even if none of the specific BIOs is
9041 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
9042 for the encoded length.
9043 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
9045 *) Add initial documentation of the X509V3 functions.
9048 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
9049 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
9050 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
9051 secure PKCS#8 private key format with a high iteration count.
9054 *) Fix determination of Perl interpreter: A perl or perl5
9055 _directory_ in $PATH was also accepted as the interpreter.
9056 [Ralf S. Engelschall]
9058 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
9059 wrong with it but it was very old and did things like calling
9060 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
9064 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
9065 to use the new extension code.
9068 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
9069 with macros. This should make it easier to change their form, add extra
9070 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
9074 *) Add to configuration table a new entry that can specify an alternative
9075 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
9076 according to Mark Crispin <MRC@Panda.COM>.
9080 *) DES CBC did not update the IV. Weird.
9083 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
9084 Changing the behaviour of the former might break existing programs --
9085 where IV updating is needed, des_ncbc_encrypt can be used.
9088 *) When bntest is run from "make test" it drives bc to check its
9089 calculations, as well as internally checking them. If an internal check
9090 fails, it needs to cause bc to give a non-zero result or make test carries
9091 on without noticing the failure. Fixed.
9094 *) DES library cleanups.
9097 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
9098 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
9099 ciphers. NOTE: although the key derivation function has been verified
9100 against some published test vectors it has not been extensively tested
9101 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
9105 *) Instead of "mkdir -p", which is not fully portable, use new
9106 Perl script "util/mkdir-p.pl".
9109 *) Rewrite the way password based encryption (PBE) is handled. It used to
9110 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
9111 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
9112 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
9113 the 'parameter' field of the AlgorithmIdentifier is passed to the
9114 underlying key generation function so it must do its own ASN1 parsing.
9115 This has also changed the EVP_PBE_CipherInit() function which now has a
9116 'parameter' argument instead of literal salt and iteration count values
9117 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
9120 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
9121 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
9122 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
9123 KEY" because this clashed with PKCS#8 unencrypted string. Since this
9124 value was just used as a "magic string" and not used directly its
9125 value doesn't matter.
9128 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9132 *) "linux-sparc64" configuration (ultrapenguin).
9133 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9134 "linux-sparc" configuration.
9135 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9137 *) config now generates no-xxx options for missing ciphers.
9140 *) Support the EBCDIC character set (work in progress).
9141 File ebcdic.c not yet included because it has a different license.
9142 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9144 *) Support BS2000/OSD-POSIX.
9145 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9147 *) Make callbacks for key generation use void * instead of char *.
9150 *) Make S/MIME samples compile (not yet tested).
9153 *) Additional typesafe stacks.
9156 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9160 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9162 *) New configuration variant "sco5-gcc".
9164 *) Updated some demos.
9165 [Sean O Riordain, Wade Scholine]
9167 *) Add missing BIO_free at exit of pkcs12 application.
9170 *) Fix memory leak in conf.c.
9173 *) Updates for Win32 to assembler version of MD5.
9176 *) Set #! path to perl in apps/der_chop to where we found it
9177 instead of using a fixed path.
9180 *) SHA library changes for irix64-mips4-cc.
9183 *) Improvements for VMS support.
9187 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9189 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9190 This also avoids the problems with SC4.2 and unpatched SC5.
9191 [Andy Polyakov <appro@fy.chalmers.se>]
9193 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9194 These are required because of the typesafe stack would otherwise break
9195 existing code. If old code used a structure member which used to be STACK
9196 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9197 sk_num or sk_value it would produce an error because the num, data members
9198 are not present in STACK_OF. Now it just produces a warning. sk_set
9199 replaces the old method of assigning a value to sk_value
9200 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9201 that does this will no longer work (and should use sk_set instead) but
9202 this could be regarded as a "questionable" behaviour anyway.
9205 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9206 correctly handle encrypted S/MIME data.
9209 *) Change type of various DES function arguments from des_cblock
9210 (which means, in function argument declarations, pointer to char)
9211 to des_cblock * (meaning pointer to array with 8 char elements),
9212 which allows the compiler to do more typechecking; it was like
9213 that back in SSLeay, but with lots of ugly casts.
9215 Introduce new type const_des_cblock.
9218 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9219 problems: find RecipientInfo structure that matches recipient certificate
9220 and initialise the ASN1 structures properly based on passed cipher.
9223 *) Belatedly make the BN tests actually check the results.
9226 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9227 to and from BNs: it was completely broken. New compilation option
9228 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9229 key elements as negative integers.
9232 *) Reorganize and speed up MD5.
