5 Changes between 1.0.1 and 1.1.0 [xx XXX xxxx]
7 *) Support for automatic EC temporary key parameter selection. If enabled
8 the most preferred EC parameters are automatically used instead of
9 hardcoded fixed parameters. Now a server just has to call:
10 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
11 support ECDH and use the most appropriate parameters.
14 *) Enhance and tidy EC curve and point format TLS extension code. Use
15 static structures instead of allocation if default values are used.
16 New ctrls to set curves we wish to support and to retrieve shared curves.
17 Print out shared curves in s_server. New options to s_server and s_client
18 to set list of supported curves.
21 *) New ctrls to retrieve supported signature algorithms and
22 supported curve values as an array of NIDs. Extend openssl utility
23 to print out received values.
26 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
27 between NIDs and the more common NIST names such as "P-256". Enhance
28 ecparam utility and ECC method to recognise the NIST names for curves.
31 *) Enhance SSL/TLS certificate chain handling to support different
32 chains for each certificate instead of one chain in the parent SSL_CTX.
35 *) Support for fixed DH ciphersuite client authentication: where both
36 server and client use DH certificates with common parameters.
39 *) Support for fixed DH ciphersuites: those requiring DH server
43 *) Transparently support X9.42 DH parameters when calling
44 PEM_read_bio_DHparameters. This means existing applications can handle
45 the new parameter format automatically.
48 *) Initial experimental support for X9.42 DH parameter format: mainly
49 to support use of 'q' parameter for RFC5114 parameters.
52 *) Add DH parameters from RFC5114 including test data to dhtest.
55 *) Update fips_test_suite to support multiple command line options. New
56 test to induce all self test errors in sequence and check expected
60 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
61 sign or verify all in one operation.
64 *) Add fips_algvs: a multicall fips utility incorporaing all the algorithm
65 test programs and fips_test_suite. Includes functionality to parse
66 the minimal script output of fipsalgest.pl directly.
69 *) Add authorisation parameter to FIPS_module_mode_set().
72 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
75 *) Use separate DRBG fields for internal and external flags. New function
76 FIPS_drbg_health_check() to perform on demand health checking. Add
77 generation tests to fips_test_suite with reduced health check interval to
78 demonstrate periodic health checking. Add "nodh" option to
79 fips_test_suite to skip very slow DH test.
82 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
86 *) More extensive health check for DRBG checking many more failure modes.
87 New function FIPS_selftest_drbg_all() to handle every possible DRBG
88 combination: call this in fips_test_suite.
91 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
92 and POST to handle Dual EC cases.
95 *) Add support for canonical generation of DSA parameter 'g'. See
98 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
99 POST to handle HMAC cases.
102 *) Add functions FIPS_module_version() and FIPS_module_version_text()
103 to return numberical and string versions of the FIPS module number.
106 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
107 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implmeneted
108 outside the validated module in the FIPS capable OpenSSL.
111 *) Minor change to DRBG entropy callback semantics. In some cases
112 there is no mutiple of the block length between min_len and
113 max_len. Allow the callback to return more than max_len bytes
114 of entropy but discard any extra: it is the callback's responsibility
115 to ensure that the extra data discarded does not impact the
116 requested amount of entropy.
119 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
120 information in FIPS186-3, SP800-57 and SP800-131A.
123 *) CCM support via EVP. Interface is very similar to GCM case except we
124 must supply all data in one chunk (i.e. no update, final) and the
125 message length must be supplied if AAD is used. Add algorithm test
129 *) Initial version of POST overhaul. Add POST callback to allow the status
130 of POST to be monitored and/or failures induced. Modify fips_test_suite
131 to use callback. Always run all selftests even if one fails.
134 *) XTS support including algorithm test driver in the fips_gcmtest program.
135 Note: this does increase the maximum key length from 32 to 64 bytes but
136 there should be no binary compatibility issues as existing applications
137 will never use XTS mode.
140 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
141 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
142 performs algorithm blocking for unapproved PRNG types. Also do not
143 set PRNG type in FIPS_mode_set(): leave this to the application.
144 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
145 the standard OpenSSL PRNG: set additional data to a date time vector.
148 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
149 This shouldn't present any incompatibility problems because applications
150 shouldn't be using these directly and any that are will need to rethink
151 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
154 *) Extensive self tests and health checking required by SP800-90 DRBG.
155 Remove strength parameter from FIPS_drbg_instantiate and always
156 instantiate at maximum supported strength.
159 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
162 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
165 *) New function DH_compute_key_padded() to compute a DH key and pad with
166 leading zeroes if needed: this complies with SP800-56A et al.
169 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
170 anything, incomplete, subject to change and largely untested at present.
173 *) Modify fipscanisteronly build option to only build the necessary object
174 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
177 *) Add experimental option FIPSSYMS to give all symbols in
178 fipscanister.o and FIPS or fips prefix. This will avoid
179 conflicts with future versions of OpenSSL. Add perl script
180 util/fipsas.pl to preprocess assembly language source files
181 and rename any affected symbols.
184 *) Add selftest checks and algorithm block of non-fips algorithms in
185 FIPS mode. Remove DES2 from selftests.
188 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
189 return internal method without any ENGINE dependencies. Add new
190 tiny fips sign and verify functions.
193 *) New build option no-ec2m to disable characteristic 2 code.
196 *) New build option "fipscanisteronly". This only builds fipscanister.o
197 and (currently) associated fips utilities. Uses the file Makefile.fips
198 instead of Makefile.org as the prototype.
201 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
202 Update fips_gcmtest to use IV generator.
205 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
206 setting output buffer to NULL. The *Final function must be
207 called although it will not retrieve any additional data. The tag
208 can be set or retrieved with a ctrl. The IV length is by default 12
209 bytes (96 bits) but can be set to an alternative value. If the IV
210 length exceeds the maximum IV length (currently 16 bytes) it cannot be
214 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
215 underlying do_cipher function handles all cipher semantics itself
216 including padding and finalisation. This is useful if (for example)
217 an ENGINE cipher handles block padding itself. The behaviour of
218 do_cipher is subtly changed if this flag is set: the return value
219 is the number of characters written to the output buffer (zero is
220 no longer an error code) or a negative error code. Also if the
221 input buffer is NULL and length 0 finalisation should be performed.
224 *) If a candidate issuer certificate is already part of the constructed
225 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
228 *) Improve forward-security support: add functions
230 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
231 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
233 for use by SSL/TLS servers; the callback function will be called whenever a
234 new session is created, and gets to decide whether the session may be
235 cached to make it resumable (return 0) or not (return 1). (As by the
236 SSL/TLS protocol specifications, the session_id sent by the server will be
237 empty to indicate that the session is not resumable; also, the server will
238 not generate RFC 4507 (RFC 5077) session tickets.)
240 A simple reasonable callback implementation is to return is_forward_secure.
241 This parameter will be set to 1 or 0 depending on the ciphersuite selected
242 by the SSL/TLS server library, indicating whether it can provide forward
244 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
246 *) New function OPENSSL_gmtime_diff to find the difference in days
247 and seconds between two tm structures. This will be used to provide
248 additional functionality for ASN1_TIME.
251 *) Add -trusted_first option which attempts to find certificates in the
252 trusted store even if an untrusted chain is also supplied.
255 *) Initial experimental support for explicitly trusted non-root CAs.
256 OpenSSL still tries to build a complete chain to a root but if an
257 intermediate CA has a trust setting included that is used. The first
258 setting is used: whether to trust or reject.
261 *) New -verify_name option in command line utilities to set verification
265 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
266 Add CMAC pkey methods.
269 *) Experimental regnegotiation in s_server -www mode. If the client
270 browses /reneg connection is renegotiated. If /renegcert it is
271 renegotiated requesting a certificate.
274 *) Add an "external" session cache for debugging purposes to s_server. This
275 should help trace issues which normally are only apparent in deployed
276 multi-process servers.
279 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
280 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
281 BIO_set_cipher() and some obscure PEM functions were changed so they
282 can now return an error. The RAND changes required a change to the
283 RAND_METHOD structure.
286 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
287 a gcc attribute to warn if the result of a function is ignored. This
288 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
289 whose return value is often ignored.
292 Changes between 1.0.1c and 1.0.1d [xx XXX xxxx]
294 *) Don't use TLS 1.0 record version number in initial client hello
298 Changes between 1.0.1b and 1.0.1c [10 May 2012]
300 *) Sanity check record length before skipping explicit IV in TLS
301 1.2, 1.1 and DTLS to avoid DoS attack.
303 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
304 fuzzing as a service testing platform.
308 *) Initialise tkeylen properly when encrypting CMS messages.
309 Thanks to Solar Designer of Openwall for reporting this issue.
312 *) In FIPS mode don't try to use composite ciphers as they are not
316 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
318 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
319 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
320 mean any application compiled against OpenSSL 1.0.0 headers setting
321 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
322 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
323 0x10000000L Any application which was previously compiled against
324 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
325 will need to be recompiled as a result. Letting be results in
326 inability to disable specifically TLS 1.1 and in client context,
327 in unlike event, limit maximum offered version to TLS 1.0 [see below].
330 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
331 disable just protocol X, but all protocols above X *if* there are
332 protocols *below* X still enabled. In more practical terms it means
333 that if application wants to disable TLS1.0 in favor of TLS1.1 and
334 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
335 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
339 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
341 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
342 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
343 in CRYPTO_realloc_clean.
345 Thanks to Tavis Ormandy, Google Security Team, for discovering this
346 issue and to Adam Langley <agl@chromium.org> for fixing it.
348 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
350 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
353 *) Workarounds for some broken servers that "hang" if a client hello
354 record length exceeds 255 bytes:
356 1. Do not use record version number > TLS 1.0 in initial client
357 hello: some (but not all) hanging servers will now work.
358 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
359 the number of ciphers sent in the client hello. This should be
360 set to an even number, such as 50, for example by passing:
361 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
362 Most broken servers should now work.
363 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
364 TLS 1.2 client support entirely.
367 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
370 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
372 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
373 STRING form instead of a DigestInfo.
376 *) The format used for MDC2 RSA signatures is inconsistent between EVP
377 and the RSA_sign/RSA_verify functions. This was made more apparent when
378 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
379 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
380 the correct format in RSA_verify so both forms transparently work.
383 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
384 STRING form instead of a DigestInfo.
387 *) Some servers which support TLS 1.0 can choke if we initially indicate
388 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
389 encrypted premaster secret. As a workaround use the maximum pemitted
390 client version in client hello, this should keep such servers happy
391 and still work with previous versions of OpenSSL.
394 *) Add support for TLS/DTLS heartbeats.
395 [Robin Seggelmann <seggelmann@fh-muenster.de>]
397 *) Add support for SCTP.
398 [Robin Seggelmann <seggelmann@fh-muenster.de>]
400 *) Improved PRNG seeding for VOS.
401 [Paul Green <Paul.Green@stratus.com>]
403 *) Extensive assembler packs updates, most notably:
405 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
406 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
407 - x86_64: bit-sliced AES implementation;
408 - ARM: NEON support, contemporary platforms optimizations;
409 - s390x: z196 support;
410 - *: GHASH and GF(2^m) multiplication implementations;
414 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
415 (removal of unnecessary code)
416 [Peter Sylvester <peter.sylvester@edelweb.fr>]
418 *) Add TLS key material exporter from RFC 5705.
421 *) Add DTLS-SRTP negotiation from RFC 5764.
424 *) Add Next Protocol Negotiation,
425 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
426 disabled with a no-npn flag to config or Configure. Code donated
428 [Adam Langley <agl@google.com> and Ben Laurie]
430 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
431 NIST-P256, NIST-P521, with constant-time single point multiplication on
432 typical inputs. Compiler support for the nonstandard type __uint128_t is
433 required to use this (present in gcc 4.4 and later, for 64-bit builds).
434 Code made available under Apache License version 2.0.
436 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
437 line to include this in your build of OpenSSL, and run "make depend" (or
438 "make update"). This enables the following EC_METHODs:
440 EC_GFp_nistp224_method()
441 EC_GFp_nistp256_method()
442 EC_GFp_nistp521_method()
444 EC_GROUP_new_by_curve_name() will automatically use these (while
445 EC_GROUP_new_curve_GFp() currently prefers the more flexible
447 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
449 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
450 all platforms. Move ssize_t definition from e_os.h to the public
451 header file e_os2.h as it now appears in public header file cms.h
454 *) New -sigopt option to the ca, req and x509 utilities. Additional
455 signature parameters can be passed using this option and in
459 *) Add RSA PSS signing function. This will generate and set the
460 appropriate AlgorithmIdentifiers for PSS based on those in the
461 corresponding EVP_MD_CTX structure. No application support yet.
464 *) Support for companion algorithm specific ASN1 signing routines.
465 New function ASN1_item_sign_ctx() signs a pre-initialised
466 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
467 the appropriate parameters.
470 *) Add new algorithm specific ASN1 verification initialisation function
471 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
472 handling will be the same no matter what EVP_PKEY_METHOD is used.
473 Add a PSS handler to support verification of PSS signatures: checked
474 against a number of sample certificates.
477 *) Add signature printing for PSS. Add PSS OIDs.
478 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
480 *) Add algorithm specific signature printing. An individual ASN1 method
481 can now print out signatures instead of the standard hex dump.
483 More complex signatures (e.g. PSS) can print out more meaningful
484 information. Include DSA version that prints out the signature
488 *) Password based recipient info support for CMS library: implementing
492 *) Split password based encryption into PBES2 and PBKDF2 functions. This
493 neatly separates the code into cipher and PBE sections and is required
494 for some algorithms that split PBES2 into separate pieces (such as
498 *) Session-handling fixes:
499 - Fix handling of connections that are resuming with a session ID,
500 but also support Session Tickets.
501 - Fix a bug that suppressed issuing of a new ticket if the client
502 presented a ticket with an expired session.
503 - Try to set the ticket lifetime hint to something reasonable.
504 - Make tickets shorter by excluding irrelevant information.
505 - On the client side, don't ignore renewed tickets.
506 [Adam Langley, Bodo Moeller (Google)]
508 *) Fix PSK session representation.
511 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
513 This work was sponsored by Intel.
516 *) Add GCM support to TLS library. Some custom code is needed to split
517 the IV between the fixed (from PRF) and explicit (from TLS record)
518 portions. This adds all GCM ciphersuites supported by RFC5288 and
519 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
520 add a special AESGCM string for GCM only.
523 *) Expand range of ctrls for AES GCM. Permit setting invocation
524 field on decrypt and retrieval of invocation field only on encrypt.
527 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
528 As required by RFC5289 these ciphersuites cannot be used if for
529 versions of TLS earlier than 1.2.
532 *) For FIPS capable OpenSSL interpret a NULL default public key method
533 as unset and return the appopriate default but do *not* set the default.
534 This means we can return the appopriate method in applications that
535 swicth between FIPS and non-FIPS modes.
538 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
539 ENGINE is used then we cannot handle that in the FIPS module so we
540 keep original code iff non-FIPS operations are allowed.
543 *) Add -attime option to openssl utilities.
544 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
546 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
549 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
550 FIPS EC methods unconditionally for now.
553 *) New build option no-ec2m to disable characteristic 2 code.
556 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
557 all cases can be covered as some introduce binary incompatibilities.
560 *) Redirect RSA operations to FIPS module including keygen,
561 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
564 *) Add similar low level API blocking to ciphers.
567 *) Low level digest APIs are not approved in FIPS mode: any attempt
568 to use these will cause a fatal error. Applications that *really* want
569 to use them can use the private_* version instead.
572 *) Redirect cipher operations to FIPS module for FIPS builds.
575 *) Redirect digest operations to FIPS module for FIPS builds.
578 *) Update build system to add "fips" flag which will link in fipscanister.o
579 for static and shared library builds embedding a signature if needed.
582 *) Output TLS supported curves in preference order instead of numerical
583 order. This is currently hardcoded for the highest order curves first.
584 This should be configurable so applications can judge speed vs strength.
587 *) Add TLS v1.2 server support for client authentication.
590 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
594 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
595 FIPS modules versions.
598 *) Add TLS v1.2 client side support for client authentication. Keep cache
599 of handshake records longer as we don't know the hash algorithm to use
600 until after the certificate request message is received.
603 *) Initial TLS v1.2 client support. Add a default signature algorithms
604 extension including all the algorithms we support. Parse new signature
605 format in client key exchange. Relax some ECC signing restrictions for
606 TLS v1.2 as indicated in RFC5246.
609 *) Add server support for TLS v1.2 signature algorithms extension. Switch
610 to new signature format when needed using client digest preference.
611 All server ciphersuites should now work correctly in TLS v1.2. No client
612 support yet and no support for client certificates.
615 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
616 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
617 ciphersuites. At present only RSA key exchange ciphersuites work with
618 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
619 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
620 and version checking.
623 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
624 with this defined it will not be affected by any changes to ssl internal
625 structures. Add several utility functions to allow openssl application
626 to work with OPENSSL_NO_SSL_INTERN defined.
630 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
632 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
635 *) Permit abbreviated handshakes when renegotiating using the function
636 SSL_renegotiate_abbreviated().
637 [Robin Seggelmann <seggelmann@fh-muenster.de>]
639 *) Add call to ENGINE_register_all_complete() to
640 ENGINE_load_builtin_engines(), so some implementations get used
641 automatically instead of needing explicit application support.
644 *) Add support for TLS key exporter as described in RFC5705.
645 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
647 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
648 a few changes are required:
650 Add SSL_OP_NO_TLSv1_1 flag.
652 Update version checking logic to handle version 1.1.
653 Add explicit IV handling (ported from DTLS code).
654 Add command line options to s_client/s_server.
657 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
659 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
660 in CMS and PKCS7 code. When RSA decryption fails use a random key for
661 content decryption and always return the same error. Note: this attack
662 needs on average 2^20 messages so it only affects automated senders. The
663 old behaviour can be reenabled in the CMS code by setting the
664 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
665 an MMA defence is not necessary.
666 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
667 this issue. (CVE-2012-0884)
670 *) Fix CVE-2011-4619: make sure we really are receiving a
671 client hello before rejecting multiple SGC restarts. Thanks to
672 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
675 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
677 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
678 Thanks to Antonio Martin, Enterprise Secure Access Research and
679 Development, Cisco Systems, Inc. for discovering this bug and
680 preparing a fix. (CVE-2012-0050)
683 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
685 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
686 of the Vaudenay padding oracle attack on CBC mode encryption
687 which enables an efficient plaintext recovery attack against
688 the OpenSSL implementation of DTLS. Their attack exploits timing
689 differences arising during decryption processing. A research
690 paper describing this attack can be found at:
691 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
692 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
693 Security Group at Royal Holloway, University of London
694 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
695 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
696 for preparing the fix. (CVE-2011-4108)
697 [Robin Seggelmann, Michael Tuexen]
699 *) Clear bytes used for block padding of SSL 3.0 records.
701 [Adam Langley (Google)]
703 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
704 Kadianakis <desnacked@gmail.com> for discovering this issue and
705 Adam Langley for preparing the fix. (CVE-2011-4619)
706 [Adam Langley (Google)]
708 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
709 [Andrey Kulikov <amdeich@gmail.com>]
711 *) Prevent malformed RFC3779 data triggering an assertion failure.
712 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
713 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
714 [Rob Austein <sra@hactrn.net>]
716 *) Improved PRNG seeding for VOS.
717 [Paul Green <Paul.Green@stratus.com>]
719 *) Fix ssl_ciph.c set-up race.
720 [Adam Langley (Google)]
722 *) Fix spurious failures in ecdsatest.c.
723 [Emilia Käsper (Google)]
725 *) Fix the BIO_f_buffer() implementation (which was mixing different
726 interpretations of the '..._len' fields).
727 [Adam Langley (Google)]
729 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
730 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
731 threads won't reuse the same blinding coefficients.
733 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
734 lock to call BN_BLINDING_invert_ex, and avoids one use of
735 BN_BLINDING_update for each BN_BLINDING structure (previously,
736 the last update always remained unused).
737 [Emilia Käsper (Google)]
739 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
740 [Bob Buckholz (Google)]
742 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
744 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
745 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
746 [Kaspar Brand <ossl@velox.ch>]
748 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
749 for multi-threaded use of ECDH. (CVE-2011-3210)
750 [Adam Langley (Google)]
752 *) Fix x509_name_ex_d2i memory leak on bad inputs.
755 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
756 signature public key algorithm by using OID xref utilities instead.
757 Before this you could only use some ECC ciphersuites with SHA1 only.
760 *) Add protection against ECDSA timing attacks as mentioned in the paper
761 by Billy Bob Brumley and Nicola Tuveri, see:
763 http://eprint.iacr.org/2011/232.pdf
765 [Billy Bob Brumley and Nicola Tuveri]
767 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
769 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
770 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
772 *) Fix bug in string printing code: if *any* escaping is enabled we must
773 escape the escape character (backslash) or the resulting string is
777 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
779 *) Disable code workaround for ancient and obsolete Netscape browsers
780 and servers: an attacker can use it in a ciphersuite downgrade attack.
781 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
784 *) Fixed J-PAKE implementation error, originally discovered by
785 Sebastien Martini, further info and confirmation from Stefan
786 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
789 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
791 *) Fix extension code to avoid race conditions which can result in a buffer
792 overrun vulnerability: resumed sessions must not be modified as they can
793 be shared by multiple threads. CVE-2010-3864
796 *) Fix WIN32 build system to correctly link an ENGINE directory into
800 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
802 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
804 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
806 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
808 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
809 context. The operation can be customised via the ctrl mechanism in
810 case ENGINEs want to include additional functionality.
813 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
816 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
817 output hashes compatible with older versions of OpenSSL.
818 [Willy Weisz <weisz@vcpc.univie.ac.at>]
820 *) Fix compression algorithm handling: if resuming a session use the
821 compression algorithm of the resumed session instead of determining
822 it from client hello again. Don't allow server to change algorithm.
825 *) Add load_crls() function to apps tidying load_certs() too. Add option
826 to verify utility to allow additional CRLs to be included.
829 *) Update OCSP request code to permit adding custom headers to the request:
830 some responders need this.
833 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
835 [Julia Lawall <julia@diku.dk>]
837 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
838 needlessly dereferenced structures, used obsolete functions and
839 didn't handle all updated verify codes correctly.
842 *) Disable MD2 in the default configuration.
845 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
846 indicate the initial BIO being pushed or popped. This makes it possible
847 to determine whether the BIO is the one explicitly called or as a result
848 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
849 it handles reference counts correctly and doesn't zero out the I/O bio
850 when it is not being explicitly popped. WARNING: applications which
851 included workarounds for the old buggy behaviour will need to be modified
852 or they could free up already freed BIOs.
855 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
856 renaming to all platforms (within the 0.9.8 branch, this was
857 done conditionally on Netware platforms to avoid a name clash).
858 [Guenter <lists@gknw.net>]
860 *) Add ECDHE and PSK support to DTLS.
861 [Michael Tuexen <tuexen@fh-muenster.de>]
863 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
867 *) Add "missing" function EVP_MD_flags() (without this the only way to
868 retrieve a digest flags is by accessing the structure directly. Update
869 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
870 or cipher is registered as in the "from" argument. Print out all
871 registered digests in the dgst usage message instead of manually
872 attempting to work them out.
875 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
876 this allows the use of compression and extensions. Change default cipher
877 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
878 by default unless an application cipher string requests it.
881 *) Alter match criteria in PKCS12_parse(). It used to try to use local
882 key ids to find matching certificates and keys but some PKCS#12 files
883 don't follow the (somewhat unwritten) rules and this strategy fails.
884 Now just gather all certificates together and the first private key
885 then look for the first certificate that matches the key.
888 *) Support use of registered digest and cipher names for dgst and cipher
889 commands instead of having to add each one as a special case. So now
896 openssl dgst -sha256 foo
898 and this works for ENGINE based algorithms too.
902 *) Update Gost ENGINE to support parameter files.
903 [Victor B. Wagner <vitus@cryptocom.ru>]
905 *) Support GeneralizedTime in ca utility.
906 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
908 *) Enhance the hash format used for certificate directory links. The new
909 form uses the canonical encoding (meaning equivalent names will work
910 even if they aren't identical) and uses SHA1 instead of MD5. This form
911 is incompatible with the older format and as a result c_rehash should
912 be used to rebuild symbolic links.
915 *) Make PKCS#8 the default write format for private keys, replacing the
916 traditional format. This form is standardised, more secure and doesn't
917 include an implicit MD5 dependency.
920 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
921 committed to OpenSSL should pass this lot as a minimum.
924 *) Add session ticket override functionality for use by EAP-FAST.
925 [Jouni Malinen <j@w1.fi>]
927 *) Modify HMAC functions to return a value. Since these can be implemented
928 in an ENGINE errors can occur.
931 *) Type-checked OBJ_bsearch_ex.
934 *) Type-checked OBJ_bsearch. Also some constification necessitated
935 by type-checking. Still to come: TXT_DB, bsearch(?),
936 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
940 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
941 seconds to a tm structure directly, instead of going through OS
942 specific date routines. This avoids any issues with OS routines such
943 as the year 2038 bug. New *_adj() functions for ASN1 time structures
944 and X509_time_adj_ex() to cover the extended range. The existing
945 X509_time_adj() is still usable and will no longer have any date issues.
948 *) Delta CRL support. New use deltas option which will attempt to locate
949 and search any appropriate delta CRLs available.
951 This work was sponsored by Google.
954 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
955 code and add additional score elements. Validate alternate CRL paths
956 as part of the CRL checking and indicate a new error "CRL path validation
957 error" in this case. Applications wanting additional details can use
958 the verify callback and check the new "parent" field. If this is not
959 NULL CRL path validation is taking place. Existing applications wont
960 see this because it requires extended CRL support which is off by
963 This work was sponsored by Google.
966 *) Support for freshest CRL extension.
968 This work was sponsored by Google.
971 *) Initial indirect CRL support. Currently only supported in the CRLs
972 passed directly and not via lookup. Process certificate issuer
973 CRL entry extension and lookup CRL entries by bother issuer name
974 and serial number. Check and process CRL issuer entry in IDP extension.
976 This work was sponsored by Google.
979 *) Add support for distinct certificate and CRL paths. The CRL issuer
980 certificate is validated separately in this case. Only enabled if
981 an extended CRL support flag is set: this flag will enable additional
982 CRL functionality in future.
984 This work was sponsored by Google.
987 *) Add support for policy mappings extension.
989 This work was sponsored by Google.
992 *) Fixes to pathlength constraint, self issued certificate handling,
993 policy processing to align with RFC3280 and PKITS tests.
995 This work was sponsored by Google.
998 *) Support for name constraints certificate extension. DN, email, DNS
999 and URI types are currently supported.
1001 This work was sponsored by Google.
1004 *) To cater for systems that provide a pointer-based thread ID rather
1005 than numeric, deprecate the current numeric thread ID mechanism and
1006 replace it with a structure and associated callback type. This
1007 mechanism allows a numeric "hash" to be extracted from a thread ID in
1008 either case, and on platforms where pointers are larger than 'long',
1009 mixing is done to help ensure the numeric 'hash' is usable even if it
1010 can't be guaranteed unique. The default mechanism is to use "&errno"
1011 as a pointer-based thread ID to distinguish between threads.
1013 Applications that want to provide their own thread IDs should now use
1014 CRYPTO_THREADID_set_callback() to register a callback that will call
1015 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
1017 Note that ERR_remove_state() is now deprecated, because it is tied
1018 to the assumption that thread IDs are numeric. ERR_remove_state(0)
1019 to free the current thread's error state should be replaced by
1020 ERR_remove_thread_state(NULL).
1022 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
1023 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
1024 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
1025 application was previously providing a numeric thread callback that
1026 was inappropriate for distinguishing threads, then uniqueness might
1027 have been obtained with &errno that happened immediately in the
1028 intermediate development versions of OpenSSL; this is no longer the
1029 case, the numeric thread callback will now override the automatic use
1031 [Geoff Thorpe, with help from Bodo Moeller]
1033 *) Initial support for different CRL issuing certificates. This covers a
1034 simple case where the self issued certificates in the chain exist and
1035 the real CRL issuer is higher in the existing chain.
1037 This work was sponsored by Google.
1040 *) Removed effectively defunct crypto/store from the build.
1043 *) Revamp of STACK to provide stronger type-checking. Still to come:
1044 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
1045 ASN1_STRING, CONF_VALUE.
1048 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
1049 RAM on SSL connections. This option can save about 34k per idle SSL.
1052 *) Revamp of LHASH to provide stronger type-checking. Still to come:
1053 STACK, TXT_DB, bsearch, qsort.
1056 *) Initial support for Cryptographic Message Syntax (aka CMS) based
1057 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
1058 support for data, signedData, compressedData, digestedData and
1059 encryptedData, envelopedData types included. Scripts to check against
1060 RFC4134 examples draft and interop and consistency checks of many
1061 content types and variants.
1064 *) Add options to enc utility to support use of zlib compression BIO.
1067 *) Extend mk1mf to support importing of options and assembly language
1068 files from Configure script, currently only included in VC-WIN32.
1069 The assembly language rules can now optionally generate the source
1070 files from the associated perl scripts.
1073 *) Implement remaining functionality needed to support GOST ciphersuites.
1074 Interop testing has been performed using CryptoPro implementations.
1075 [Victor B. Wagner <vitus@cryptocom.ru>]
1077 *) s390x assembler pack.
1080 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
1084 *) Implement Opaque PRF Input TLS extension as specified in
1085 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
1086 official specification yet and no extension type assignment by
1087 IANA exists, this extension (for now) will have to be explicitly
1088 enabled when building OpenSSL by providing the extension number
1089 to use. For example, specify an option
1091 -DTLSEXT_TYPE_opaque_prf_input=0x9527
1093 to the "config" or "Configure" script to enable the extension,
1094 assuming extension number 0x9527 (which is a completely arbitrary
1095 and unofficial assignment based on the MD5 hash of the Internet
1096 Draft). Note that by doing so, you potentially lose
1097 interoperability with other TLS implementations since these might
1098 be using the same extension number for other purposes.
1100 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
1101 opaque PRF input value to use in the handshake. This will create
1102 an interal copy of the length-'len' string at 'src', and will
1103 return non-zero for success.
1105 To get more control and flexibility, provide a callback function
1108 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
1109 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
1113 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
1116 Callback function 'cb' will be called in handshakes, and is
1117 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
1118 Argument 'arg' is for application purposes (the value as given to
1119 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
1120 be provided to the callback function). The callback function
1121 has to return non-zero to report success: usually 1 to use opaque
1122 PRF input just if possible, or 2 to enforce use of the opaque PRF
1123 input. In the latter case, the library will abort the handshake
1124 if opaque PRF input is not successfully negotiated.
1126 Arguments 'peerinput' and 'len' given to the callback function
1127 will always be NULL and 0 in the case of a client. A server will
1128 see the client's opaque PRF input through these variables if
1129 available (NULL and 0 otherwise). Note that if the server
1130 provides an opaque PRF input, the length must be the same as the
1131 length of the client's opaque PRF input.
1133 Note that the callback function will only be called when creating
1134 a new session (session resumption can resume whatever was
1135 previously negotiated), and will not be called in SSL 2.0
1136 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
1137 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
1138 for applications that need to enforce opaque PRF input.
1142 *) Update ssl code to support digests other than SHA1+MD5 for handshake
1145 [Victor B. Wagner <vitus@cryptocom.ru>]
1147 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1148 RFC4507bis. The encrypted ticket format is an encrypted encoded
1149 SSL_SESSION structure, that way new session features are automatically
1152 If a client application caches session in an SSL_SESSION structure
1153 support is transparent because tickets are now stored in the encoded
1156 The SSL_CTX structure automatically generates keys for ticket
1157 protection in servers so again support should be possible
1158 with no application modification.
1160 If a client or server wishes to disable RFC4507 support then the option
1161 SSL_OP_NO_TICKET can be set.
1163 Add a TLS extension debugging callback to allow the contents of any client
1164 or server extensions to be examined.
1166 This work was sponsored by Google.
1169 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1170 OpenSSL should now compile cleanly on gcc 4.2
1171 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1173 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1174 support including streaming MAC support: this is required for GOST
1175 ciphersuite support.
1176 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1178 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1179 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1180 to output in BER and PEM format.
1183 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1184 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1185 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1186 ENGINE support for HMAC keys which are unextractable. New -mac and
1187 -macopt options to dgst utility.
1190 *) New option -sigopt to dgst utility. Update dgst to use
1191 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1192 alternative signing paramaters such as X9.31 or PSS in the dgst
1196 *) Change ssl_cipher_apply_rule(), the internal function that does
1197 the work each time a ciphersuite string requests enabling
1198 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1199 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1200 the order of disabled ciphersuites such that those ciphersuites
1201 that most recently went from enabled to disabled not only stay
1202 in order with respect to each other, but also have higher priority
1203 than other disabled ciphersuites the next time ciphersuites are
1206 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1207 the same ciphersuites as with "HIGH" alone, but in a specific
1208 order where the PSK ciphersuites come first (since they are the
1209 most recently disabled ciphersuites when "HIGH" is parsed).
1211 Also, change ssl_create_cipher_list() (using this new
1212 funcionality) such that between otherwise identical
1213 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
1217 *) Change ssl_create_cipher_list() so that it automatically
1218 arranges the ciphersuites in reasonable order before starting
1219 to process the rule string. Thus, the definition for "DEFAULT"
1220 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
1221 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
1222 This makes it much easier to arrive at a reasonable default order
1223 in applications for which anonymous ciphers are OK (meaning
1224 that you can't actually use DEFAULT).
1225 [Bodo Moeller; suggested by Victor Duchovni]
1227 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
1228 processing) into multiple integers instead of setting
1229 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
1230 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
1231 (These masks as well as the individual bit definitions are hidden
1232 away into the non-exported interface ssl/ssl_locl.h, so this
1233 change to the definition of the SSL_CIPHER structure shouldn't
1234 affect applications.) This give us more bits for each of these
1235 categories, so there is no longer a need to coagulate AES128 and
1236 AES256 into a single algorithm bit, and to coagulate Camellia128
1237 and Camellia256 into a single algorithm bit, which has led to all
1240 Thus, among other things, the kludge introduced in 0.9.7m and
1241 0.9.8e for masking out AES256 independently of AES128 or masking
1242 out Camellia256 independently of AES256 is not needed here in 0.9.9.
1244 With the change, we also introduce new ciphersuite aliases that
1245 so far were missing: "AES128", "AES256", "CAMELLIA128", and
1249 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
1250 Use the leftmost N bytes of the signature input if the input is
1251 larger than the prime q (with N being the size in bytes of q).
1254 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1255 it yet and it is largely untested.
1258 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1261 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1262 some compilers (gcc 4.2 and later) reject their use. Safestack is
1263 reimplemented. Update ASN1 to avoid use of legacy functions.
1266 *) Win32/64 targets are linked with Winsock2.
1269 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1270 to external functions. This can be used to increase CRL handling
1271 efficiency especially when CRLs are very large by (for example) storing
1272 the CRL revoked certificates in a database.
1275 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1276 new CRLs added to a directory can be used. New command line option
1277 -verify_return_error to s_client and s_server. This causes real errors
1278 to be returned by the verify callback instead of carrying on no matter
1279 what. This reflects the way a "real world" verify callback would behave.
1282 *) GOST engine, supporting several GOST algorithms and public key formats.
1283 Kindly donated by Cryptocom.
1286 *) Partial support for Issuing Distribution Point CRL extension. CRLs
1287 partitioned by DP are handled but no indirect CRL or reason partitioning
1288 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
1289 selected via a scoring technique which handles IDP and AKID in CRLs.
1292 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1293 will ultimately be used for all verify operations: this will remove the
1294 X509_STORE dependency on certificate verification and allow alternative
1295 lookup methods. X509_STORE based implementations of these two callbacks.
1298 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1299 Modify get_crl() to find a valid (unexpired) CRL if possible.
1302 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1303 this would be called X509_CRL_cmp() but that name is already used by
1304 a function that just compares CRL issuer names. Cache several CRL
1305 extensions in X509_CRL structure and cache CRLDP in X509.
1308 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1309 this maps equivalent X509_NAME structures into a consistent structure.
1310 Name comparison can then be performed rapidly using memcmp().
1313 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1317 *) Allow digests to supply their own micalg string for S/MIME type using
1318 the ctrl EVP_MD_CTRL_MICALG.
1321 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1322 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1323 ctrl. It can then customise the structure before and/or after signing
1327 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1328 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1329 to free up any added signature OIDs.
1332 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1333 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1334 digest and cipher tables. New options added to openssl utility:
1335 list-message-digest-algorithms and list-cipher-algorithms.
1338 *) Change the array representation of binary polynomials: the list
1339 of degrees of non-zero coefficients is now terminated with -1.
1340 Previously it was terminated with 0, which was also part of the
1341 value; thus, the array representation was not applicable to
1342 polynomials where t^0 has coefficient zero. This change makes
1343 the array representation useful in a more general context.
1346 *) Various modifications and fixes to SSL/TLS cipher string
1347 handling. For ECC, the code now distinguishes between fixed ECDH
1348 with RSA certificates on the one hand and with ECDSA certificates
1349 on the other hand, since these are separate ciphersuites. The
1350 unused code for Fortezza ciphersuites has been removed.