9233 [Andy Polyakov <appro@fy.chalmers.se>]
9236 [Richard Levitte <richard@levitte.org>]
9238 *) New option -out to asn1parse to allow the parsed structure to be
9239 output to a file. This is most useful when combined with the -strparse
9240 option to examine the output of things like OCTET STRINGS.
9243 *) Make SSL library a little more fool-proof by not requiring any longer
9244 that SSL_set_{accept,connect}_state be called before
9245 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9246 in many applications because usually everything *appeared* to work as
9247 intended anyway -- now it really works as intended).
9250 *) Move openssl.cnf out of lib/.
9253 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
9254 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
9255 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
9256 [Ralf S. Engelschall]
9258 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
9259 handle PKCS#7 enveloped data properly.
9260 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
9262 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
9263 copying pointers. The cert_st handling is changed by this in
9264 various ways (and thus what used to be known as ctx->default_cert
9265 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
9266 any longer when s->cert does not give us what we need).
9267 ssl_cert_instantiate becomes obsolete by this change.
9268 As soon as we've got the new code right (possibly it already is?),
9269 we have solved a couple of bugs of the earlier code where s->cert
9270 was used as if it could not have been shared with other SSL structures.
9272 Note that using the SSL API in certain dirty ways now will result
9273 in different behaviour than observed with earlier library versions:
9274 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
9275 does not influence s as it used to.
9277 In order to clean up things more thoroughly, inside SSL_SESSION
9278 we don't use CERT any longer, but a new structure SESS_CERT
9279 that holds per-session data (if available); currently, this is
9280 the peer's certificate chain and, for clients, the server's certificate
9281 and temporary key. CERT holds only those values that can have
9282 meaningful defaults in an SSL_CTX.
9285 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
9286 from the internal representation. Various PKCS#7 fixes: remove some
9287 evil casts and set the enc_dig_alg field properly based on the signing
9291 *) Allow PKCS#12 password to be set from the command line or the
9292 environment. Let 'ca' get its config file name from the environment
9293 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
9297 *) Allow certificate policies extension to use an IA5STRING for the
9298 organization field. This is contrary to the PKIX definition but
9299 VeriSign uses it and IE5 only recognises this form. Document 'x509'
9303 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
9304 without disallowing inline assembler and the like for non-pedantic builds.
9307 *) Support Borland C++ builder.
9308 [Janez Jere <jj@void.si>, modified by Ulf Möller]
9313 *) SHA-1 cleanups and performance enhancements.
9314 [Andy Polyakov <appro@fy.chalmers.se>]
9316 *) Sparc v8plus assembler for the bignum library.
9317 [Andy Polyakov <appro@fy.chalmers.se>]
9319 *) Accept any -xxx and +xxx compiler options in Configure.
9322 *) Update HPUX configuration.
9325 *) Add missing sk_<type>_unshift() function to safestack.h
9326 [Ralf S. Engelschall]
9328 *) New function SSL_CTX_use_certificate_chain_file that sets the
9329 "extra_cert"s in addition to the certificate. (This makes sense
9330 only for "PEM" format files, as chains as a whole are not
9334 *) Support verify_depth from the SSL API.
9335 x509_vfy.c had what can be considered an off-by-one-error:
9336 Its depth (which was not part of the external interface)
9337 was actually counting the number of certificates in a chain;
9338 now it really counts the depth.
9341 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
9342 instead of X509err, which often resulted in confusing error
9343 messages since the error codes are not globally unique
9344 (e.g. an alleged error in ssl3_accept when a certificate
9345 didn't match the private key).
9347 *) New function SSL_CTX_set_session_id_context that allows to set a default
9348 value (so that you don't need SSL_set_session_id_context for each
9349 connection using the SSL_CTX).
9352 *) OAEP decoding bug fix.
9355 *) Support INSTALL_PREFIX for package builders, as proposed by
9359 *) New Configure options "threads" and "no-threads". For systems
9360 where the proper compiler options are known (currently Solaris
9361 and Linux), "threads" is the default.
9364 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
9367 *) Install various scripts to $(OPENSSLDIR)/misc, not to
9368 $(INSTALLTOP)/bin -- they shouldn't clutter directories
9369 such as /usr/local/bin.
9372 *) "make linux-shared" to build shared libraries.
9373 [Niels Poppe <niels@netbox.org>]
9375 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
9378 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
9379 extension adding in x509 utility.
9382 *) Remove NOPROTO sections and error code comments.
9385 *) Partial rewrite of the DEF file generator to now parse the ANSI
9389 *) New Configure options --prefix=DIR and --openssldir=DIR.