1352 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1353 (not "ECDHE"). For consistency with the code for DH
1354 certificates, use of ECDH certificates is now considered ECDH
1355 authentication, not RSA or ECDSA authentication (the latter is
1356 merely the CA's signing algorithm and not actively used in the
1359 The temporary ciphersuite alias "ECCdraft" is no longer
1360 available, and ECC ciphersuites are no longer excluded from "ALL"
1361 and "DEFAULT". The following aliases now exist for RFC 4492
1362 ciphersuites, most of these by analogy with the DH case:
1364 kECDHr - ECDH cert, signed with RSA
1365 kECDHe - ECDH cert, signed with ECDSA
1366 kECDH - ECDH cert (signed with either RSA or ECDSA)
1367 kEECDH - ephemeral ECDH
1368 ECDH - ECDH cert or ephemeral ECDH
1374 AECDH - anonymous ECDH
1375 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1379 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1380 Use correct micalg parameters depending on digest(s) in signed message.
1383 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1384 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1387 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1388 an engine to register a method. Add ENGINE lookups for methods and
1389 functional reference processing.
1392 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1393 EVP_{Sign,Verify}* which allow an application to customise the signature
1397 *) New -resign option to smime utility. This adds one or more signers
1398 to an existing PKCS#7 signedData structure. Also -md option to use an
1399 alternative message digest algorithm for signing.
1402 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1403 create PKCS7 structures containing multiple signers. Update smime
1404 application to support multiple signers.
1407 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1411 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1412 Reorganize PBE internals to lookup from a static table using NIDs,
1413 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1414 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1415 PRF which will be automatically used with PBES2.
1418 *) Replace the algorithm specific calls to generate keys in "req" with the
1422 *) Update PKCS#7 enveloped data routines to use new API. This is now
1423 supported by any public key method supporting the encrypt operation. A
1424 ctrl is added to allow the public key algorithm to examine or modify
1425 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1429 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1430 a default digest type to use. In most cases this will be SHA1 but some
1431 algorithms (such as GOST) need to specify an alternative digest. The
1432 return value indicates how strong the prefernce is 1 means optional and
1433 2 is mandatory (that is it is the only supported type). Modify
1434 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1435 use the default md. Update openssl utilities to use the default digest
1436 type for signing if it is not explicitly indicated.
1439 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1440 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1441 signing method from the key type. This effectively removes the link
1442 between digests and public key types.
1445 *) Add an OID cross reference table and utility functions. Its purpose is to
1446 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1447 rsaEncryption. This will allow some of the algorithm specific hackery
1448 needed to use the correct OID to be removed.
1451 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1452 structures for PKCS7_sign(). They are now set up by the relevant public
1456 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1459 *) Add support for key derivation (agreement) in the API, DH method and
1463 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1464 public and private key formats. As a side effect these add additional
1465 command line functionality not previously available: DSA signatures can be
1466 generated and verified using pkeyutl and DH key support and generation in
1471 [Oliver Tappe <zooey@hirschkaefer.de>]
1473 *) New make target "install_html_docs" installs HTML renditions of the
1475 [Oliver Tappe <zooey@hirschkaefer.de>]
1477 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1478 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1479 support key and parameter generation and add initial key generation
1480 functionality for RSA.
1483 *) Add functions for main EVP_PKEY_method operations. The undocumented
1484 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1485 EVP_PKEY_{encrypt,decrypt}_old.
1488 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1489 key API, doesn't do much yet.
1492 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1493 public key algorithms. New option to openssl utility:
1494 "list-public-key-algorithms" to print out info.
1497 *) Implement the Supported Elliptic Curves Extension for
1498 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1501 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1502 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1505 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1506 utilities such as rsa, dsa, dsaparam etc except they process any key
1510 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1511 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1512 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1516 *) Initial support for pluggable public key ASN1.
1517 De-spaghettify the public key ASN1 handling. Move public and private
1518 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1519 algorithm specific handling to a single module within the relevant
1520 algorithm directory. Add functions to allow (near) opaque processing
1521 of public and private key structures.
1524 *) Implement the Supported Point Formats Extension for
1525 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1528 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1529 for the psk identity [hint] and the psk callback functions to the
1530 SSL_SESSION, SSL and SSL_CTX structure.
1533 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1537 SSL_CTX_use_psk_identity_hint
1538 SSL_get_psk_identity_hint
1539 SSL_get_psk_identity
1540 SSL_use_psk_identity_hint
1542 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1544 *) Add RFC 3161 compliant time stamp request creation, response generation
1545 and response verification functionality.
1546 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1548 *) Add initial support for TLS extensions, specifically for the server_name
1549 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1550 have new members for a host name. The SSL data structure has an
1551 additional member SSL_CTX *initial_ctx so that new sessions can be
1552 stored in that context to allow for session resumption, even after the
1553 SSL has been switched to a new SSL_CTX in reaction to a client's
1554 server_name extension.
1556 New functions (subject to change):
1558 SSL_get_servername()
1559 SSL_get_servername_type()
1562 New CTRL codes and macros (subject to change):
1564 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1565 - SSL_CTX_set_tlsext_servername_callback()
1566 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1567 - SSL_CTX_set_tlsext_servername_arg()
1568 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1570 openssl s_client has a new '-servername ...' option.
1572 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1573 '-key2 ...', '-servername_fatal' (subject to change). This allows
1574 testing the HostName extension for a specific single host name ('-cert'
1575 and '-key' remain fallbacks for handshakes without HostName
1576 negotiation). If the unrecogninzed_name alert has to be sent, this by
1577 default is a warning; it becomes fatal with the '-servername_fatal'
1580 [Peter Sylvester, Remy Allais, Christophe Renou]
1582 *) Whirlpool hash implementation is added.
1585 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1586 bn(64,32). Because of instruction set limitations it doesn't have
1587 any negative impact on performance. This was done mostly in order
1588 to make it possible to share assembler modules, such as bn_mul_mont
1589 implementations, between 32- and 64-bit builds without hassle.
1592 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1593 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1597 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1598 dedicated Montgomery multiplication procedure, is introduced.
1599 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1600 "64-bit" performance on certain 32-bit targets.
1603 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1604 in SSL structures. New SSL ctrl to set maximum send fragment size.
1605 Save memory by seeting the I/O buffer sizes dynamically instead of
1606 using the maximum available value.
1609 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1610 in addition to the text details.
1613 *) Very, very preliminary EXPERIMENTAL support for printing of general
1614 ASN1 structures. This currently produces rather ugly output and doesn't
1615 handle several customised structures at all.
1618 *) Integrated support for PVK file format and some related formats such
1619 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1620 these in the 'rsa' and 'dsa' utilities.
1623 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1626 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1627 place for the (very old) "NETSCAPE" format certificates which are now
1628 handled using new ASN1 code equivalents.
1631 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1632 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1633 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1636 *) Modify CRL distribution points extension code to print out previously
1637 unsupported fields. Enhance extension setting code to allow setting of
1641 *) Add print and set support for Issuing Distribution Point CRL extension.
1644 *) Change 'Configure' script to enable Camellia by default.
1647 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
1649 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1650 Thanks to Antonio Martin, Enterprise Secure Access Research and
1651 Development, Cisco Systems, Inc. for discovering this bug and
1652 preparing a fix. (CVE-2012-0050)
1655 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
1657 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1658 of the Vaudenay padding oracle attack on CBC mode encryption
1659 which enables an efficient plaintext recovery attack against
1660 the OpenSSL implementation of DTLS. Their attack exploits timing
1661 differences arising during decryption processing. A research
1662 paper describing this attack can be found at:
1663 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1664 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1665 Security Group at Royal Holloway, University of London
1666 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1667 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1668 for preparing the fix. (CVE-2011-4108)
1669 [Robin Seggelmann, Michael Tuexen]
1671 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
1672 [Ben Laurie, Kasper <ekasper@google.com>]
1674 *) Clear bytes used for block padding of SSL 3.0 records.
1676 [Adam Langley (Google)]
1678 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1679 Kadianakis <desnacked@gmail.com> for discovering this issue and
1680 Adam Langley for preparing the fix. (CVE-2011-4619)
1681 [Adam Langley (Google)]
1683 *) Prevent malformed RFC3779 data triggering an assertion failure.
1684 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1685 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1686 [Rob Austein <sra@hactrn.net>]
1688 *) Fix ssl_ciph.c set-up race.
1689 [Adam Langley (Google)]
1691 *) Fix spurious failures in ecdsatest.c.
1692 [Emilia Käsper (Google)]
1694 *) Fix the BIO_f_buffer() implementation (which was mixing different
1695 interpretations of the '..._len' fields).
1696 [Adam Langley (Google)]
1698 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1699 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1700 threads won't reuse the same blinding coefficients.
1702 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1703 lock to call BN_BLINDING_invert_ex, and avoids one use of
1704 BN_BLINDING_update for each BN_BLINDING structure (previously,
1705 the last update always remained unused).
1706 [Emilia Käsper (Google)]
1708 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1709 for multi-threaded use of ECDH.
1710 [Adam Langley (Google)]
1712 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1715 *) Add protection against ECDSA timing attacks as mentioned in the paper
1716 by Billy Bob Brumley and Nicola Tuveri, see:
1718 http://eprint.iacr.org/2011/232.pdf
1720 [Billy Bob Brumley and Nicola Tuveri]
1722 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
1724 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1725 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1727 *) Fix bug in string printing code: if *any* escaping is enabled we must
1728 escape the escape character (backslash) or the resulting string is
1732 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
1734 *) Disable code workaround for ancient and obsolete Netscape browsers
1735 and servers: an attacker can use it in a ciphersuite downgrade attack.
1736 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1739 *) Fixed J-PAKE implementation error, originally discovered by
1740 Sebastien Martini, further info and confirmation from Stefan
1741 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1744 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
1746 *) Fix extension code to avoid race conditions which can result in a buffer
1747 overrun vulnerability: resumed sessions must not be modified as they can
1748 be shared by multiple threads. CVE-2010-3864
1751 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
1754 *) Don't reencode certificate when calculating signature: cache and use
1755 the original encoding instead. This makes signature verification of
1756 some broken encodings work correctly.
1759 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
1760 is also one of the inputs.
1761 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1763 *) Don't repeatedly append PBE algorithms to table if they already exist.
1764 Sort table on each new add. This effectively makes the table read only
1765 after all algorithms are added and subsequent calls to PKCS12_pbe_add
1769 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
1771 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
1774 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
1775 access or freeing data twice (CVE-2010-0742)
1776 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
1778 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
1779 common in certificates and some applications which only call
1780 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
1784 Reduce copying into .apps and .test in makevms.com
1785 Don't try to use blank CA certificate in CA.com
1786 Allow use of C files from original directories in maketests.com
1787 [Steven M. Schweda" <sms@antinode.info>]
1789 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1791 *) When rejecting SSL/TLS records due to an incorrect version number, never
1792 update s->server with a new major version number. As of
1793 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1794 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1795 the previous behavior could result in a read attempt at NULL when
1796 receiving specific incorrect SSL/TLS records once record payload
1797 protection is active. (CVE-2010-0740)
1798 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1800 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1801 could be crashed if the relevant tables were not present (e.g. chrooted).
1802 [Tomas Hoger <thoger@redhat.com>]
1804 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1806 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1807 [Martin Olsson, Neel Mehta]
1809 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1810 accommodate for stack sorting, always a write lock!).
1813 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1814 excessive delays in the RAND_poll(): over a minute. As a workaround
1815 include a time check in the inner Heap32Next loop too.
1818 *) The code that handled flushing of data in SSL/TLS originally used the
1819 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1820 the problem outlined in PR#1949. The fix suggested there however can
1821 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1822 of Apache). So instead simplify the code to flush unconditionally.
1823 This should be fine since flushing with no data to flush is a no op.
1826 *) Handle TLS versions 2.0 and later properly and correctly use the
1827 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1828 off ancient servers have a habit of sticking around for a while...
1831 *) Modify compression code so it frees up structures without using the
1832 ex_data callbacks. This works around a problem where some applications
1833 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1834 restarting) then use compression (e.g. SSL with compression) later.
1835 This results in significant per-connection memory leaks and
1836 has caused some security issues including CVE-2008-1678 and
1840 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1841 change when encrypting or decrypting.
1844 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1845 connect and renegotiate with servers which do not support RI.
1846 Until RI is more widely deployed this option is enabled by default.
1849 *) Add "missing" ssl ctrls to clear options and mode.
1852 *) If client attempts to renegotiate and doesn't support RI respond with
1853 a no_renegotiation alert as required by RFC5746. Some renegotiating
1854 TLS clients will continue a connection gracefully when they receive
1855 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1856 waiting for a server hello which it will never receive. Now we treat a
1857 received no_renegotiation alert as a fatal error. This is because
1858 applications requesting a renegotiation might well expect it to succeed
1859 and would have no code in place to handle the server denying it so the
1860 only safe thing to do is to terminate the connection.
1863 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1864 peer supports secure renegotiation and 0 otherwise. Print out peer
1865 renegotiation support in s_client/s_server.
1868 *) Replace the highly broken and deprecated SPKAC certification method with
1869 the updated NID creation version. This should correctly handle UTF8.
1872 *) Implement RFC5746. Re-enable renegotiation but require the extension
1873 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
1874 turns out to be a bad idea. It has been replaced by
1875 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
1876 SSL_CTX_set_options(). This is really not recommended unless you
1877 know what you are doing.
1878 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
1880 *) Fixes to stateless session resumption handling. Use initial_ctx when
1881 issuing and attempting to decrypt tickets in case it has changed during
1882 servername handling. Use a non-zero length session ID when attempting
1883 stateless session resumption: this makes it possible to determine if
1884 a resumption has occurred immediately after receiving server hello
1885 (several places in OpenSSL subtly assume this) instead of later in
1889 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1890 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1891 fixes for a few places where the return code is not checked
1893 [Julia Lawall <julia@diku.dk>]
1895 *) Add --strict-warnings option to Configure script to include devteam
1896 warnings in other configurations.
1899 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1900 makes it possible to install openssl libraries in locations which
1901 have names other than "lib", for example "/usr/lib64" which some
1903 [Steve Henson, based on patch from Jeremy Utley]
1905 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1906 X690 8.9.12 and can produce some misleading textual output of OIDs.
1907 [Steve Henson, reported by Dan Kaminsky]
1909 *) Delete MD2 from algorithm tables. This follows the recommendation in
1910 several standards that it is not used in new applications due to
1911 several cryptographic weaknesses. For binary compatibility reasons
1912 the MD2 API is still compiled in by default.
1915 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1919 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1920 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1922 [Guenter <lists@gknw.net>]
1924 *) Fix the server certificate chain building code to use X509_verify_cert(),
1925 it used to have an ad-hoc builder which was unable to cope with anything
1926 other than a simple chain.
1927 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1929 *) Don't check self signed certificate signatures in X509_verify_cert()
1930 by default (a flag can override this): it just wastes time without
1931 adding any security. As a useful side effect self signed root CAs
1932 with non-FIPS digests are now usable in FIPS mode.
1935 *) In dtls1_process_out_of_seq_message() the check if the current message
1936 is already buffered was missing. For every new message was memory
1937 allocated, allowing an attacker to perform an denial of service attack
1938 with sending out of seq handshake messages until there is no memory
1939 left. Additionally every future messege was buffered, even if the
1940 sequence number made no sense and would be part of another handshake.
1941 So only messages with sequence numbers less than 10 in advance will be
1942 buffered. (CVE-2009-1378)
1943 [Robin Seggelmann, discovered by Daniel Mentz]
1945 *) Records are buffered if they arrive with a future epoch to be
1946 processed after finishing the corresponding handshake. There is
1947 currently no limitation to this buffer allowing an attacker to perform
1948 a DOS attack with sending records with future epochs until there is no
1949 memory left. This patch adds the pqueue_size() function to detemine
1950 the size of a buffer and limits the record buffer to 100 entries.
1952 [Robin Seggelmann, discovered by Daniel Mentz]
1954 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
1955 parent structure is freed. (CVE-2009-1379)
1958 *) Handle non-blocking I/O properly in SSL_shutdown() call.
1959 [Darryl Miles <darryl-mailinglists@netbauds.net>]
1962 [Ilya O. <vrghost@gmail.com>]
1964 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
1966 *) Disable renegotiation completely - this fixes a severe security
1967 problem (CVE-2009-3555) at the cost of breaking all
1968 renegotiation. Renegotiation can be re-enabled by setting
1969 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
1970 run-time. This is really not recommended unless you know what
1974 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
1976 *) Don't set val to NULL when freeing up structures, it is freed up by
1977 underlying code. If sizeof(void *) > sizeof(long) this can result in
1978 zeroing past the valid field. (CVE-2009-0789)
1979 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
1981 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
1982 checked correctly. This would allow some invalid signed attributes to
1983 appear to verify correctly. (CVE-2009-0591)
1984 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1986 *) Reject UniversalString and BMPString types with invalid lengths. This
1987 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
1988 a legal length. (CVE-2009-0590)
1991 *) Set S/MIME signing as the default purpose rather than setting it
1992 unconditionally. This allows applications to override it at the store
1996 *) Permit restricted recursion of ASN1 strings. This is needed in practice
1997 to handle some structures.
2000 *) Improve efficiency of mem_gets: don't search whole buffer each time
2002 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2004 *) New -hex option for openssl rand.
2007 *) Print out UTF8String and NumericString when parsing ASN1.
2010 *) Support NumericString type for name components.
2013 *) Allow CC in the environment to override the automatically chosen
2014 compiler. Note that nothing is done to ensure flags work with the
2018 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2020 *) Properly check EVP_VerifyFinal() and similar return values
2022 [Ben Laurie, Bodo Moeller, Google Security Team]
2024 *) Enable TLS extensions by default.
2027 *) Allow the CHIL engine to be loaded, whether the application is
2028 multithreaded or not. (This does not release the developer from the
2029 obligation to set up the dynamic locking callbacks.)
2030 [Sander Temme <sander@temme.net>]
2032 *) Use correct exit code if there is an error in dgst command.
2033 [Steve Henson; problem pointed out by Roland Dirlewanger]
2035 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2036 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2039 *) Add experimental JPAKE support, including demo authentication in
2040 s_client and s_server.
2043 *) Set the comparison function in v3_addr_canonize().
2044 [Rob Austein <sra@hactrn.net>]
2046 *) Add support for XMPP STARTTLS in s_client.
2047 [Philip Paeps <philip@freebsd.org>]
2049 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2050 to ensure that even with this option, only ciphersuites in the
2051 server's preference list will be accepted. (Note that the option
2052 applies only when resuming a session, so the earlier behavior was
2053 just about the algorithm choice for symmetric cryptography.)