9392 *) Complete rewrite of the error code script(s). It is all now handled
9393 by one script at the top level which handles error code gathering,
9394 header rewriting and C source file generation. It should be much better
9395 than the old method: it now uses a modified version of Ulf's parser to
9396 read the ANSI prototypes in all header files (thus the old K&R definitions
9397 aren't needed for error creation any more) and do a better job of
9398 translating function codes into names. The old 'ASN1 error code imbedded
9399 in a comment' is no longer necessary and it doesn't use .err files which
9400 have now been deleted. Also the error code call doesn't have to appear all
9401 on one line (which resulted in some large lines...).
9404 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
9407 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
9408 0 (which usually indicates a closed connection), but continue reading.
9411 *) Fix some race conditions.
9414 *) Add support for CRL distribution points extension. Add Certificate
9415 Policies and CRL distribution points documentation.
9418 *) Move the autogenerated header file parts to crypto/opensslconf.h.
9421 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
9422 8 of keying material. Merlin has also confirmed interop with this fix
9423 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
9424 [Merlin Hughes <merlin@baltimore.ie>]
9426 *) Fix lots of warnings.
9427 [Richard Levitte <levitte@stacken.kth.se>]
9429 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
9430 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9431 [Richard Levitte <levitte@stacken.kth.se>]
9433 *) Fix problems with sizeof(long) == 8.
9434 [Andy Polyakov <appro@fy.chalmers.se>]
9436 *) Change functions to ANSI C.
9439 *) Fix typos in error codes.
9440 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9442 *) Remove defunct assembler files from Configure.
9445 *) SPARC v8 assembler BIGNUM implementation.
9446 [Andy Polyakov <appro@fy.chalmers.se>]
9448 *) Support for Certificate Policies extension: both print and set.
9449 Various additions to support the r2i method this uses.
9452 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9453 return a const string when you are expecting an allocated buffer.
9456 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9457 types DirectoryString and DisplayText.
9460 *) Add code to allow r2i extensions to access the configuration database,
9461 add an LHASH database driver and add several ctx helper functions.
9464 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9465 fail when they extended the size of a BIGNUM.
9468 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9469 support typesafe stack.
9472 *) Fix typo in SSL_[gs]et_options().
9473 [Nils Frostberg <nils@medcom.se>]
9475 *) Delete various functions and files that belonged to the (now obsolete)
9476 old X509V3 handling code.
9479 *) New Configure option "rsaref".
9482 *) Don't auto-generate pem.h.
9485 *) Introduce type-safe ASN.1 SETs.
9488 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9489 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9491 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9492 that links with OpenSSL (well at least cause lots of warnings), but fear
9493 not: the conversion is trivial, and it eliminates loads of evil casts. A
9494 few STACKed things have been converted already. Feel free to convert more.
9495 In the fullness of time, I'll do away with the STACK type altogether.
9498 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9499 specified in <certfile> by updating the entry in the index.txt file.
9500 This way one no longer has to edit the index.txt file manually for
9501 revoking a certificate. The -revoke option does the gory details now.
9502 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9504 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9505 `-text' option at all and this way the `-noout -text' combination was
9506 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9507 [Ralf S. Engelschall]
9509 *) Make sure a corresponding plain text error message exists for the
9510 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9511 verify callback function determined that a certificate was revoked.
9512 [Ralf S. Engelschall]
9514 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9515 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9516 all available cipers including rc5, which was forgotten until now.
9517 In order to let the testing shell script know which algorithms
9518 are available, a new (up to now undocumented) command
9519 "openssl list-cipher-commands" is used.
9522 *) Bugfix: s_client occasionally would sleep in select() when
9523 it should have checked SSL_pending() first.
9526 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9527 the raw DSA values prior to ASN.1 encoding.
9530 *) Tweaks to Configure
9531 [Niels Poppe <niels@netbox.org>]
9533 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9537 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9540 *) New config option to avoid instructions that are illegal on the 80386.
9541 The default code is faster, but requires at least a 486.
9544 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9545 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9546 same as SSL2_VERSION anyway.
9549 *) New "-showcerts" option for s_client.
9552 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9553 application. Various cleanups and fixes.
9556 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9557 modify error routines to work internally. Add error codes and PBE init
9558 to library startup routines.
9561 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9562 packing functions to asn1 and evp. Changed function names and error
9563 codes along the way.
9566 *) PKCS12 integration: and so it begins... First of several patches to
9567 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9568 objects to objects.h
9571 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9572 and display support for Thawte strong extranet extension.
9575 *) Add LinuxPPC support.