2056 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2058 *) Fix NULL pointer dereference if a DTLS server received
2059 ChangeCipherSpec as first record (CVE-2009-1386).
2062 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
2063 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2066 *) The fix in 0.9.8c that supposedly got rid of unsafe
2067 double-checked locking was incomplete for RSA blinding,
2068 addressing just one layer of what turns out to have been
2069 doubly unsafe triple-checked locking.
2071 So now fix this for real by retiring the MONT_HELPER macro
2072 in crypto/rsa/rsa_eay.c.
2074 [Bodo Moeller; problem pointed out by Marius Schilder]
2076 *) Various precautionary measures:
2078 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2080 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2081 (NB: This would require knowledge of the secret session ticket key
2082 to exploit, in which case you'd be SOL either way.)
2084 - Change bn_nist.c so that it will properly handle input BIGNUMs
2085 outside the expected range.
2087 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2090 [Neel Mehta, Bodo Moeller]
2092 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2093 the load fails. Useful for distros.
2094 [Ben Laurie and the FreeBSD team]
2096 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2099 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2102 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2104 This work was sponsored by Logica.
2107 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2108 keystores. Support for SSL/TLS client authentication too.
2109 Not compiled unless enable-capieng specified to Configure.
2111 This work was sponsored by Logica.
2114 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2115 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2116 attribute creation routines such as certifcate requests and PKCS#12
2120 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2122 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2123 handshake which could lead to a cilent crash as found using the
2124 Codenomicon TLS test suite (CVE-2008-1672)
2125 [Steve Henson, Mark Cox]
2127 *) Fix double free in TLS server name extensions which could lead to
2128 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
2131 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
2133 Clear the error queue to ensure that error entries left from
2134 older function calls do not interfere with the correct operation.
2135 [Lutz Jaenicke, Erik de Castro Lopo]
2137 *) Remove root CA certificates of commercial CAs:
2139 The OpenSSL project does not recommend any specific CA and does not
2140 have any policy with respect to including or excluding any CA.
2141 Therefore it does not make any sense to ship an arbitrary selection
2142 of root CA certificates with the OpenSSL software.
2145 *) RSA OAEP patches to fix two separate invalid memory reads.
2146 The first one involves inputs when 'lzero' is greater than
2147 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
2148 before the beginning of from). The second one involves inputs where
2149 the 'db' section contains nothing but zeroes (there is a one-byte
2150 invalid read after the end of 'db').
2151 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2153 *) Partial backport from 0.9.9-dev:
2155 Introduce bn_mul_mont (dedicated Montgomery multiplication
2156 procedure) as a candidate for BIGNUM assembler implementation.
2157 While 0.9.9-dev uses assembler for various architectures, only
2158 x86_64 is available by default here in the 0.9.8 branch, and
2159 32-bit x86 is available through a compile-time setting.
2161 To try the 32-bit x86 assembler implementation, use Configure
2162 option "enable-montasm" (which exists only for this backport).
2164 As "enable-montasm" for 32-bit x86 disclaims code stability
2165 anyway, in this constellation we activate additional code
2166 backported from 0.9.9-dev for further performance improvements,
2167 namely BN_from_montgomery_word. (To enable this otherwise,
2168 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
2170 [Andy Polyakov (backport partially by Bodo Moeller)]
2172 *) Add TLS session ticket callback. This allows an application to set
2173 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
2174 values. This is useful for key rollover for example where several key
2175 sets may exist with different names.
2178 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
2179 This was broken until now in 0.9.8 releases, such that the only way
2180 a registered ENGINE could be used (assuming it initialises
2181 successfully on the host) was to explicitly set it as the default
2182 for the relevant algorithms. This is in contradiction with 0.9.7
2183 behaviour and the documentation. With this fix, when an ENGINE is
2184 registered into a given algorithm's table of implementations, the
2185 'uptodate' flag is reset so that auto-discovery will be used next
2186 time a new context for that algorithm attempts to select an
2188 [Ian Lister (tweaked by Geoff Thorpe)]
2190 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
2191 implemention in the following ways:
2193 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
2196 Lack of BER streaming support means one pass streaming processing is
2197 only supported if data is detached: setting the streaming flag is
2198 ignored for embedded content.
2200 CMS support is disabled by default and must be explicitly enabled
2201 with the enable-cms configuration option.
2204 *) Update the GMP engine glue to do direct copies between BIGNUM and
2205 mpz_t when openssl and GMP use the same limb size. Otherwise the
2206 existing "conversion via a text string export" trick is still used.
2207 [Paul Sheer <paulsheer@gmail.com>]
2209 *) Zlib compression BIO. This is a filter BIO which compressed and
2210 uncompresses any data passed through it.
2213 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
2214 RFC3394 compatible AES key wrapping.
2217 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
2218 sets string data without copying. X509_ALGOR_set0() and
2219 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
2220 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
2221 from an X509_ATTRIBUTE structure optionally checking it occurs only
2222 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
2226 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
2227 to get the expected BN_FLG_CONSTTIME behavior.
2228 [Bodo Moeller (Google)]
2232 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
2233 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
2234 - added some more tests to do_tests.pl
2235 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
2236 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
2237 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
2238 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
2239 - various changes to netware.pl to enable gcc-cross builds on Win32
2241 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
2242 - various changes to fix missing prototype warnings
2243 - fixed x86nasm.pl to create correct asm files for NASM COFF output
2244 - added AES, WHIRLPOOL and CPUID assembler code to build files
2245 - added missing AES assembler make rules to mk1mf.pl
2246 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
2247 [Guenter Knauf <eflash@gmx.net>]
2249 *) Implement certificate status request TLS extension defined in RFC3546.
2250 A client can set the appropriate parameters and receive the encoded
2251 OCSP response via a callback. A server can query the supplied parameters
2252 and set the encoded OCSP response in the callback. Add simplified examples
2253 to s_client and s_server.
2256 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
2258 *) Fix various bugs:
2259 + Binary incompatibility of ssl_ctx_st structure
2260 + DTLS interoperation with non-compliant servers
2261 + Don't call get_session_cb() without proposed session
2262 + Fix ia64 assembler code
2263 [Andy Polyakov, Steve Henson]
2265 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
2267 *) DTLS Handshake overhaul. There were longstanding issues with
2268 OpenSSL DTLS implementation, which were making it impossible for
2269 RFC 4347 compliant client to communicate with OpenSSL server.
2270 Unfortunately just fixing these incompatibilities would "cut off"
2271 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
2272 server keeps tolerating non RFC compliant syntax. The opposite is
2273 not true, 0.9.8f client can not communicate with earlier server.
2274 This update even addresses CVE-2007-4995.
2277 *) Changes to avoid need for function casts in OpenSSL: some compilers
2278 (gcc 4.2 and later) reject their use.
2279 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
2282 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2283 RFC4507bis. The encrypted ticket format is an encrypted encoded
2284 SSL_SESSION structure, that way new session features are automatically
2287 If a client application caches session in an SSL_SESSION structure
2288 support is transparent because tickets are now stored in the encoded
2291 The SSL_CTX structure automatically generates keys for ticket
2292 protection in servers so again support should be possible
2293 with no application modification.
2295 If a client or server wishes to disable RFC4507 support then the option
2296 SSL_OP_NO_TICKET can be set.
2298 Add a TLS extension debugging callback to allow the contents of any client
2299 or server extensions to be examined.
2301 This work was sponsored by Google.
2304 *) Add initial support for TLS extensions, specifically for the server_name
2305 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2306 have new members for a host name. The SSL data structure has an
2307 additional member SSL_CTX *initial_ctx so that new sessions can be
2308 stored in that context to allow for session resumption, even after the
2309 SSL has been switched to a new SSL_CTX in reaction to a client's
2310 server_name extension.
2312 New functions (subject to change):
2314 SSL_get_servername()
2315 SSL_get_servername_type()
2318 New CTRL codes and macros (subject to change):
2320 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2321 - SSL_CTX_set_tlsext_servername_callback()
2322 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2323 - SSL_CTX_set_tlsext_servername_arg()
2324 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2326 openssl s_client has a new '-servername ...' option.
2328 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2329 '-key2 ...', '-servername_fatal' (subject to change). This allows
2330 testing the HostName extension for a specific single host name ('-cert'
2331 and '-key' remain fallbacks for handshakes without HostName
2332 negotiation). If the unrecogninzed_name alert has to be sent, this by
2333 default is a warning; it becomes fatal with the '-servername_fatal'
2336 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2338 *) Add AES and SSE2 assembly language support to VC++ build.
2341 *) Mitigate attack on final subtraction in Montgomery reduction.
2344 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2345 (which previously caused an internal error).
2348 *) Squeeze another 10% out of IGE mode when in != out.
2351 *) AES IGE mode speedup.
2352 [Dean Gaudet (Google)]
2354 *) Add the Korean symmetric 128-bit cipher SEED (see
2355 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2356 add SEED ciphersuites from RFC 4162:
2358 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2359 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2360 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2361 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2363 To minimize changes between patchlevels in the OpenSSL 0.9.8
2364 series, SEED remains excluded from compilation unless OpenSSL
2365 is configured with 'enable-seed'.
2366 [KISA, Bodo Moeller]
2368 *) Mitigate branch prediction attacks, which can be practical if a
2369 single processor is shared, allowing a spy process to extract
2370 information. For detailed background information, see
2371 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2372 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2373 and Necessary Software Countermeasures"). The core of the change
2374 are new versions BN_div_no_branch() and
2375 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2376 respectively, which are slower, but avoid the security-relevant
2377 conditional branches. These are automatically called by BN_div()
2378 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2379 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2380 remove a conditional branch.
2382 BN_FLG_CONSTTIME is the new name for the previous
2383 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2384 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2385 in the exponent causes BN_mod_exp_mont() to use the alternative
2386 implementation in BN_mod_exp_mont_consttime().) The old name
2387 remains as a deprecated alias.
2389 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2390 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2391 constant-time implementations for more than just exponentiation.
2392 Here too the old name is kept as a deprecated alias.
2394 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2395 the BN_BLINDING structure gets an independent copy of the
2396 modulus. This means that the previous "BIGNUM *m" argument to
2397 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2398 essentially becomes "const BIGNUM *m", although we can't actually
2399 change this in the header file before 0.9.9. It allows
2400 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2401 enable BN_FLG_CONSTTIME.
2403 [Matthew D Wood (Intel Corp)]
2405 *) In the SSL/TLS server implementation, be strict about session ID
2406 context matching (which matters if an application uses a single
2407 external cache for different purposes). Previously,
2408 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2409 set. This did ensure strict client verification, but meant that,
2410 with applications using a single external cache for quite
2411 different requirements, clients could circumvent ciphersuite
2412 restrictions for a given session ID context by starting a session
2413 in a different context.
2416 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2417 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2418 authentication-only ciphersuites.
2421 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2422 not complete and could lead to a possible single byte overflow
2423 (CVE-2007-5135) [Ben Laurie]
2425 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2427 *) Since AES128 and AES256 (and similarly Camellia128 and
2428 Camellia256) share a single mask bit in the logic of
2429 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2430 kludge to work properly if AES128 is available and AES256 isn't
2431 (or if Camellia128 is available and Camellia256 isn't).
2434 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2435 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2436 When a point or a seed is encoded in a BIT STRING, we need to
2437 prevent the removal of trailing zero bits to get the proper DER
2438 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2439 of a NamedBitList, for which trailing 0 bits need to be removed.)
2442 *) Have SSL/TLS server implementation tolerate "mismatched" record
2443 protocol version while receiving ClientHello even if the
2444 ClientHello is fragmented. (The server can't insist on the
2445 particular protocol version it has chosen before the ServerHello
2446 message has informed the client about his choice.)
2449 *) Add RFC 3779 support.
2450 [Rob Austein for ARIN, Ben Laurie]
2452 *) Load error codes if they are not already present instead of using a
2453 static variable. This allows them to be cleanly unloaded and reloaded.
2454 Improve header file function name parsing.
2457 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2458 or CAPABILITY handshake as required by RFCs.
2461 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2463 *) Introduce limits to prevent malicious keys being able to
2464 cause a denial of service. (CVE-2006-2940)
2465 [Steve Henson, Bodo Moeller]
2467 *) Fix ASN.1 parsing of certain invalid structures that can result
2468 in a denial of service. (CVE-2006-2937) [Steve Henson]
2470 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2471 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2473 *) Fix SSL client code which could crash if connecting to a
2474 malicious SSLv2 server. (CVE-2006-4343)
2475 [Tavis Ormandy and Will Drewry, Google Security Team]
2477 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2478 match only those. Before that, "AES256-SHA" would be interpreted
2479 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2480 the same strength classification in 0.9.7h) as we currently only
2481 have a single AES bit in the ciphersuite description bitmap.
2482 That change, however, also applied to ciphersuite strings such as
2483 "RC4-MD5" that intentionally matched multiple ciphersuites --
2484 namely, SSL 2.0 ciphersuites in addition to the more common ones
2485 from SSL 3.0/TLS 1.0.
2487 So we change the selection algorithm again: Naming an explicit
2488 ciphersuite selects this one ciphersuite, and any other similar
2489 ciphersuite (same bitmap) from *other* protocol versions.
2490 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2491 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2493 Since SSL 2.0 does not have any ciphersuites for which the
2494 128/256 bit distinction would be relevant, this works for now.
2495 The proper fix will be to use different bits for AES128 and
2496 AES256, which would have avoided the problems from the beginning;
2497 however, bits are scarce, so we can only do this in a new release
2498 (not just a patchlevel) when we can change the SSL_CIPHER
2499 definition to split the single 'unsigned long mask' bitmap into
2500 multiple values to extend the available space.
2504 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2506 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2507 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2509 *) Add AES IGE and biIGE modes.
2512 *) Change the Unix randomness entropy gathering to use poll() when
2513 possible instead of select(), since the latter has some
2514 undesirable limitations.
2515 [Darryl Miles via Richard Levitte and Bodo Moeller]
2517 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2518 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2519 cannot be implicitly activated as part of, e.g., the "AES" alias.
2520 However, please upgrade to OpenSSL 0.9.9[-dev] for
2521 non-experimental use of the ECC ciphersuites to get TLS extension
2522 support, which is required for curve and point format negotiation
2523 to avoid potential handshake problems.
2526 *) Disable rogue ciphersuites:
2528 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2529 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2530 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2532 The latter two were purportedly from
2533 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2536 Also deactivate the remaining ciphersuites from
2537 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2538 unofficial, and the ID has long expired.
2541 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2542 dual-core machines) and other potential thread-safety issues.
2545 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2546 versions), which is now available for royalty-free use
2547 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2548 Also, add Camellia TLS ciphersuites from RFC 4132.
2550 To minimize changes between patchlevels in the OpenSSL 0.9.8
2551 series, Camellia remains excluded from compilation unless OpenSSL
2552 is configured with 'enable-camellia'.
2555 *) Disable the padding bug check when compression is in use. The padding
2556 bug check assumes the first packet is of even length, this is not
2557 necessarily true if compresssion is enabled and can result in false
2558 positives causing handshake failure. The actual bug test is ancient
2559 code so it is hoped that implementations will either have fixed it by
2560 now or any which still have the bug do not support compression.
2563 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2565 *) When applying a cipher rule check to see if string match is an explicit
2566 cipher suite and only match that one cipher suite if it is.
2569 *) Link in manifests for VC++ if needed.
2570 [Austin Ziegler <halostatue@gmail.com>]
2572 *) Update support for ECC-based TLS ciphersuites according to
2573 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2574 TLS extensions, which are supported starting with the 0.9.9
2575 branch, not in the OpenSSL 0.9.8 branch).
2578 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2579 opaque EVP_CIPHER_CTX handling.
2582 *) Fixes and enhancements to zlib compression code. We now only use
2583 "zlib1.dll" and use the default __cdecl calling convention on Win32
2584 to conform with the standards mentioned here:
2585 http://www.zlib.net/DLL_FAQ.txt
2586 Static zlib linking now works on Windows and the new --with-zlib-include
2587 --with-zlib-lib options to Configure can be used to supply the location
2588 of the headers and library. Gracefully handle case where zlib library
2592 *) Several fixes and enhancements to the OID generation code. The old code
2593 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2594 handle numbers larger than ULONG_MAX, truncated printing and had a
2595 non standard OBJ_obj2txt() behaviour.
2598 *) Add support for building of engines under engine/ as shared libraries
2599 under VC++ build system.
2602 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2603 Hopefully, we will not see any false combination of paths any more.
2606 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2608 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2609 (part of SSL_OP_ALL). This option used to disable the
2610 countermeasure against man-in-the-middle protocol-version
2611 rollback in the SSL 2.0 server implementation, which is a bad
2612 idea. (CVE-2005-2969)
2614 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2615 for Information Security, National Institute of Advanced Industrial
2616 Science and Technology [AIST], Japan)]
2618 *) Add two function to clear and return the verify parameter flags.
2621 *) Keep cipherlists sorted in the source instead of sorting them at
2622 runtime, thus removing the need for a lock.
2625 *) Avoid some small subgroup attacks in Diffie-Hellman.
2626 [Nick Mathewson and Ben Laurie]
2628 *) Add functions for well-known primes.
2631 *) Extended Windows CE support.
2632 [Satoshi Nakamura and Andy Polyakov]
2634 *) Initialize SSL_METHOD structures at compile time instead of during
2635 runtime, thus removing the need for a lock.
2638 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2639 attempting to decrypt each encrypted key in turn. Add support to
2643 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2645 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2648 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2651 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2652 key into the same file any more.
2655 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2658 *) Add -utf8 command line and config file option to 'ca'.
2659 [Stefan <stf@udoma.org]
2661 *) Removed the macro des_crypt(), as it seems to conflict with some
2662 libraries. Use DES_crypt().
2665 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2666 involves renaming the source and generated shared-libs for
2667 both. The engines will accept the corrected or legacy ids
2668 ('ncipher' and '4758_cca' respectively) when binding. NB,
2669 this only applies when building 'shared'.
2670 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2672 *) Add attribute functions to EVP_PKEY structure. Modify
2673 PKCS12_create() to recognize a CSP name attribute and
2674 use it. Make -CSP option work again in pkcs12 utility.
2677 *) Add new functionality to the bn blinding code:
2678 - automatic re-creation of the BN_BLINDING parameters after
2679 a fixed number of uses (currently 32)
2680 - add new function for parameter creation
2681 - introduce flags to control the update behaviour of the
2682 BN_BLINDING parameters
2683 - hide BN_BLINDING structure
2684 Add a second BN_BLINDING slot to the RSA structure to improve
2685 performance when a single RSA object is shared among several
2689 *) Add support for DTLS.
2690 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2692 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2693 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2696 *) Remove buggy and incompletet DH cert support from
2697 ssl/ssl_rsa.c and ssl/s3_both.c
2700 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2701 the apps/openssl applications.