9576 [Jeff Dubrule <igor@pobox.org>]
9578 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9579 bn_div_words in alpha.s.
9580 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9582 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9583 OAEP isn't supported when OpenSSL is built with RSAref.
9584 [Ulf Moeller <ulf@fitug.de>]
9586 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9587 so they no longer are missing under -DNOPROTO.
9588 [Soren S. Jorvang <soren@t.dk>]
9591 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9593 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9594 doesn't work when the session is reused. Coming soon!
9597 *) Fix a security hole, that allows sessions to be reused in the wrong
9598 context thus bypassing client cert protection! All software that uses
9599 client certs and session caches in multiple contexts NEEDS PATCHING to
9600 allow session reuse! A fuller solution is in the works.
9601 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9603 *) Some more source tree cleanups (removed obsolete files
9604 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9605 permission on "config" script to be executable) and a fix for the INSTALL
9607 [Ulf Moeller <ulf@fitug.de>]
9609 *) Remove some legacy and erroneous uses of malloc, free instead of
9611 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9613 *) Make rsa_oaep_test return non-zero on error.
9614 [Ulf Moeller <ulf@fitug.de>]
9616 *) Add support for native Solaris shared libraries. Configure
9617 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9618 if someone would make that last step automatic.
9619 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9621 *) ctx_size was not built with the right compiler during "make links". Fixed.
9624 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9625 except NULL ciphers". This means the default cipher list will no longer
9626 enable NULL ciphers. They need to be specifically enabled e.g. with
9627 the string "DEFAULT:eNULL".
9630 *) Fix to RSA private encryption routines: if p < q then it would
9631 occasionally produce an invalid result. This will only happen with
9632 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9635 *) Be less restrictive and allow also `perl util/perlpath.pl
9636 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9637 because this way one can also use an interpreter named `perl5' (which is
9638 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9639 installed as `perl').
9640 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9642 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9643 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9645 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9646 advapi32.lib to Win32 build and change the pem test comparision
9647 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9648 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9649 and crypto/des/ede_cbcm_enc.c.
9652 *) DES quad checksum was broken on big-endian architectures. Fixed.
9655 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9656 Win32 test batch file so it (might) work again. The Win32 test batch file
9657 is horrible: I feel ill....
9660 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9661 in e_os.h. Audit of header files to check ANSI and non ANSI
9662 sections: 10 functions were absent from non ANSI section and not exported
9663 from Windows DLLs. Fixed up libeay.num for new functions.
9666 *) Make `openssl version' output lines consistent.
9667 [Ralf S. Engelschall]
9669 *) Fix Win32 symbol export lists for BIO functions: Added
9670 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9671 to ms/libeay{16,32}.def.
9672 [Ralf S. Engelschall]
9674 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9675 fine under Unix and passes some trivial tests I've now added. But the
9676 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9677 added to make sure no one expects that this stuff really works in the
9678 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9679 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9681 [Ralf S. Engelschall]
9683 *) Fix the generation of two part addresses in perl.
9684 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9686 *) Add config entry for Linux on MIPS.
9687 [John Tobey <jtobey@channel1.com>]
9689 *) Make links whenever Configure is run, unless we are on Windoze.
9692 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9693 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9697 *) Add a useful kludge to allow package maintainers to specify compiler and
9698 other platforms details on the command line without having to patch the
9699 Configure script everytime: One now can use ``perl Configure
9700 <id>:<details>'', i.e. platform ids are allowed to have details appended
9701 to them (seperated by colons). This is treated as there would be a static
9702 pre-configured entry in Configure's %table under key <id> with value
9703 <details> and ``perl Configure <id>'' is called. So, when you want to
9704 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9705 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9706 now, which overrides the FreeBSD-elf entry on-the-fly.
9707 [Ralf S. Engelschall]
9709 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9712 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9713 on the `perl Configure ...' command line. This way one can compile
9714 OpenSSL libraries with Position Independent Code (PIC) which is needed
9715 for linking it into DSOs.
9716 [Ralf S. Engelschall]
9718 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9722 *) Cleaned up the LICENSE document: The official contact for any license
9723 questions now is the OpenSSL core team under openssl-core@openssl.org.
9724 And add a paragraph about the dual-license situation to make sure people
9725 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9726 to the OpenSSL toolkit.
9727 [Ralf S. Engelschall]
9729 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9730 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9731 Additonally cleaned up the `make links' target: Remove unnecessary
9732 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9733 to speed processing and no longer clutter the display with confusing
9734 stuff. Instead only the actually done links are displayed.