2704 *) Compile clean with "-Wall -Wmissing-prototypes
2705 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2706 DEBUG_SAFESTACK must also be set.
2709 *) Change ./Configure so that certain algorithms can be disabled by default.
2710 The new counterpiece to "no-xxx" is "enable-xxx".
2712 The patented RC5 and MDC2 algorithms will now be disabled unless
2713 "enable-rc5" and "enable-mdc2", respectively, are specified.
2715 (IDEA remains enabled despite being patented. This is because IDEA
2716 is frequently required for interoperability, and there is no license
2717 fee for non-commercial use. As before, "no-idea" can be used to
2718 avoid this algorithm.)
2722 *) Add processing of proxy certificates (see RFC 3820). This work was
2723 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2724 EGEE (Enabling Grids for E-science in Europe).
2727 *) RC4 performance overhaul on modern architectures/implementations, such
2728 as Intel P4, IA-64 and AMD64.
2731 *) New utility extract-section.pl. This can be used specify an alternative
2732 section number in a pod file instead of having to treat each file as
2733 a separate case in Makefile. This can be done by adding two lines to the
2736 =for comment openssl_section:XXX
2738 The blank line is mandatory.
2742 *) New arguments -certform, -keyform and -pass for s_client and s_server
2743 to allow alternative format key and certificate files and passphrase
2747 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2748 update associated structures and add various utility functions.
2750 Add new policy related verify parameters, include policy checking in
2751 standard verify code. Enhance 'smime' application with extra parameters
2752 to support policy checking and print out.
2755 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2756 Nehemiah processors. These extensions support AES encryption in hardware
2757 as well as RNG (though RNG support is currently disabled).
2758 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2760 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2763 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2764 [Andy Polyakov and a number of other people]
2766 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2767 implementation contributed by IBM.
2768 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2770 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2771 exponent rather than 'unsigned long'. There is a corresponding change to
2772 the new 'rsa_keygen' element of the RSA_METHOD structure.
2773 [Jelte Jansen, Geoff Thorpe]
2775 *) Functionality for creating the initial serial number file is now
2776 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2778 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2779 number file to 1, which is bound to cause problems. To avoid
2780 the problems while respecting compatibility between different 0.9.7
2781 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2782 CA.pl for serial number initialization. With the new release 0.9.8,
2783 we can fix the problem directly in the 'ca' utility.)
2786 *) Reduced header interdepencies by declaring more opaque objects in
2787 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2788 give fewer recursive includes, which could break lazy source code - so
2789 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2790 developers should define this symbol when building and using openssl to
2791 ensure they track the recommended behaviour, interfaces, [etc], but
2792 backwards-compatible behaviour prevails when this isn't defined.
2795 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2798 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2799 This will generate a random key of the appropriate length based on the
2800 cipher context. The EVP_CIPHER can provide its own random key generation
2801 routine to support keys of a specific form. This is used in the des and
2802 3des routines to generate a key of the correct parity. Update S/MIME
2803 code to use new functions and hence generate correct parity DES keys.
2804 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2805 valid (weak or incorrect parity).
2808 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2809 as looking them up. This is useful when the verified structure may contain
2810 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2811 present unless the new PKCS7_NO_CRL flag is asserted.
2814 *) Extend ASN1 oid configuration module. It now additionally accepts the
2817 shortName = some long name, 1.2.3.4
2820 *) Reimplemented the BN_CTX implementation. There is now no more static
2821 limitation on the number of variables it can handle nor the depth of the
2822 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2823 information can now expand as required, and rather than having a single
2824 static array of bignums, BN_CTX now uses a linked-list of such arrays
2825 allowing it to expand on demand whilst maintaining the usefulness of
2826 BN_CTX's "bundling".
2829 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2830 to allow all RSA operations to function using a single BN_CTX.
2833 *) Preliminary support for certificate policy evaluation and checking. This
2834 is initially intended to pass the tests outlined in "Conformance Testing
2835 of Relying Party Client Certificate Path Processing Logic" v1.07.
2838 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2839 remained unused and not that useful. A variety of other little bignum
2840 tweaks and fixes have also been made continuing on from the audit (see
2844 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2845 associated ASN1, EVP and SSL functions and old ASN1 macros.
2848 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2849 and this should never fail. So the return value from the use of
2850 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2851 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2854 *) BN_CTX_get() should return zero-valued bignums, providing the same
2855 initialised value as BN_new().
2856 [Geoff Thorpe, suggested by Ulf Möller]
2858 *) Support for inhibitAnyPolicy certificate extension.
2861 *) An audit of the BIGNUM code is underway, for which debugging code is
2862 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2863 is considered valid when processing BIGNUMs, and causes execution to
2864 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2865 further steps are taken to deliberately pollute unused data in BIGNUM
2866 structures to try and expose faulty code further on. For now, openssl will
2867 (in its default mode of operation) continue to tolerate the inconsistent
2868 forms that it has tolerated in the past, but authors and packagers should
2869 consider trying openssl and their own applications when compiled with
2870 these debugging symbols defined. It will help highlight potential bugs in
2871 their own code, and will improve the test coverage for OpenSSL itself. At
2872 some point, these tighter rules will become openssl's default to improve
2873 maintainability, though the assert()s and other overheads will remain only
2874 in debugging configurations. See bn.h for more details.
2875 [Geoff Thorpe, Nils Larsch, Ulf Möller]
2877 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
2878 that can only be obtained through BN_CTX_new() (which implicitly
2879 initialises it). The presence of this function only made it possible
2880 to overwrite an existing structure (and cause memory leaks).
2883 *) Because of the callback-based approach for implementing LHASH as a
2884 template type, lh_insert() adds opaque objects to hash-tables and
2885 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2886 to clean up those corresponding objects before destroying the hash table
2887 (and losing the object pointers). So some over-zealous constifications in
2888 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2889 objects as "const" and the lh_doall[_arg] callback wrappers are not
2890 prototyped to have "const" restrictions on the object pointers they are
2891 given (and so aren't required to cast them away any more).
2894 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2895 (speed) prefers to use its own implementation. The two implementations
2896 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2897 its object type properly exposed (MS_TM) instead of casting to/from "char
2898 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2899 aren't necessarily the greatest nomenclatures - but this is what was used
2900 internally to the implementation so I've used that for now.
2903 *) Ensure that deprecated functions do not get compiled when
2904 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2905 the self-tests were still using deprecated key-generation functions so
2906 these have been updated also.
2909 *) Reorganise PKCS#7 code to separate the digest location functionality
2910 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2911 New function PKCS7_set_digest() to set the digest type for PKCS#7
2912 digestedData type. Add additional code to correctly generate the
2913 digestedData type and add support for this type in PKCS7 initialization
2917 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2918 structure of type "other".
2921 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2922 sure the loop does correctly stop and breaking ("division by zero")
2923 modulus operations are not performed. The (pre-generated) prime
2924 table crypto/bn/bn_prime.h was already correct, but it could not be
2925 re-generated on some platforms because of the "division by zero"
2926 situation in the script.
2927 [Ralf S. Engelschall]
2929 *) Update support for ECC-based TLS ciphersuites according to
2930 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2931 SHA-1 now is only used for "small" curves (where the
2932 representation of a field element takes up to 24 bytes); for
2933 larger curves, the field element resulting from ECDH is directly
2934 used as premaster secret.
2935 [Douglas Stebila (Sun Microsystems Laboratories)]
2937 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
2938 curve secp160r1 to the tests.
2939 [Douglas Stebila (Sun Microsystems Laboratories)]
2941 *) Add the possibility to load symbols globally with DSO.
2942 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
2944 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
2945 control of the error stack.
2948 *) Add support for STORE in ENGINE.
2951 *) Add the STORE type. The intention is to provide a common interface
2952 to certificate and key stores, be they simple file-based stores, or
2953 HSM-type store, or LDAP stores, or...
2954 NOTE: The code is currently UNTESTED and isn't really used anywhere.
2957 *) Add a generic structure called OPENSSL_ITEM. This can be used to
2958 pass a list of arguments to any function as well as provide a way
2959 for a function to pass data back to the caller.
2962 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
2963 works like BUF_strdup() but can be used to duplicate a portion of
2964 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
2968 *) Add the function sk_find_ex() which works like sk_find(), but will
2969 return an index to an element even if an exact match couldn't be
2970 found. The index is guaranteed to point at the element where the
2971 searched-for key would be inserted to preserve sorting order.
2974 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
2975 takes an extra flags argument for optional functionality. Currently,
2976 the following flags are defined:
2978 OBJ_BSEARCH_VALUE_ON_NOMATCH
2979 This one gets OBJ_bsearch_ex() to return a pointer to the first
2980 element where the comparing function returns a negative or zero
2983 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
2984 This one gets OBJ_bsearch_ex() to return a pointer to the first
2985 element where the comparing function returns zero. This is useful
2986 if there are more than one element where the comparing function
2990 *) Make it possible to create self-signed certificates with 'openssl ca'
2991 in such a way that the self-signed certificate becomes part of the
2992 CA database and uses the same mechanisms for serial number generation
2993 as all other certificate signing. The new flag '-selfsign' enables
2994 this functionality. Adapt CA.sh and CA.pl.in.
2997 *) Add functionality to check the public key of a certificate request
2998 against a given private. This is useful to check that a certificate
2999 request can be signed by that key (self-signing).
3002 *) Make it possible to have multiple active certificates with the same
3003 subject in the CA index file. This is done only if the keyword
3004 'unique_subject' is set to 'no' in the main CA section (default
3005 if 'CA_default') of the configuration file. The value is saved
3006 with the database itself in a separate index attribute file,
3007 named like the index file with '.attr' appended to the name.
3010 *) Generate muti valued AVAs using '+' notation in config files for
3014 *) Support for nameConstraints certificate extension.
3017 *) Support for policyConstraints certificate extension.
3020 *) Support for policyMappings certificate extension.
3023 *) Make sure the default DSA_METHOD implementation only uses its
3024 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3025 and change its own handlers to be NULL so as to remove unnecessary
3026 indirection. This lets alternative implementations fallback to the
3027 default implementation more easily.
3030 *) Support for directoryName in GeneralName related extensions
3034 *) Make it possible to link applications using Makefile.shared.
3035 Make that possible even when linking against static libraries!
3038 *) Support for single pass processing for S/MIME signing. This now
3039 means that S/MIME signing can be done from a pipe, in addition
3040 cleartext signing (multipart/signed type) is effectively streaming
3041 and the signed data does not need to be all held in memory.
3043 This is done with a new flag PKCS7_STREAM. When this flag is set
3044 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3045 is done after the data is output (and digests calculated) in
3046 SMIME_write_PKCS7().
3049 *) Add full support for -rpath/-R, both in shared libraries and
3050 applications, at least on the platforms where it's known how
3054 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3055 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3056 will now compute a table of multiples of the generator that
3057 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3058 faster (notably in the case of a single point multiplication,
3059 scalar * generator).
3060 [Nils Larsch, Bodo Moeller]
3062 *) IPv6 support for certificate extensions. The various extensions
3063 which use the IP:a.b.c.d can now take IPv6 addresses using the
3064 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3068 *) Added an ENGINE that implements RSA by performing private key
3069 exponentiations with the GMP library. The conversions to and from
3070 GMP's mpz_t format aren't optimised nor are any montgomery forms
3071 cached, and on x86 it appears OpenSSL's own performance has caught up.
3072 However there are likely to be other architectures where GMP could
3073 provide a boost. This ENGINE is not built in by default, but it can be
3074 specified at Configure time and should be accompanied by the necessary
3075 linker additions, eg;
3076 ./config -DOPENSSL_USE_GMP -lgmp
3079 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3080 testing availability of engines with "-t" - the old behaviour is
3081 produced by increasing the feature's verbosity with "-tt".
3084 *) ECDSA routines: under certain error conditions uninitialized BN objects
3085 could be freed. Solution: make sure initialization is performed early
3086 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3090 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3091 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3092 software implementations. For DSA and DH, parameter generation can
3093 also be overriden by providing the appropriate method callbacks.
3096 *) Change the "progress" mechanism used in key-generation and
3097 primality testing to functions that take a new BN_GENCB pointer in
3098 place of callback/argument pairs. The new API functions have "_ex"
3099 postfixes and the older functions are reimplemented as wrappers for
3100 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3101 declarations of the old functions to help (graceful) attempts to
3102 migrate to the new functions. Also, the new key-generation API
3103 functions operate on a caller-supplied key-structure and return
3104 success/failure rather than returning a key or NULL - this is to
3105 help make "keygen" another member function of RSA_METHOD etc.
3107 Example for using the new callback interface:
3109 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3113 BN_GENCB_set(&my_cb, my_callback, my_arg);
3115 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3116 /* For the meaning of a, b in calls to my_callback(), see the
3117 * documentation of the function that calls the callback.
3118 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3119 * my_callback should return 1 if it wants BN_is_prime_ex()
3120 * to continue, or 0 to stop.
3125 *) Change the ZLIB compression method to be stateful, and make it
3126 available to TLS with the number defined in
3127 draft-ietf-tls-compression-04.txt.
3130 *) Add the ASN.1 structures and functions for CertificatePair, which
3131 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
3133 CertificatePair ::= SEQUENCE {
3134 forward [0] Certificate OPTIONAL,
3135 reverse [1] Certificate OPTIONAL,
3136 -- at least one of the pair shall be present -- }
3138 Also implement the PEM functions to read and write certificate
3139 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
3141 This needed to be defined, mostly for the sake of the LDAP
3142 attribute crossCertificatePair, but may prove useful elsewhere as
3146 *) Make it possible to inhibit symlinking of shared libraries in
3147 Makefile.shared, for Cygwin's sake.
3150 *) Extend the BIGNUM API by creating a function
3151 void BN_set_negative(BIGNUM *a, int neg);
3152 and a macro that behave like
3153 int BN_is_negative(const BIGNUM *a);
3155 to avoid the need to access 'a->neg' directly in applications.
3158 *) Implement fast modular reduction for pseudo-Mersenne primes
3159 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
3160 EC_GROUP_new_curve_GFp() will now automatically use this
3162 [Nils Larsch <nla@trustcenter.de>]
3164 *) Add new lock type (CRYPTO_LOCK_BN).
3167 *) Change the ENGINE framework to automatically load engines
3168 dynamically from specific directories unless they could be
3169 found to already be built in or loaded. Move all the
3170 current engines except for the cryptodev one to a new
3172 The engines in engines/ are built as shared libraries if
3173 the "shared" options was given to ./Configure or ./config.
3174 Otherwise, they are inserted in libcrypto.a.
3175 /usr/local/ssl/engines is the default directory for dynamic
3176 engines, but that can be overriden at configure time through
3177 the usual use of --prefix and/or --openssldir, and at run
3178 time with the environment variable OPENSSL_ENGINES.
3179 [Geoff Thorpe and Richard Levitte]
3181 *) Add Makefile.shared, a helper makefile to build shared
3182 libraries. Addapt Makefile.org.
3185 *) Add version info to Win32 DLLs.
3186 [Peter 'Luna' Runestig" <peter@runestig.com>]
3188 *) Add new 'medium level' PKCS#12 API. Certificates and keys
3189 can be added using this API to created arbitrary PKCS#12
3190 files while avoiding the low level API.
3192 New options to PKCS12_create(), key or cert can be NULL and
3193 will then be omitted from the output file. The encryption
3194 algorithm NIDs can be set to -1 for no encryption, the mac
3195 iteration count can be set to 0 to omit the mac.
3197 Enhance pkcs12 utility by making the -nokeys and -nocerts
3198 options work when creating a PKCS#12 file. New option -nomac
3199 to omit the mac, NONE can be set for an encryption algorithm.
3200 New code is modified to use the enhanced PKCS12_create()
3201 instead of the low level API.
3204 *) Extend ASN1 encoder to support indefinite length constructed
3205 encoding. This can output sequences tags and octet strings in
3206 this form. Modify pk7_asn1.c to support indefinite length
3207 encoding. This is experimental and needs additional code to
3208 be useful, such as an ASN1 bio and some enhanced streaming
3211 Extend template encode functionality so that tagging is passed
3212 down to the template encoder.
3215 *) Let 'openssl req' fail if an argument to '-newkey' is not
3216 recognized instead of using RSA as a default.
3219 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
3220 As these are not official, they are not included in "ALL";
3221 the "ECCdraft" ciphersuite group alias can be used to select them.
3222 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
3224 *) Add ECDH engine support.
3225 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
3227 *) Add ECDH in new directory crypto/ecdh/.
3228 [Douglas Stebila (Sun Microsystems Laboratories)]
3230 *) Let BN_rand_range() abort with an error after 100 iterations
3231 without success (which indicates a broken PRNG).
3234 *) Change BN_mod_sqrt() so that it verifies that the input value
3235 is really the square of the return value. (Previously,
3236 BN_mod_sqrt would show GIGO behaviour.)
3239 *) Add named elliptic curves over binary fields from X9.62, SECG,
3240 and WAP/WTLS; add OIDs that were still missing.
3242 [Sheueling Chang Shantz and Douglas Stebila
3243 (Sun Microsystems Laboratories)]
3245 *) Extend the EC library for elliptic curves over binary fields
3246 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
3249 EC_GF2m_simple_method
3253 EC_GROUP_new_curve_GF2m
3254 EC_GROUP_set_curve_GF2m
3255 EC_GROUP_get_curve_GF2m
3256 EC_POINT_set_affine_coordinates_GF2m
3257 EC_POINT_get_affine_coordinates_GF2m
3258 EC_POINT_set_compressed_coordinates_GF2m
3260 Point compression for binary fields is disabled by default for
3261 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
3264 As binary polynomials are represented as BIGNUMs, various members
3265 of the EC_GROUP and EC_POINT data structures can be shared
3266 between the implementations for prime fields and binary fields;
3267 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
3268 are essentially identical to their ..._GFp counterparts.
3269 (For simplicity, the '..._GFp' prefix has been dropped from
3270 various internal method names.)
3272 An internal 'field_div' method (similar to 'field_mul' and
3273 'field_sqr') has been added; this is used only for binary fields.
3275 [Sheueling Chang Shantz and Douglas Stebila
3276 (Sun Microsystems Laboratories)]
3278 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
3279 through methods ('mul', 'precompute_mult').
3281 The generic implementations (now internally called 'ec_wNAF_mul'
3282 and 'ec_wNAF_precomputed_mult') remain the default if these
3283 methods are undefined.
3285 [Sheueling Chang Shantz and Douglas Stebila
3286 (Sun Microsystems Laboratories)]
3288 *) New function EC_GROUP_get_degree, which is defined through
3289 EC_METHOD. For curves over prime fields, this returns the bit
3290 length of the modulus.
3292 [Sheueling Chang Shantz and Douglas Stebila
3293 (Sun Microsystems Laboratories)]
3295 *) New functions EC_GROUP_dup, EC_POINT_dup.
3296 (These simply call ..._new and ..._copy).
3298 [Sheueling Chang Shantz and Douglas Stebila
3299 (Sun Microsystems Laboratories)]
3301 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3302 Polynomials are represented as BIGNUMs (where the sign bit is not
3303 used) in the following functions [macros]:
3306 BN_GF2m_sub [= BN_GF2m_add]
3307 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3308 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3309 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3311 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3312 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3313 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3314 BN_GF2m_cmp [= BN_ucmp]
3316 (Note that only the 'mod' functions are actually for fields GF(2^m).
3317 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3319 For some functions, an the irreducible polynomial defining a
3320 field can be given as an 'unsigned int[]' with strictly
3321 decreasing elements giving the indices of those bits that are set;
3322 i.e., p[] represents the polynomial
3323 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3325 p[0] > p[1] > ... > p[k] = 0.
3326 This applies to the following functions:
3331 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3332 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3334 BN_GF2m_mod_sqrt_arr
3335 BN_GF2m_mod_solve_quad_arr
3339 Conversion can be performed by the following functions:
3344 bntest.c has additional tests for binary polynomial arithmetic.
3346 Two implementations for BN_GF2m_mod_div() are available.
3347 The default algorithm simply uses BN_GF2m_mod_inv() and
3348 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3349 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3350 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3352 [Sheueling Chang Shantz and Douglas Stebila
3353 (Sun Microsystems Laboratories)]
3355 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3356 functionality is disabled at compile-time.
3357 [Douglas Stebila <douglas.stebila@sun.com>]
3359 *) Change default behaviour of 'openssl asn1parse' so that more
3360 information is visible when viewing, e.g., a certificate:
3362 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3363 mode the content of non-printable OCTET STRINGs is output in a
3364 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3365 avoid the appearance of a printable string.
3366 [Nils Larsch <nla@trustcenter.de>]
3368 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3370 EC_GROUP_set_asn1_flag()
3371 EC_GROUP_get_asn1_flag()
3372 EC_GROUP_set_point_conversion_form()
3373 EC_GROUP_get_point_conversion_form()
3374 These control ASN1 encoding details:
3375 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3376 has been set to OPENSSL_EC_NAMED_CURVE.
3377 - Points are encoded in uncompressed form by default; options for
3378 asn1_for are as for point2oct, namely
3379 POINT_CONVERSION_COMPRESSED
3380 POINT_CONVERSION_UNCOMPRESSED
3381 POINT_CONVERSION_HYBRID
3383 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3386 EC_GROUP_get0_seed()
3387 EC_GROUP_get_seed_len()
3388 This is used only for ASN1 purposes (so far).
3389 [Nils Larsch <nla@trustcenter.de>]
3391 *) Add 'field_type' member to EC_METHOD, which holds the NID
3392 of the appropriate field type OID. The new function
3393 EC_METHOD_get_field_type() returns this value.
3394 [Nils Larsch <nla@trustcenter.de>]
3399 EC_POINT_point2hex()
3400 EC_POINT_hex2point()
3401 providing useful interfaces to EC_POINT_point2oct() and
3402 EC_POINT_oct2point().
3403 [Nils Larsch <nla@trustcenter.de>]
3405 *) Change internals of the EC library so that the functions
3406 EC_GROUP_set_generator()
3407 EC_GROUP_get_generator()
3408 EC_GROUP_get_order()
3409 EC_GROUP_get_cofactor()
3410 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3411 to methods, which would lead to unnecessary code duplication when
3412 adding different types of curves.
3413 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3415 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3416 arithmetic, and such that modified wNAFs are generated
3417 (which avoid length expansion in many cases).
3420 *) Add a function EC_GROUP_check_discriminant() (defined via
3421 EC_METHOD) that verifies that the curve discriminant is non-zero.
3423 Add a function EC_GROUP_check() that makes some sanity tests
3424 on a EC_GROUP, its generator and order. This includes
3425 EC_GROUP_check_discriminant().
3426 [Nils Larsch <nla@trustcenter.de>]
3428 *) Add ECDSA in new directory crypto/ecdsa/.
3430 Add applications 'openssl ecparam' and 'openssl ecdsa'
3431 (these are based on 'openssl dsaparam' and 'openssl dsa').
3433 ECDSA support is also included in various other files across the
3434 library. Most notably,
3435 - 'openssl req' now has a '-newkey ecdsa:file' option;
3436 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3437 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3438 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3439 them suitable for ECDSA where domain parameters must be
3440 extracted before the specific public key;
3441 - ECDSA engine support has been added.
3442 [Nils Larsch <nla@trustcenter.de>]
3444 *) Include some named elliptic curves, and add OIDs from X9.62,
3445 SECG, and WAP/WTLS. Each curve can be obtained from the new
3447 EC_GROUP_new_by_curve_name(),
3448 and the list of available named curves can be obtained with
3449 EC_get_builtin_curves().
3450 Also add a 'curve_name' member to EC_GROUP objects, which can be
3452 EC_GROUP_set_curve_name()
3453 EC_GROUP_get_curve_name()
3454 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3456 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3457 was actually never needed) and in BN_mul(). The removal in BN_mul()
3458 required a small change in bn_mul_part_recursive() and the addition
3459 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3460 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3461 bn_sub_words() and bn_add_words() except they take arrays with
3465 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3467 *) Cleanse PEM buffers before freeing them since they may contain
3469 [Benjamin Bennett <ben@psc.edu>]
3471 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3472 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3473 authentication-only ciphersuites.
3476 *) Since AES128 and AES256 share a single mask bit in the logic of
3477 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3478 kludge to work properly if AES128 is available and AES256 isn't.
3481 *) Expand security boundary to match 1.1.1 module.
3484 *) Remove redundant features: hash file source, editing of test vectors
3485 modify fipsld to use external fips_premain.c signature.
3488 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3489 run algorithm test programs.
3492 *) Make algorithm test programs more tolerant of whitespace.
3495 *) Have SSL/TLS server implementation tolerate "mismatched" record
3496 protocol version while receiving ClientHello even if the
3497 ClientHello is fragmented. (The server can't insist on the
3498 particular protocol version it has chosen before the ServerHello
3499 message has informed the client about his choice.)
3502 *) Load error codes if they are not already present instead of using a
3503 static variable. This allows them to be cleanly unloaded and reloaded.
3506 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3508 *) Introduce limits to prevent malicious keys being able to
3509 cause a denial of service. (CVE-2006-2940)
3510 [Steve Henson, Bodo Moeller]
3512 *) Fix ASN.1 parsing of certain invalid structures that can result
3513 in a denial of service. (CVE-2006-2937) [Steve Henson]
3515 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3516 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3518 *) Fix SSL client code which could crash if connecting to a
3519 malicious SSLv2 server. (CVE-2006-4343)
3520 [Tavis Ormandy and Will Drewry, Google Security Team]
3522 *) Change ciphersuite string processing so that an explicit
3523 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3524 will no longer include "AES128-SHA"), and any other similar
3525 ciphersuite (same bitmap) from *other* protocol versions (so that
3526 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3527 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3528 changes from 0.9.8b and 0.9.8d.
3531 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3533 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3534 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3536 *) Change the Unix randomness entropy gathering to use poll() when
3537 possible instead of select(), since the latter has some
3538 undesirable limitations.
3539 [Darryl Miles via Richard Levitte and Bodo Moeller]
3541 *) Disable rogue ciphersuites:
3543 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3544 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3545 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3547 The latter two were purportedly from
3548 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3551 Also deactive the remaining ciphersuites from
3552 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3553 unofficial, and the ID has long expired.
3556 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3557 dual-core machines) and other potential thread-safety issues.
3560 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3562 *) Adapt fipsld and the build system to link against the validated FIPS
3563 module in FIPS mode.
3566 *) Fixes for VC++ 2005 build under Windows.
3569 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3570 from a Windows bash shell such as MSYS. It is autodetected from the
3571 "config" script when run from a VC++ environment. Modify standard VC++
3572 build to use fipscanister.o from the GNU make build.
3575 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3577 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3578 The value now differs depending on if you build for FIPS or not.
3579 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3580 safely run with a non-FIPSed libcrypto, as it may crash because of
3581 the difference induced by this change.
3584 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3586 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3587 (part of SSL_OP_ALL). This option used to disable the
3588 countermeasure against man-in-the-middle protocol-version
3589 rollback in the SSL 2.0 server implementation, which is a bad
3590 idea. (CVE-2005-2969)
3592 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3593 for Information Security, National Institute of Advanced Industrial
3594 Science and Technology [AIST], Japan)]
3596 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3597 mainly for FIPS compliance and not fully integrated at this stage.
3600 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3601 the exponentiation using a fixed-length exponent. (Otherwise,
3602 the information leaked through timing could expose the secret key
3603 after many signatures; cf. Bleichenbacher's attack on DSA with
3607 *) Make a new fixed-window mod_exp implementation the default for
3608 RSA, DSA, and DH private-key operations so that the sequence of
3609 squares and multiplies and the memory access pattern are
3610 independent of the particular secret key. This will mitigate
3611 cache-timing and potential related attacks.
3613 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3614 and this is automatically used by BN_mod_exp_mont() if the new flag
3615 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3616 will use this BN flag for private exponents unless the flag
3617 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3618 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3620 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3622 *) Change the client implementation for SSLv23_method() and
3623 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3624 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3625 (Previously, the SSL 2.0 backwards compatible Client Hello
3626 message format would be used even with SSL_OP_NO_SSLv2.)
3629 *) Add support for smime-type MIME parameter in S/MIME messages which some
3633 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3634 a threadsafe manner. Modify rsa code to use new function and add calls
3635 to dsa and dh code (which had race conditions before).
3638 *) Include the fixed error library code in the C error file definitions
3639 instead of fixing them up at runtime. This keeps the error code
3640 structures constant.
3643 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3645 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3648 *) Fixes for newer kerberos headers. NB: the casts are needed because
3649 the 'length' field is signed on one version and unsigned on another
3650 with no (?) obvious way to tell the difference, without these VC++
3651 complains. Also the "definition" of FAR (blank) is no longer included
3652 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3653 some needed definitions.
3656 *) Undo Cygwin change.
3659 *) Added support for proxy certificates according to RFC 3820.
3660 Because they may be a security thread to unaware applications,
3661 they must be explicitely allowed in run-time. See
3662 docs/HOWTO/proxy_certificates.txt for further information.
3665 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3667 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3668 server and client random values. Previously
3669 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3670 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3672 This change has negligible security impact because:
3674 1. Server and client random values still have 24 bytes of pseudo random
3677 2. Server and client random values are sent in the clear in the initial
3680 3. The master secret is derived using the premaster secret (48 bytes in
3681 size for static RSA ciphersuites) as well as client server and random
3684 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3687 [Stephen Henson, reported by UK NISCC]
3689 *) Use Windows randomness collection on Cygwin.
3692 *) Fix hang in EGD/PRNGD query when communication socket is closed
3693 prematurely by EGD/PRNGD.
3694 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3696 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3699 *) Back-port of selected performance improvements from development
3700 branch, as well as improved support for PowerPC platforms.
3703 *) Add lots of checks for memory allocation failure, error codes to indicate
3704 failure and freeing up memory if a failure occurs.
3705 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3707 *) Add new -passin argument to dgst.
3710 *) Perform some character comparisons of different types in X509_NAME_cmp:
3711 this is needed for some certificates that reencode DNs into UTF8Strings
3712 (in violation of RFC3280) and can't or wont issue name rollover
3716 *) Make an explicit check during certificate validation to see that
3717 the CA setting in each certificate on the chain is correct. As a
3718 side effect always do the following basic checks on extensions,
3719 not just when there's an associated purpose to the check:
3721 - if there is an unhandled critical extension (unless the user
3722 has chosen to ignore this fault)
3723 - if the path length has been exceeded (if one is set at all)
3724 - that certain extensions fit the associated purpose (if one has
3728 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3730 *) Avoid a race condition when CRLs are checked in a multi threaded
3731 environment. This would happen due to the reordering of the revoked
3732 entries during signature checking and serial number lookup. Now the
3733 encoding is cached and the serial number sort performed under a lock.
3734 Add new STACK function sk_is_sorted().
3737 *) Add Delta CRL to the extension code.
3740 *) Various fixes to s3_pkt.c so alerts are sent properly.
3741 [David Holmes <d.holmes@f5.com>]
3743 *) Reduce the chances of duplicate issuer name and serial numbers (in
3744 violation of RFC3280) using the OpenSSL certificate creation utilities.
3745 This is done by creating a random 64 bit value for the initial serial
3746 number when a serial number file is created or when a self signed
3747 certificate is created using 'openssl req -x509'. The initial serial
3748 number file is created using 'openssl x509 -next_serial' in CA.pl
3749 rather than being initialized to 1.
3752 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3754 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3755 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3756 [Joe Orton, Steve Henson]
3758 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3760 [Joe Orton, Steve Henson]
3762 *) Make it possible to have multiple active certificates with the same
3763 subject in the CA index file. This is done only if the keyword
3764 'unique_subject' is set to 'no' in the main CA section (default
3765 if 'CA_default') of the configuration file. The value is saved
3766 with the database itself in a separate index attribute file,
3767 named like the index file with '.attr' appended to the name.
3770 *) X509 verify fixes. Disable broken certificate workarounds when
3771 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3772 keyUsage extension present. Don't accept CRLs with unhandled critical
3773 extensions: since verify currently doesn't process CRL extensions this
3774 rejects a CRL with *any* critical extensions. Add new verify error codes
3778 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3779 A clarification of RFC2560 will require the use of OCTET STRINGs and
3780 some implementations cannot handle the current raw format. Since OpenSSL
3781 copies and compares OCSP nonces as opaque blobs without any attempt at
3782 parsing them this should not create any compatibility issues.
3785 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3786 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3787 this HMAC (and other) operations are several times slower than OpenSSL
3791 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3792 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3794 *) Use the correct content when signing type "other".
3797 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3799 *) Fix various bugs revealed by running the NISCC test suite:
3801 Stop out of bounds reads in the ASN1 code when presented with
3802 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3804 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3806 If verify callback ignores invalid public key errors don't try to check
3807 certificate signature with the NULL public key.
3811 *) New -ignore_err option in ocsp application to stop the server
3812 exiting on the first error in a request.
3815 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3816 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3820 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3821 extra data after the compression methods not only for TLS 1.0
3822 but also for SSL 3.0 (as required by the specification).
3823 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3825 *) Change X509_certificate_type() to mark the key as exported/exportable
3826 when it's 512 *bits* long, not 512 bytes.
3829 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3830 blocks during encryption.
3833 *) Various fixes to base64 BIO and non blocking I/O. On write
3834 flushes were not handled properly if the BIO retried. On read
3835 data was not being buffered properly and had various logic bugs.
3836 This also affects blocking I/O when the data being decoded is a
3840 *) Various S/MIME bugfixes and compatibility changes:
3841 output correct application/pkcs7 MIME type if
3842 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3843 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3844 of files as .eml work). Correctly handle very long lines in MIME
3848 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3850 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3851 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3852 a protocol version number mismatch like a decryption error
3853 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3856 *) Turn on RSA blinding by default in the default implementation
3857 to avoid a timing attack. Applications that don't want it can call
3858 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3859 They would be ill-advised to do so in most cases.
3860 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3862 *) Change RSA blinding code so that it works when the PRNG is not
3863 seeded (in this case, the secret RSA exponent is abused as
3864 an unpredictable seed -- if it is not unpredictable, there
3865 is no point in blinding anyway). Make RSA blinding thread-safe
3866 by remembering the creator's thread ID in rsa->blinding and
3867 having all other threads use local one-time blinding factors
3868 (this requires more computation than sharing rsa->blinding, but
3869 avoids excessive locking; and if an RSA object is not shared
3870 between threads, blinding will still be very fast).
3873 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
3874 ENGINE as defaults for all supported algorithms irrespective of
3875 the 'flags' parameter. 'flags' is now honoured, so applications
3876 should make sure they are passing it correctly.
3879 *) Target "mingw" now allows native Windows code to be generated in
3880 the Cygwin environment as well as with the MinGW compiler.
3883 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
3885 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3886 via timing by performing a MAC computation even if incorrrect
3887 block cipher padding has been found. This is a countermeasure
3888 against active attacks where the attacker has to distinguish
3889 between bad padding and a MAC verification error. (CVE-2003-0078)
3891 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3892 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3893 Martin Vuagnoux (EPFL, Ilion)]
3895 *) Make the no-err option work as intended. The intention with no-err
3896 is not to have the whole error stack handling routines removed from
3897 libcrypto, it's only intended to remove all the function name and
3898 reason texts, thereby removing some of the footprint that may not
3899 be interesting if those errors aren't displayed anyway.
3901 NOTE: it's still possible for any application or module to have it's
3902 own set of error texts inserted. The routines are there, just not
3903 used by default when no-err is given.
3906 *) Add support for FreeBSD on IA64.
3907 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
3909 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
3910 Kerberos function mit_des_cbc_cksum(). Before this change,
3911 the value returned by DES_cbc_cksum() was like the one from
3912 mit_des_cbc_cksum(), except the bytes were swapped.
3913 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
3915 *) Allow an application to disable the automatic SSL chain building.
3916 Before this a rather primitive chain build was always performed in
3917 ssl3_output_cert_chain(): an application had no way to send the
3918 correct chain if the automatic operation produced an incorrect result.
3920 Now the chain builder is disabled if either:
3922 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
3924 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
3926 The reasoning behind this is that an application would not want the
3927 auto chain building to take place if extra chain certificates are
3928 present and it might also want a means of sending no additional
3929 certificates (for example the chain has two certificates and the
3933 *) Add the possibility to build without the ENGINE framework.
3934 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3936 *) Under Win32 gmtime() can return NULL: check return value in
3937 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
3940 *) DSA routines: under certain error conditions uninitialized BN objects
3941 could be freed. Solution: make sure initialization is performed early
3942 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
3943 Nils Larsch <nla@trustcenter.de> via PR#459)
3946 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
3947 checked on reconnect on the client side, therefore session resumption
3948 could still fail with a "ssl session id is different" error. This
3949 behaviour is masked when SSL_OP_ALL is used due to
3950 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
3951 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3952 followup to PR #377.
3955 *) IA-32 assembler support enhancements: unified ELF targets, support
3956 for SCO/Caldera platforms, fix for Cygwin shared build.
3959 *) Add support for FreeBSD on sparc64. As a consequence, support for
3960 FreeBSD on non-x86 processors is separate from x86 processors on
3961 the config script, much like the NetBSD support.