9735 [Ralf S. Engelschall]
9737 *) Permit null encryption ciphersuites, used for authentication only. It used
9738 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9739 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9743 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9744 signed attributes when verifying signatures (this would break them),
9745 the detached data encoding was wrong and public keys obtained using
9746 X509_get_pubkey() weren't freed.
9749 *) Add text documentation for the BUFFER functions. Also added a work around
9750 to a Win95 console bug. This was triggered by the password read stuff: the
9751 last character typed gets carried over to the next fread(). If you were
9752 generating a new cert request using 'req' for example then the last
9753 character of the passphrase would be CR which would then enter the first
9757 *) Added the new `Includes OpenSSL Cryptography Software' button as
9758 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9759 button and can be used by applications based on OpenSSL to show the
9760 relationship to the OpenSSL project.
9761 [Ralf S. Engelschall]
9763 *) Remove confusing variables in function signatures in files
9764 ssl/ssl_lib.c and ssl/ssl.h.
9765 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9767 *) Don't install bss_file.c under PREFIX/include/
9768 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9770 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9771 functions that return function pointers and has support for NT specific
9772 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9773 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9774 unsigned to signed types: this was killing the Win32 compile.
9777 *) Add new certificate file to stack functions,
9778 SSL_add_dir_cert_subjects_to_stack() and
9779 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9780 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9781 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9782 This means that Apache-SSL and similar packages don't have to mess around
9783 to add as many CAs as they want to the preferred list.
9786 *) Experiment with doxygen documentation. Currently only partially applied to
9788 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9789 openssl.doxy as the configuration file.
9792 *) Get rid of remaining C++-style comments which strict C compilers hate.
9793 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9795 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9796 compiled in by default: it has problems with large keys.
9799 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9800 DH private keys and/or callback functions which directly correspond to
9801 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9802 is needed for applications which have to configure certificates on a
9803 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9805 For the RSA certificate situation is makes no difference, but
9806 for the DSA certificate situation this fixes the "no shared cipher"
9807 problem where the OpenSSL cipher selection procedure failed because the
9808 temporary keys were not overtaken from the context and the API provided
9809 no way to reconfigure them.
9810 The new functions now let applications reconfigure the stuff and they
9811 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9812 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9813 non-public-API function ssl_cert_instantiate() is used as a helper
9814 function and also to reduce code redundancy inside ssl_rsa.c.
9815 [Ralf S. Engelschall]
9817 *) Move s_server -dcert and -dkey options out of the undocumented feature
9818 area because they are useful for the DSA situation and should be
9819 recognized by the users.
9820 [Ralf S. Engelschall]
9822 *) Fix the cipher decision scheme for export ciphers: the export bits are
9823 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9824 SSL_EXP_MASK. So, the original variable has to be used instead of the
9825 already masked variable.
9826 [Richard Levitte <levitte@stacken.kth.se>]
9828 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9829 [Richard Levitte <levitte@stacken.kth.se>]
9831 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9832 from `int' to `unsigned int' because it's a length and initialized by
9833 EVP_DigestFinal() which expects an `unsigned int *'.
9834 [Richard Levitte <levitte@stacken.kth.se>]
9836 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9837 script. Instead use the usual Shell->Perl transition trick.
9838 [Ralf S. Engelschall]
9840 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9841 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9842 -noout -modulus' as it's already the case for `openssl rsa -noout
9843 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9844 currently the public key is printed (a decision which was already done by
9845 `openssl dsa -modulus' in the past) which serves a similar purpose.
9846 Additionally the NO_RSA no longer completely removes the whole -modulus
9847 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9849 [Ralf S. Engelschall]
9851 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9852 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9853 [Arne Ansper <arne@ats.cyber.ee>]
9855 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9856 to be added. Now both 'req' and 'ca' can use new objects defined in the
9860 *) Add cool BIO that does syslog (or event log on NT).
9861 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9863 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9864 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9865 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9866 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9869 *) Add preliminary config info for new extension code.
9872 *) Make RSA_NO_PADDING really use no padding.
9873 [Ulf Moeller <ulf@fitug.de>]
9875 *) Generate errors when private/public key check is done.
9878 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9879 for some CRL extensions and new objects added.
9882 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9883 key usage extension and fuller support for authority key id.
9886 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9887 padding method for RSA, which is recommended for new applications in PKCS
9888 #1 v2.0 (RFC 2437, October 1998).
9889 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9890 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9891 against Bleichbacher's attack on RSA.
9892 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9895 *) Updates to the new SSL compression code
9896 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9898 *) Fix so that the version number in the master secret, when passed
9899 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9900 (because the server will not accept higher), that the version number
9901 is 0x03,0x01, not 0x03,0x00
9902 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9904 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9905 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9906 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9909 *) Support for RAW extensions where an arbitrary extension can be
9910 created by including its DER encoding. See apps/openssl.cnf for
9914 *) Make sure latest Perl versions don't interpret some generated C array
9915 code as Perl array code in the crypto/err/err_genc.pl script.
9916 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9918 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9919 not many people have the assembler. Various Win32 compilation fixes and
9920 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9924 *) Modify configure script 'Configure' to automatically create crypto/date.h
9925 file under Win32 and also build pem.h from pem.org. New script
9926 util/mkfiles.pl to create the MINFO file on environments that can't do a
9927 'make files': perl util/mkfiles.pl >MINFO should work.
9930 *) Major rework of DES function declarations, in the pursuit of correctness
9931 and purity. As a result, many evil casts evaporated, and some weirdness,
9932 too. You may find this causes warnings in your code. Zapping your evil
9933 casts will probably fix them. Mostly.
9936 *) Fix for a typo in asn1.h. Bug fix to object creation script
9937 obj_dat.pl. It considered a zero in an object definition to mean
9938 "end of object": none of the objects in objects.h have any zeros
9939 so it wasn't spotted.
9940 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9942 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9943 Masking (CBCM). In the absence of test vectors, the best I have been able
9944 to do is check that the decrypt undoes the encrypt, so far. Send me test
9945 vectors if you have them.
9948 *) Correct calculation of key length for export ciphers (too much space was
9949 allocated for null ciphers). This has not been tested!
9952 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9953 message is now correct (it understands "crypto" and "ssl" on its
9954 command line). There is also now an "update" option. This will update
9955 the util/ssleay.num and util/libeay.num files with any new functions.
9957 perl util/mkdef.pl crypto ssl update
9958 it will update them.
9961 *) Overhauled the Perl interface (perl/*):
9962 - ported BN stuff to OpenSSL's different BN library
9963 - made the perl/ source tree CVS-aware
9964 - renamed the package from SSLeay to OpenSSL (the files still contain
9965 their history because I've copied them in the repository)
9966 - removed obsolete files (the test scripts will be replaced
9967 by better Test::Harness variants in the future)
9968 [Ralf S. Engelschall]
9970 *) First cut for a very conservative source tree cleanup:
9971 1. merge various obsolete readme texts into doc/ssleay.txt
9972 where we collect the old documents and readme texts.
9973 2. remove the first part of files where I'm already sure that we no
9974 longer need them because of three reasons: either they are just temporary
9975 files which were left by Eric or they are preserved original files where
9976 I've verified that the diff is also available in the CVS via "cvs diff
9977 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9978 the crypto/md/ stuff).
9979 [Ralf S. Engelschall]
9981 *) More extension code. Incomplete support for subject and issuer alt
9982 name, issuer and authority key id. Change the i2v function parameters
9983 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9984 what that's for :-) Fix to ASN1 macro which messed up
9985 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9988 *) Preliminary support for ENUMERATED type. This is largely copied from the
9992 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9993 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9995 *) Make sure `make rehash' target really finds the `openssl' program.
9996 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9998 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9999 like to hear about it if this slows down other processors.
10002 *) Add CygWin32 platform information to Configure script.
10003 [Alan Batie <batie@aahz.jf.intel.com>]
10005 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
10006 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
10008 *) New program nseq to manipulate netscape certificate sequences
10011 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
10015 *) Fixes to BN code. Previously the default was to define BN_RECURSION
10016 but the BN code had some problems that would cause failures when
10017 doing certificate verification and some other functions.
10018 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10020 *) Add ASN1 and PEM code to support netscape certificate sequences.
10023 *) Add ASN1 and PEM code to support netscape certificate sequences.
10026 *) Add several PKIX and private extended key usage OIDs.
10029 *) Modify the 'ca' program to handle the new extension code. Modify
10030 openssl.cnf for new extension format, add comments.
10033 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
10034 and add a sample to openssl.cnf so req -x509 now adds appropriate
10038 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
10039 error code, add initial support to X509_print() and x509 application.
10042 *) Takes a deep breath and start addding X509 V3 extension support code. Add
10043 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
10044 stuff is currently isolated and isn't even compiled yet.
10047 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
10048 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
10049 Removed the versions check from X509 routines when loading extensions:
10050 this allows certain broken certificates that don't set the version
10051 properly to be processed.
10054 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
10055 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
10056 can still be regenerated with "make depend".
10059 *) Spelling mistake in C version of CAST-128.