3962 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
3964 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
3966 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
3969 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
3970 code (06) was taken as the first octet of the session ID and the last
3971 octet was ignored consequently. As a result SSLv2 client side session
3972 caching could not have worked due to the session ID mismatch between
3974 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3978 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
3979 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
3983 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
3984 seems that in spite of existing for more than a year, many application
3985 author have done nothing to provide the necessary callbacks, which
3986 means that this particular engine will not work properly anywhere.
3987 This is a very unfortunate situation which forces us, in the name
3988 of usability, to give the hw_ncipher.c a static lock, which is part
3990 NOTE: This is for the 0.9.7 series ONLY. This hack will never
3991 appear in 0.9.8 or later. We EXPECT application authors to have
3992 dealt properly with this when 0.9.8 is released (unless we actually
3993 make such changes in the libcrypto locking code that changes will
3994 have to be made anyway).
3997 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
3998 octets have been read, EOF or an error occurs. Without this change
3999 some truncated ASN1 structures will not produce an error.
4002 *) Disable Heimdal support, since it hasn't been fully implemented.
4003 Still give the possibility to force the use of Heimdal, but with
4004 warnings and a request that patches get sent to openssl-dev.
4007 *) Add the VC-CE target, introduce the WINCE sysname, and add
4008 INSTALL.WCE and appropriate conditionals to make it build.
4009 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4011 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4012 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4013 edit numbers of the version.
4014 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4016 *) Introduce safe string copy and catenation functions
4017 (BUF_strlcpy() and BUF_strlcat()).
4018 [Ben Laurie (CHATS) and Richard Levitte]
4020 *) Avoid using fixed-size buffers for one-line DNs.
4021 [Ben Laurie (CHATS)]
4023 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4024 resizing buffers containing secrets, and use where appropriate.
4025 [Ben Laurie (CHATS)]
4027 *) Avoid using fixed size buffers for configuration file location.
4028 [Ben Laurie (CHATS)]
4030 *) Avoid filename truncation for various CA files.
4031 [Ben Laurie (CHATS)]
4033 *) Use sizeof in preference to magic numbers.
4034 [Ben Laurie (CHATS)]
4036 *) Avoid filename truncation in cert requests.
4037 [Ben Laurie (CHATS)]
4039 *) Add assertions to check for (supposedly impossible) buffer
4041 [Ben Laurie (CHATS)]
4043 *) Don't cache truncated DNS entries in the local cache (this could
4044 potentially lead to a spoofing attack).
4045 [Ben Laurie (CHATS)]
4047 *) Fix various buffers to be large enough for hex/decimal
4048 representations in a platform independent manner.
4049 [Ben Laurie (CHATS)]
4051 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4052 resizing buffers containing secrets, and use where appropriate.
4053 [Ben Laurie (CHATS)]
4055 *) Add BIO_indent() to avoid much slightly worrying code to do
4057 [Ben Laurie (CHATS)]
4059 *) Convert sprintf()/BIO_puts() to BIO_printf().
4060 [Ben Laurie (CHATS)]
4062 *) buffer_gets() could terminate with the buffer only half
4064 [Ben Laurie (CHATS)]
4066 *) Add assertions to prevent user-supplied crypto functions from
4067 overflowing internal buffers by having large block sizes, etc.
4068 [Ben Laurie (CHATS)]
4070 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4072 [Ben Laurie (CHATS)]
4074 *) Eliminate unused copy of key in RC4.
4075 [Ben Laurie (CHATS)]
4077 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4078 [Ben Laurie (CHATS)]
4080 *) Fix off-by-one error in EGD path.
4081 [Ben Laurie (CHATS)]
4083 *) If RANDFILE path is too long, ignore instead of truncating.
4084 [Ben Laurie (CHATS)]
4086 *) Eliminate unused and incorrectly sized X.509 structure
4088 [Ben Laurie (CHATS)]
4090 *) Eliminate unused and dangerous function knumber().
4091 [Ben Laurie (CHATS)]
4093 *) Eliminate unused and dangerous structure, KSSL_ERR.
4094 [Ben Laurie (CHATS)]
4096 *) Protect against overlong session ID context length in an encoded
4097 session object. Since these are local, this does not appear to be
4099 [Ben Laurie (CHATS)]
4101 *) Change from security patch (see 0.9.6e below) that did not affect
4102 the 0.9.6 release series:
4104 Remote buffer overflow in SSL3 protocol - an attacker could
4105 supply an oversized master key in Kerberos-enabled versions.
4107 [Ben Laurie (CHATS)]
4109 *) Change the SSL kerb5 codes to match RFC 2712.
4112 *) Make -nameopt work fully for req and add -reqopt switch.
4113 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4115 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4116 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4118 *) Make sure tests can be performed even if the corresponding algorithms
4119 have been removed entirely. This was also the last step to make
4120 OpenSSL compilable with DJGPP under all reasonable conditions.
4121 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
4123 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
4124 to allow version independent disabling of normally unselected ciphers,
4125 which may be activated as a side-effect of selecting a single cipher.
4127 (E.g., cipher list string "RSA" enables ciphersuites that are left
4128 out of "ALL" because they do not provide symmetric encryption.
4129 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
4130 [Lutz Jaenicke, Bodo Moeller]
4132 *) Add appropriate support for separate platform-dependent build
4133 directories. The recommended way to make a platform-dependent
4134 build directory is the following (tested on Linux), maybe with
4137 # Place yourself outside of the OpenSSL source tree. In
4138 # this example, the environment variable OPENSSL_SOURCE
4139 # is assumed to contain the absolute OpenSSL source directory.
4140 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
4141 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
4142 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
4143 mkdir -p `dirname $F`
4144 ln -s $OPENSSL_SOURCE/$F $F
4147 To be absolutely sure not to disturb the source tree, a "make clean"
4148 is a good thing. If it isn't successfull, don't worry about it,
4149 it probably means the source directory is very clean.
4152 *) Make sure any ENGINE control commands make local copies of string
4153 pointers passed to them whenever necessary. Otherwise it is possible
4154 the caller may have overwritten (or deallocated) the original string
4155 data when a later ENGINE operation tries to use the stored values.
4156 [Götz Babin-Ebell <babinebell@trustcenter.de>]
4158 *) Improve diagnostics in file reading and command-line digests.
4159 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
4161 *) Add AES modes CFB and OFB to the object database. Correct an
4162 error in AES-CFB decryption.
4165 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
4166 allows existing EVP_CIPHER_CTX structures to be reused after
4167 calling EVP_*Final(). This behaviour is used by encryption
4168 BIOs and some applications. This has the side effect that
4169 applications must explicitly clean up cipher contexts with
4170 EVP_CIPHER_CTX_cleanup() or they will leak memory.
4173 *) Check the values of dna and dnb in bn_mul_recursive before calling
4174 bn_mul_comba (a non zero value means the a or b arrays do not contain
4175 n2 elements) and fallback to bn_mul_normal if either is not zero.
4178 *) Fix escaping of non-ASCII characters when using the -subj option
4179 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
4182 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
4183 form for "surname", serialNumber has no short form.
4184 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
4185 therefore remove "mail" short name for "internet 7".
4186 The OID for unique identifiers in X509 certificates is
4187 x500UniqueIdentifier, not uniqueIdentifier.
4188 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
4191 *) Add an "init" command to the ENGINE config module and auto initialize
4192 ENGINEs. Without any "init" command the ENGINE will be initialized
4193 after all ctrl commands have been executed on it. If init=1 the
4194 ENGINE is initailized at that point (ctrls before that point are run
4195 on the uninitialized ENGINE and after on the initialized one). If
4196 init=0 then the ENGINE will not be iniatialized at all.
4199 *) Fix the 'app_verify_callback' interface so that the user-defined
4200 argument is actually passed to the callback: In the
4201 SSL_CTX_set_cert_verify_callback() prototype, the callback
4202 declaration has been changed from
4205 int (*cb)(X509_STORE_CTX *,void *);
4206 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
4207 i=s->ctx->app_verify_callback(&ctx)
4208 has been changed into
4209 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
4211 To update applications using SSL_CTX_set_cert_verify_callback(),
4212 a dummy argument can be added to their callback functions.
4213 [D. K. Smetters <smetters@parc.xerox.com>]
4215 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
4216 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
4218 *) Add and OPENSSL_LOAD_CONF define which will cause
4219 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
4220 This allows older applications to transparently support certain
4221 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
4222 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
4223 load the config file and OPENSSL_add_all_algorithms_conf() which will
4224 always load it have also been added.
4227 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
4228 Adjust NIDs and EVP layer.
4229 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4231 *) Config modules support in openssl utility.
4233 Most commands now load modules from the config file,
4234 though in a few (such as version) this isn't done
4235 because it couldn't be used for anything.
4237 In the case of ca and req the config file used is
4238 the same as the utility itself: that is the -config
4239 command line option can be used to specify an
4243 *) Move default behaviour from OPENSSL_config(). If appname is NULL
4244 use "openssl_conf" if filename is NULL use default openssl config file.
4247 *) Add an argument to OPENSSL_config() to allow the use of an alternative
4248 config section name. Add a new flag to tolerate a missing config file
4249 and move code to CONF_modules_load_file().
4252 *) Support for crypto accelerator cards from Accelerated Encryption
4253 Processing, www.aep.ie. (Use engine 'aep')
4254 The support was copied from 0.9.6c [engine] and adapted/corrected
4255 to work with the new engine framework.
4256 [AEP Inc. and Richard Levitte]
4258 *) Support for SureWare crypto accelerator cards from Baltimore
4259 Technologies. (Use engine 'sureware')
4260 The support was copied from 0.9.6c [engine] and adapted
4261 to work with the new engine framework.
4264 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
4265 make the newer ENGINE framework commands for the CHIL engine work.
4266 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
4268 *) Make it possible to produce shared libraries on ReliantUNIX.
4269 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
4271 *) Add the configuration target debug-linux-ppro.
4272 Make 'openssl rsa' use the general key loading routines
4273 implemented in apps.c, and make those routines able to
4274 handle the key format FORMAT_NETSCAPE and the variant
4276 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4278 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4279 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4281 *) Add -keyform to rsautl, and document -engine.
4282 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
4284 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
4285 BIO_R_NO_SUCH_FILE error code rather than the generic
4286 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
4289 *) Add new functions
4291 ERR_peek_last_error_line
4292 ERR_peek_last_error_line_data.
4293 These are similar to
4296 ERR_peek_error_line_data,
4297 but report on the latest error recorded rather than the first one
4298 still in the error queue.
4299 [Ben Laurie, Bodo Moeller]
4301 *) default_algorithms option in ENGINE config module. This allows things
4303 default_algorithms = ALL
4304 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4307 *) Prelminary ENGINE config module.
4310 *) New experimental application configuration code.
4313 *) Change the AES code to follow the same name structure as all other
4314 symmetric ciphers, and behave the same way. Move everything to
4315 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4316 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4318 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4319 [Ben Laurie and Theo de Raadt]
4321 *) Add option to output public keys in req command.
4322 [Massimiliano Pala madwolf@openca.org]
4324 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4325 (up to about 10% better than before for P-192 and P-224).
4328 *) New functions/macros
4330 SSL_CTX_set_msg_callback(ctx, cb)
4331 SSL_CTX_set_msg_callback_arg(ctx, arg)
4332 SSL_set_msg_callback(ssl, cb)
4333 SSL_set_msg_callback_arg(ssl, arg)
4335 to request calling a callback function
4337 void cb(int write_p, int version, int content_type,
4338 const void *buf, size_t len, SSL *ssl, void *arg)
4340 whenever a protocol message has been completely received
4341 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4342 protocol version according to which the SSL library interprets
4343 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4344 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4345 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4346 specification (change_cipher_spec(20), alert(21), handshake(22)).
4347 'buf' and 'len' point to the actual message, 'ssl' to the
4348 SSL object, and 'arg' is the application-defined value set by
4349 SSL[_CTX]_set_msg_callback_arg().
4351 'openssl s_client' and 'openssl s_server' have new '-msg' options
4352 to enable a callback that displays all protocol messages.
4355 *) Change the shared library support so shared libraries are built as
4356 soon as the corresponding static library is finished, and thereby get
4357 openssl and the test programs linked against the shared library.
4358 This still only happens when the keyword "shard" has been given to
4359 the configuration scripts.
4361 NOTE: shared library support is still an experimental thing, and
4362 backward binary compatibility is still not guaranteed.
4363 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4365 *) Add support for Subject Information Access extension.
4366 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4368 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4369 additional bytes when new memory had to be allocated, not just
4370 when reusing an existing buffer.
4373 *) New command line and configuration option 'utf8' for the req command.
4374 This allows field values to be specified as UTF8 strings.
4377 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4378 runs for the former and machine-readable output for the latter.
4381 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4382 of the e-mail address in the DN (i.e., it will go into a certificate
4383 extension only). The new configuration file option 'email_in_dn = no'
4384 has the same effect.
4385 [Massimiliano Pala madwolf@openca.org]
4387 *) Change all functions with names starting with des_ to be starting
4388 with DES_ instead. Add wrappers that are compatible with libdes,
4389 but are named _ossl_old_des_*. Finally, add macros that map the
4390 des_* symbols to the corresponding _ossl_old_des_* if libdes
4391 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4392 desired, the des_* symbols will be mapped to DES_*, with one
4395 Since we provide two compatibility mappings, the user needs to
4396 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4397 compatibility is desired. The default (i.e., when that macro
4398 isn't defined) is OpenSSL 0.9.6c compatibility.
4400 There are also macros that enable and disable the support of old
4401 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4402 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4403 are defined, the default will apply: to support the old des routines.
4405 In either case, one must include openssl/des.h to get the correct
4406 definitions. Do not try to just include openssl/des_old.h, that
4409 NOTE: This is a major break of an old API into a new one. Software
4410 authors are encouraged to switch to the DES_ style functions. Some
4411 time in the future, des_old.h and the libdes compatibility functions
4412 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4413 default), and then completely removed.
4416 *) Test for certificates which contain unsupported critical extensions.
4417 If such a certificate is found during a verify operation it is
4418 rejected by default: this behaviour can be overridden by either
4419 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4420 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4421 X509_supported_extension() has also been added which returns 1 if a
4422 particular extension is supported.
4425 *) Modify the behaviour of EVP cipher functions in similar way to digests
4426 to retain compatibility with existing code.
4429 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4430 compatibility with existing code. In particular the 'ctx' parameter does
4431 not have to be to be initialized before the call to EVP_DigestInit() and
4432 it is tidied up after a call to EVP_DigestFinal(). New function
4433 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4434 EVP_MD_CTX_copy() changed to not require the destination to be
4435 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4436 requires the destination to be valid.
4438 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4439 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4442 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4443 so that complete 'Handshake' protocol structures are kept in memory
4444 instead of overwriting 'msg_type' and 'length' with 'body' data.
4447 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4448 [Massimo Santin via Richard Levitte]
4450 *) Major restructuring to the underlying ENGINE code. This includes
4451 reduction of linker bloat, separation of pure "ENGINE" manipulation
4452 (initialisation, etc) from functionality dealing with implementations
4453 of specific crypto iterfaces. This change also introduces integrated
4454 support for symmetric ciphers and digest implementations - so ENGINEs
4455 can now accelerate these by providing EVP_CIPHER and EVP_MD
4456 implementations of their own. This is detailed in crypto/engine/README
4457 as it couldn't be adequately described here. However, there are a few
4458 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4459 were changed in the original introduction of ENGINE code have now
4460 reverted back - the hooking from this code to ENGINE is now a good
4461 deal more passive and at run-time, operations deal directly with
4462 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4463 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4464 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4465 they were not being used by the framework as there is no concept of a
4466 BIGNUM_METHOD and they could not be generalised to the new
4467 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4468 ENGINE_cpy() has been removed as it cannot be consistently defined in
4472 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4475 *) Change mkdef.pl to sort symbols that get the same entry number,
4476 and make sure the automatically generated functions ERR_load_*
4477 become part of libeay.num as well.
4480 *) New function SSL_renegotiate_pending(). This returns true once
4481 renegotiation has been requested (either SSL_renegotiate() call
4482 or HelloRequest/ClientHello receveived from the peer) and becomes
4483 false once a handshake has been completed.
4484 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4485 sends a HelloRequest, but does not ensure that a handshake takes
4486 place. SSL_renegotiate_pending() is useful for checking if the
4487 client has followed the request.)
4490 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4491 By default, clients may request session resumption even during
4492 renegotiation (if session ID contexts permit); with this option,
4493 session resumption is possible only in the first handshake.
4495 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4496 more bits available for options that should not be part of
4497 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4500 *) Add some demos for certificate and certificate request creation.
4503 *) Make maximum certificate chain size accepted from the peer application
4504 settable (SSL*_get/set_max_cert_list()), as proposed by
4505 "Douglas E. Engert" <deengert@anl.gov>.
4508 *) Add support for shared libraries for Unixware-7
4509 (Boyd Lynn Gerber <gerberb@zenez.com>).
4512 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4513 be done prior to destruction. Use this to unload error strings from
4514 ENGINEs that load their own error strings. NB: This adds two new API
4515 functions to "get" and "set" this destroy handler in an ENGINE.
4518 *) Alter all existing ENGINE implementations (except "openssl" and
4519 "openbsd") to dynamically instantiate their own error strings. This
4520 makes them more flexible to be built both as statically-linked ENGINEs
4521 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4522 Also, add stub code to each that makes building them as self-contained
4523 shared-libraries easier (see README.ENGINE).
4526 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4527 implementations into applications that are completely implemented in
4528 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4529 commands that can be used to configure what shared-library to load and
4530 to control aspects of the way it is handled. Also, made an update to
4531 the README.ENGINE file that brings its information up-to-date and
4532 provides some information and instructions on the "dynamic" ENGINE
4533 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4536 *) Make it possible to unload ranges of ERR strings with a new
4537 "ERR_unload_strings" function.
4540 *) Add a copy() function to EVP_MD.
4543 *) Make EVP_MD routines take a context pointer instead of just the
4544 md_data void pointer.
4547 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4548 that the digest can only process a single chunk of data
4549 (typically because it is provided by a piece of
4550 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4551 is only going to provide a single chunk of data, and hence the
4552 framework needn't accumulate the data for oneshot drivers.
4555 *) As with "ERR", make it possible to replace the underlying "ex_data"
4556 functions. This change also alters the storage and management of global
4557 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4558 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4559 index counters. The API functions that use this state have been changed
4560 to take a "class_index" rather than pointers to the class's local STACK
4561 and counter, and there is now an API function to dynamically create new
4562 classes. This centralisation allows us to (a) plug a lot of the
4563 thread-safety problems that existed, and (b) makes it possible to clean
4564 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4565 such data would previously have always leaked in application code and
4566 workarounds were in place to make the memory debugging turn a blind eye
4567 to it. Application code that doesn't use this new function will still
4568 leak as before, but their memory debugging output will announce it now
4569 rather than letting it slide.