10060 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
10062 *) Changes to the error generation code. The perl script err-code.pl
10063 now reads in the old error codes and retains the old numbers, only
10064 adding new ones if necessary. It also only changes the .err files if new
10065 codes are added. The makefiles have been modified to only insert errors
10066 when needed (to avoid needlessly modifying header files). This is done
10067 by only inserting errors if the .err file is newer than the auto generated
10068 C file. To rebuild all the error codes from scratch (the old behaviour)
10069 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
10070 or delete all the .err files.
10073 *) CAST-128 was incorrectly implemented for short keys. The C version has
10074 been fixed, but is untested. The assembler versions are also fixed, but
10075 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
10076 to regenerate it if needed.
10077 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
10078 Hagino <itojun@kame.net>]
10080 *) File was opened incorrectly in randfile.c.
10081 [Ulf Möller <ulf@fitug.de>]
10083 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
10084 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
10085 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
10086 al: it's just almost always a UTCTime. Note this patch adds new error
10087 codes so do a "make errors" if there are problems.
10090 *) Correct Linux 1 recognition in config.
10091 [Ulf Möller <ulf@fitug.de>]
10093 *) Remove pointless MD5 hash when using DSA keys in ca.
10094 [Anonymous <nobody@replay.com>]
10096 *) Generate an error if given an empty string as a cert directory. Also
10097 generate an error if handed NULL (previously returned 0 to indicate an
10098 error, but didn't set one).
10099 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
10101 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
10104 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
10105 parameters. This was causing a warning which killed off the Win32 compile.
10108 *) Remove C++ style comments from crypto/bn/bn_local.h.
10109 [Neil Costigan <neil.costigan@celocom.com>]
10111 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
10112 based on a text string, looking up short and long names and finally
10113 "dot" format. The "dot" format stuff didn't work. Added new function
10114 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
10115 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
10116 OID is not part of the table.
10119 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
10120 X509_LOOKUP_by_alias().
10123 *) Sort openssl functions by name.
10126 *) Get the gendsa program working (hopefully) and add it to app list. Remove
10127 encryption from sample DSA keys (in case anyone is interested the password
10131 *) Make _all_ *_free functions accept a NULL pointer.
10132 [Frans Heymans <fheymans@isaserver.be>]
10134 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10136 [Anonymous <nobody@replay.com>]
10138 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10139 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10141 *) Don't blow it for numeric -newkey arguments to apps/req.
10142 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10144 *) Temp key "for export" tests were wrong in s3_srvr.c.
10145 [Anonymous <nobody@replay.com>]
10147 *) Add prototype for temp key callback functions
10148 SSL_CTX_set_tmp_{rsa,dh}_callback().
10151 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10152 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10155 *) X509_name_add_entry() freed the wrong thing after an error.
10156 [Arne Ansper <arne@ats.cyber.ee>]
10158 *) rsa_eay.c would attempt to free a NULL context.
10159 [Arne Ansper <arne@ats.cyber.ee>]
10161 *) BIO_s_socket() had a broken should_retry() on Windoze.
10162 [Arne Ansper <arne@ats.cyber.ee>]
10164 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10165 [Arne Ansper <arne@ats.cyber.ee>]
10167 *) Make sure the already existing X509_STORE->depth variable is initialized
10168 in X509_STORE_new(), but document the fact that this variable is still
10169 unused in the certificate verification process.
10170 [Ralf S. Engelschall]
10172 *) Fix the various library and apps files to free up pkeys obtained from
10173 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10176 *) Fix reference counting in X509_PUBKEY_get(). This makes
10177 demos/maurice/example2.c work, amongst others, probably.
10178 [Steve Henson and Ben Laurie]
10180 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10181 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10182 are no longer created. This way we have a single and consistent command
10183 line interface `openssl <command>', similar to `cvs <command>'.
10184 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10186 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10187 BIT STRING wrapper always have zero unused bits.
10190 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10193 *) Make the top-level INSTALL documentation easier to understand.
10196 *) Makefiles updated to exit if an error occurs in a sub-directory
10197 make (including if user presses ^C) [Paul Sutton]
10199 *) Make Montgomery context stuff explicit in RSA data structure.
10202 *) Fix build order of pem and err to allow for generated pem.h.
10205 *) Fix renumbering bug in X509_NAME_delete_entry().
10208 *) Enhanced the err-ins.pl script so it makes the error library number
10209 global and can add a library name. This is needed for external ASN1 and
10210 other error libraries.
10213 *) Fixed sk_insert which never worked properly.
10216 *) Fix ASN1 macros so they can handle indefinite length construted
10217 EXPLICIT tags. Some non standard certificates use these: they can now
10221 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10222 into a single doc/ssleay.txt bundle. This way the information is still
10223 preserved but no longer messes up this directory. Now it's new room for
10224 the new set of documenation files.
10225 [Ralf S. Engelschall]
10227 *) SETs were incorrectly DER encoded. This was a major pain, because they
10228 shared code with SEQUENCEs, which aren't coded the same. This means that
10229 almost everything to do with SETs or SEQUENCEs has either changed name or
10230 number of arguments.
10231 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10233 *) Fix test data to work with the above.
10236 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10237 was already fixed by Eric for 0.9.1 it seems.
10238 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10240 *) Autodetect FreeBSD3.
10243 *) Fix various bugs in Configure. This affects the following platforms:
10247 unixware-2.0-pentium
10251 *) Eliminate generated files from CVS. Reorder tests to regenerate files
10252 before they are needed.
10255 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
10259 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
10261 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
10262 changed SSLeay to OpenSSL in version strings.
10263 [Ralf S. Engelschall]
10265 *) Some fixups to the top-level documents.
10268 *) Fixed the nasty bug where rsaref.h was not found under compile-time
10269 because the symlink to include/ was missing.
10270 [Ralf S. Engelschall]
10272 *) Incorporated the popular no-RSA/DSA-only patches
10273 which allow to compile a RSA-free SSLeay.
10274 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
10276 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
10277 when "ssleay" is still not found.
10278 [Ralf S. Engelschall]
10280 *) Added more platforms to Configure: Cray T3E, HPUX 11,
10281 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
10283 *) Updated the README file.
10284 [Ralf S. Engelschall]
10286 *) Added various .cvsignore files in the CVS repository subdirs
10287 to make a "cvs update" really silent.
10288 [Ralf S. Engelschall]
10290 *) Recompiled the error-definition header files and added
10291 missing symbols to the Win32 linker tables.
10292 [Ralf S. Engelschall]
10294 *) Cleaned up the top-level documents;
10295 o new files: CHANGES and LICENSE
10296 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
10297 o merged COPYRIGHT into LICENSE
10298 o removed obsolete TODO file
10299 o renamed MICROSOFT to INSTALL.W32
10300 [Ralf S. Engelschall]
10302 *) Removed dummy files from the 0.9.1b source tree:
10303 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
10304 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
10305 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
10306 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
10307 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
10308 [Ralf S. Engelschall]
10310 *) Added various platform portability fixes.
10313 *) The Genesis of the OpenSSL rpject:
10314 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
10315 Young and Tim J. Hudson created while they were working for C2Net until
10317 [The OpenSSL Project]
10320 Changes between 0.9.0b and 0.9.1b [not released]
10322 *) Updated a few CA certificates under certs/
10325 *) Changed some BIGNUM api stuff.
10328 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
10329 DGUX x86, Linux Alpha, etc.
10332 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
10333 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
10337 *) Add -strparse option to asn1pars program which parses nested
10339 [Dr Stephen Henson <shenson@bigfoot.com>]
10341 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
10344 *) DSA fix for "ca" program.
10347 *) Added "-genkey" option to "dsaparam" program.
10350 *) Added RIPE MD160 (rmd160) message digest.
10353 *) Added -a (all) option to "ssleay version" command.
10356 *) Added PLATFORM define which is the id given to Configure.
10359 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
10362 *) Extended the ASN.1 parser routines.
10365 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
10368 *) Added a BN_CTX to the BN library.
10371 *) Fixed the weak key values in DES library
10374 *) Changed API in EVP library for cipher aliases.
10377 *) Added support for RC2/64bit cipher.
10380 *) Converted the lhash library to the crypto/mem.c functions.
10383 *) Added more recognized ASN.1 object ids.
10386 *) Added more RSA padding checks for SSL/TLS.
10389 *) Added BIO proxy/filter functionality.
10392 *) Added extra_certs to SSL_CTX which can be used
10393 send extra CA certificates to the client in the CA cert chain sending
10394 process. It can be configured with SSL_CTX_add_extra_chain_cert().
10397 *) Now Fortezza is denied in the authentication phase because
10398 this is key exchange mechanism is not supported by SSLeay at all.
10401 *) Additional PKCS1 checks.
10404 *) Support the string "TLSv1" for all TLS v1 ciphers.
10407 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
10408 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
10411 *) Fixed a few memory leaks.
10414 *) Fixed various code and comment typos.
10417 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
10418 bytes sent in the client random.
10419 [Edward Bishop <ebishop@spyglass.com>]