4571 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4572 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4573 has a return value to indicate success or failure.
4576 *) Make it possible to replace the underlying "ERR" functions such that the
4577 global state (2 LHASH tables and 2 locks) is only used by the "default"
4578 implementation. This change also adds two functions to "get" and "set"
4579 the implementation prior to it being automatically set the first time
4580 any other ERR function takes place. Ie. an application can call "get",
4581 pass the return value to a module it has just loaded, and that module
4582 can call its own "set" function using that value. This means the
4583 module's "ERR" operations will use (and modify) the error state in the
4584 application and not in its own statically linked copy of OpenSSL code.
4587 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4588 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4589 the operation, and provides a more encapsulated way for external code
4590 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4591 to use these functions rather than manually incrementing the counts.
4593 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4596 *) Add EVP test program.
4599 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4602 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4603 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4604 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4605 These allow a CRL to be built without having to access X509_CRL fields
4606 directly. Modify 'ca' application to use new functions.
4609 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4610 bug workarounds. Rollback attack detection is a security feature.
4611 The problem will only arise on OpenSSL servers when TLSv1 is not
4612 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4613 Software authors not wanting to support TLSv1 will have special reasons
4614 for their choice and can explicitly enable this option.
4615 [Bodo Moeller, Lutz Jaenicke]
4617 *) Rationalise EVP so it can be extended: don't include a union of
4618 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4619 (similar to those existing for EVP_CIPHER_CTX).
4624 EVP_MD_CTX_init(&md); /* new function call */
4625 EVP_DigestInit(&md, EVP_sha1());
4626 EVP_DigestUpdate(&md, in, len);
4627 EVP_DigestFinal(&md, out, NULL);
4628 EVP_MD_CTX_cleanup(&md); /* new function call */
4632 *) Make DES key schedule conform to the usual scheme, as well as
4633 correcting its structure. This means that calls to DES functions
4634 now have to pass a pointer to a des_key_schedule instead of a
4635 plain des_key_schedule (which was actually always a pointer
4638 des_key_schedule ks;
4640 des_set_key_checked(..., &ks);
4641 des_ncbc_encrypt(..., &ks, ...);
4643 (Note that a later change renames 'des_...' into 'DES_...'.)
4646 *) Initial reduction of linker bloat: the use of some functions, such as
4647 PEM causes large amounts of unused functions to be linked in due to
4648 poor organisation. For example pem_all.c contains every PEM function
4649 which has a knock on effect of linking in large amounts of (unused)
4650 ASN1 code. Grouping together similar functions and splitting unrelated
4651 functions prevents this.
4654 *) Cleanup of EVP macros.
4657 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4658 correct _ecb suffix.
4661 *) Add initial OCSP responder support to ocsp application. The
4662 revocation information is handled using the text based index
4663 use by the ca application. The responder can either handle
4664 requests generated internally, supplied in files (for example
4665 via a CGI script) or using an internal minimal server.
4668 *) Add configuration choices to get zlib compression for TLS.
4671 *) Changes to Kerberos SSL for RFC 2712 compliance:
4672 1. Implemented real KerberosWrapper, instead of just using
4673 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4674 2. Implemented optional authenticator field of KerberosWrapper.
4676 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4677 and authenticator structs; see crypto/krb5/.
4679 Generalized Kerberos calls to support multiple Kerberos libraries.
4680 [Vern Staats <staatsvr@asc.hpc.mil>,
4681 Jeffrey Altman <jaltman@columbia.edu>
4682 via Richard Levitte]
4684 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4685 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4686 values for each of the key sizes rather than having just
4687 parameters (and 'speed' generating keys each time).
4690 *) Speed up EVP routines.
4693 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4694 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4695 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4696 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4698 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4699 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4700 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4703 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4705 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4708 *) Added the OS2-EMX target.
4709 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4711 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4712 to support NCONF routines in extension code. New function CONF_set_nconf()
4713 to allow functions which take an NCONF to also handle the old LHASH
4714 structure: this means that the old CONF compatible routines can be
4715 retained (in particular wrt extensions) without having to duplicate the
4716 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4719 *) Enhance the general user interface with mechanisms for inner control
4720 and with possibilities to have yes/no kind of prompts.
4723 *) Change all calls to low level digest routines in the library and
4724 applications to use EVP. Add missing calls to HMAC_cleanup() and
4725 don't assume HMAC_CTX can be copied using memcpy().
4726 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4728 *) Add the possibility to control engines through control names but with
4729 arbitrary arguments instead of just a string.
4730 Change the key loaders to take a UI_METHOD instead of a callback
4731 function pointer. NOTE: this breaks binary compatibility with earlier
4732 versions of OpenSSL [engine].
4733 Adapt the nCipher code for these new conditions and add a card insertion
4737 *) Enhance the general user interface with mechanisms to better support
4738 dialog box interfaces, application-defined prompts, the possibility
4739 to use defaults (for example default passwords from somewhere else)
4740 and interrupts/cancellations.
4743 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4744 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4747 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4748 tidy up some unnecessarily weird code in 'sk_new()').
4749 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4751 *) Change the key loading routines for ENGINEs to use the same kind
4752 callback (pem_password_cb) as all other routines that need this
4756 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4757 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4758 than this minimum value is recommended.
4761 *) New random seeder for OpenVMS, using the system process statistics
4762 that are easily reachable.
4765 *) Windows apparently can't transparently handle global
4766 variables defined in DLLs. Initialisations such as:
4768 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4770 wont compile. This is used by the any applications that need to
4771 declare their own ASN1 modules. This was fixed by adding the option
4772 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4773 needed for static libraries under Win32.
4776 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4777 setting of purpose and trust fields. New X509_STORE trust and
4778 purpose functions and tidy up setting in other SSL functions.
4781 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4782 structure. These are inherited by X509_STORE_CTX when it is
4783 initialised. This allows various defaults to be set in the
4784 X509_STORE structure (such as flags for CRL checking and custom
4785 purpose or trust settings) for functions which only use X509_STORE_CTX
4786 internally such as S/MIME.
4788 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4789 trust settings if they are not set in X509_STORE. This allows X509_STORE
4790 purposes and trust (in S/MIME for example) to override any set by default.
4792 Add command line options for CRL checking to smime, s_client and s_server
4796 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4797 are set then the CRL is looked up in the X509_STORE structure and
4798 its validity and signature checked, then if the certificate is found
4799 in the CRL the verify fails with a revoked error.
4801 Various new CRL related callbacks added to X509_STORE_CTX structure.
4803 Command line options added to 'verify' application to support this.
4805 This needs some additional work, such as being able to handle multiple
4806 CRLs with different times, extension based lookup (rather than just
4807 by subject name) and ultimately more complete V2 CRL extension
4811 *) Add a general user interface API (crypto/ui/). This is designed
4812 to replace things like des_read_password and friends (backward
4813 compatibility functions using this new API are provided).
4814 The purpose is to remove prompting functions from the DES code
4815 section as well as provide for prompting through dialog boxes in
4816 a window system and the like.
4819 *) Add "ex_data" support to ENGINE so implementations can add state at a
4820 per-structure level rather than having to store it globally.
4823 *) Make it possible for ENGINE structures to be copied when retrieved by
4824 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4825 This causes the "original" ENGINE structure to act like a template,
4826 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4827 operational state can be localised to each ENGINE structure, despite the
4828 fact they all share the same "methods". New ENGINE structures returned in
4829 this case have no functional references and the return value is the single
4830 structural reference. This matches the single structural reference returned
4831 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4835 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4836 needs to match any other type at all we need to manually clear the
4840 *) Changes to the "openssl engine" utility to include;
4841 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4842 about an ENGINE's available control commands.
4843 - executing control commands from command line arguments using the
4844 '-pre' and '-post' switches. '-post' is only used if '-t' is
4845 specified and the ENGINE is successfully initialised. The syntax for
4846 the individual commands are colon-separated, for example;
4847 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4850 *) New dynamic control command support for ENGINEs. ENGINEs can now
4851 declare their own commands (numbers), names (strings), descriptions,
4852 and input types for run-time discovery by calling applications. A
4853 subset of these commands are implicitly classed as "executable"
4854 depending on their input type, and only these can be invoked through
4855 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4856 can be based on user input, config files, etc). The distinction is
4857 that "executable" commands cannot return anything other than a boolean
4858 result and can only support numeric or string input, whereas some
4859 discoverable commands may only be for direct use through
4860 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4861 pointers, or other custom uses. The "executable" commands are to
4862 support parameterisations of ENGINE behaviour that can be
4863 unambiguously defined by ENGINEs and used consistently across any
4864 OpenSSL-based application. Commands have been added to all the
4865 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4866 control over shared-library paths without source code alterations.
4869 *) Changed all ENGINE implementations to dynamically allocate their
4870 ENGINEs rather than declaring them statically. Apart from this being
4871 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
4872 this also allows the implementations to compile without using the
4873 internal engine_int.h header.
4876 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
4877 'const' value. Any code that should be able to modify a RAND_METHOD
4878 should already have non-const pointers to it (ie. they should only
4879 modify their own ones).
4882 *) Made a variety of little tweaks to the ENGINE code.
4883 - "atalla" and "ubsec" string definitions were moved from header files
4884 to C code. "nuron" string definitions were placed in variables
4885 rather than hard-coded - allowing parameterisation of these values
4886 later on via ctrl() commands.
4887 - Removed unused "#if 0"'d code.
4888 - Fixed engine list iteration code so it uses ENGINE_free() to release
4889 structural references.
4890 - Constified the RAND_METHOD element of ENGINE structures.
4891 - Constified various get/set functions as appropriate and added
4892 missing functions (including a catch-all ENGINE_cpy that duplicates
4893 all ENGINE values onto a new ENGINE except reference counts/state).
4894 - Removed NULL parameter checks in get/set functions. Setting a method
4895 or function to NULL is a way of cancelling out a previously set
4896 value. Passing a NULL ENGINE parameter is just plain stupid anyway
4897 and doesn't justify the extra error symbols and code.
4898 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
4899 flags from engine_int.h to engine.h.
4900 - Changed prototypes for ENGINE handler functions (init(), finish(),
4901 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
4904 *) Implement binary inversion algorithm for BN_mod_inverse in addition
4905 to the algorithm using long division. The binary algorithm can be
4906 used only if the modulus is odd. On 32-bit systems, it is faster
4907 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
4908 roughly 5-15% for 256-bit moduli), so we use it only for moduli
4909 up to 450 bits. In 64-bit environments, the binary algorithm
4910 appears to be advantageous for much longer moduli; here we use it
4911 for moduli up to 2048 bits.
4914 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
4915 could not support the combine flag in choice fields.
4918 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
4919 extensions from a certificate request to the certificate.
4922 *) Allow multiple 'certopt' and 'nameopt' options to be separated
4923 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
4924 file: this allows the display of the certificate about to be
4925 signed to be customised, to allow certain fields to be included
4926 or excluded and extension details. The old system didn't display
4927 multicharacter strings properly, omitted fields not in the policy
4928 and couldn't display additional details such as extensions.
4931 *) Function EC_POINTs_mul for multiple scalar multiplication
4932 of an arbitrary number of elliptic curve points
4933 \sum scalars[i]*points[i],
4934 optionally including the generator defined for the EC_GROUP:
4935 scalar*generator + \sum scalars[i]*points[i].
4937 EC_POINT_mul is a simple wrapper function for the typical case
4938 that the point list has just one item (besides the optional
4942 *) First EC_METHODs for curves over GF(p):
4944 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
4945 operations and provides various method functions that can also
4946 operate with faster implementations of modular arithmetic.
4948 EC_GFp_mont_method() reuses most functions that are part of
4949 EC_GFp_simple_method, but uses Montgomery arithmetic.
4951 [Bodo Moeller; point addition and point doubling
4952 implementation directly derived from source code provided by
4953 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
4955 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
4956 crypto/ec/ec_lib.c):
4958 Curves are EC_GROUP objects (with an optional group generator)
4959 based on EC_METHODs that are built into the library.
4961 Points are EC_POINT objects based on EC_GROUP objects.
4963 Most of the framework would be able to handle curves over arbitrary
4964 finite fields, but as there are no obvious types for fields other
4965 than GF(p), some functions are limited to that for now.
4968 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
4969 that the file contains a complete HTTP response.
4972 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
4973 change the def and num file printf format specifier from "%-40sXXX"
4974 to "%-39s XXX". The latter will always guarantee a space after the
4975 field while the former will cause them to run together if the field
4976 is 40 of more characters long.
4979 *) Constify the cipher and digest 'method' functions and structures
4980 and modify related functions to take constant EVP_MD and EVP_CIPHER
4984 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
4985 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
4988 *) Modify EVP_Digest*() routines so they now return values. Although the
4989 internal software routines can never fail additional hardware versions
4993 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
4995 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
4996 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
4999 ERR_R_NESTED_ASN1_ERROR
5001 ERR_R_MISSING_ASN1_EOS
5002 were 4 .. 9, conflicting with
5003 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5005 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5006 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5008 Add new error code 'ERR_R_INTERNAL_ERROR'.
5011 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5015 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5016 sets the subject name for a new request or supersedes the
5017 subject name in a given request. Formats that can be parsed are
5018 'CN=Some Name, OU=myOU, C=IT'
5020 'CN=Some Name/OU=myOU/C=IT'.
5022 Add options '-batch' and '-verbose' to 'openssl req'.
5023 [Massimiliano Pala <madwolf@hackmasters.net>]
5025 *) Introduce the possibility to access global variables through
5026 functions on platform were that's the best way to handle exporting
5027 global variables in shared libraries. To enable this functionality,
5028 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5029 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5030 is normally done by Configure or something similar).
5032 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5033 in the source file (foo.c) like this:
5035 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5036 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5038 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5039 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5041 OPENSSL_DECLARE_GLOBAL(int,foo);
5042 #define foo OPENSSL_GLOBAL_REF(foo)
5043 OPENSSL_DECLARE_GLOBAL(double,bar);
5044 #define bar OPENSSL_GLOBAL_REF(bar)
5046 The #defines are very important, and therefore so is including the
5047 header file everywhere where the defined globals are used.
5049 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5050 of ASN.1 items, but that structure is a bit different.
5052 The largest change is in util/mkdef.pl which has been enhanced with
5053 better and easier to understand logic to choose which symbols should
5054 go into the Windows .def files as well as a number of fixes and code
5055 cleanup (among others, algorithm keywords are now sorted
5056 lexicographically to avoid constant rewrites).
5059 *) In BN_div() keep a copy of the sign of 'num' before writing the
5060 result to 'rm' because if rm==num the value will be overwritten
5061 and produce the wrong result if 'num' is negative: this caused
5062 problems with BN_mod() and BN_nnmod().
5065 *) Function OCSP_request_verify(). This checks the signature on an
5066 OCSP request and verifies the signer certificate. The signer
5067 certificate is just checked for a generic purpose and OCSP request
5071 *) Add OCSP_check_validity() function to check the validity of OCSP
5072 responses. OCSP responses are prepared in real time and may only
5073 be a few seconds old. Simply checking that the current time lies
5074 between thisUpdate and nextUpdate max reject otherwise valid responses
5075 caused by either OCSP responder or client clock inaccuracy. Instead
5076 we allow thisUpdate and nextUpdate to fall within a certain period of
5077 the current time. The age of the response can also optionally be
5078 checked. Two new options -validity_period and -status_age added to
5082 *) If signature or public key algorithm is unrecognized print out its
5083 OID rather that just UNKNOWN.
5086 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5087 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5088 ID to be generated from the issuer certificate alone which can then be
5089 passed to OCSP_id_issuer_cmp().
5092 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5093 ASN1 modules to export functions returning ASN1_ITEM pointers
5094 instead of the ASN1_ITEM structures themselves. This adds several
5095 new macros which allow the underlying ASN1 function/structure to
5096 be accessed transparently. As a result code should not use ASN1_ITEM
5097 references directly (such as &X509_it) but instead use the relevant
5098 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5099 use of the new ASN1 code on platforms where exporting structures
5100 is problematical (for example in shared libraries) but exporting
5101 functions returning pointers to structures is not.
5104 *) Add support for overriding the generation of SSL/TLS session IDs.
5105 These callbacks can be registered either in an SSL_CTX or per SSL.
5106 The purpose of this is to allow applications to control, if they wish,
5107 the arbitrary values chosen for use as session IDs, particularly as it
5108 can be useful for session caching in multiple-server environments. A
5109 command-line switch for testing this (and any client code that wishes
5110 to use such a feature) has been added to "s_server".
5111 [Geoff Thorpe, Lutz Jaenicke]
5113 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5114 of the form '#if defined(...) || defined(...) || ...' and
5115 '#if !defined(...) && !defined(...) && ...'. This also avoids
5116 the growing number of special cases it was previously handling.
5119 *) Make all configuration macros available for application by making
5120 sure they are available in opensslconf.h, by giving them names starting
5121 with "OPENSSL_" to avoid conflicts with other packages and by making
5122 sure e_os2.h will cover all platform-specific cases together with
5124 Additionally, it is now possible to define configuration/platform-
5125 specific names (called "system identities"). In the C code, these
5126 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
5127 macro with the name beginning with "OPENSSL_SYS_", which is determined
5128 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
5132 *) New option -set_serial to 'req' and 'x509' this allows the serial
5133 number to use to be specified on the command line. Previously self
5134 signed certificates were hard coded with serial number 0 and the
5135 CA options of 'x509' had to use a serial number in a file which was
5139 *) New options to 'ca' utility to support V2 CRL entry extensions.
5140 Currently CRL reason, invalidity date and hold instruction are
5141 supported. Add new CRL extensions to V3 code and some new objects.
5144 *) New function EVP_CIPHER_CTX_set_padding() this is used to
5145 disable standard block padding (aka PKCS#5 padding) in the EVP
5146 API, which was previously mandatory. This means that the data is
5147 not padded in any way and so the total length much be a multiple
5148 of the block size, otherwise an error occurs.
5151 *) Initial (incomplete) OCSP SSL support.
5154 *) New function OCSP_parse_url(). This splits up a URL into its host,
5155 port and path components: primarily to parse OCSP URLs. New -url
5156 option to ocsp utility.
5159 *) New nonce behavior. The return value of OCSP_check_nonce() now
5160 reflects the various checks performed. Applications can decide
5161 whether to tolerate certain situations such as an absent nonce
5162 in a response when one was present in a request: the ocsp application
5163 just prints out a warning. New function OCSP_add1_basic_nonce()
5164 this is to allow responders to include a nonce in a response even if
5165 the request is nonce-less.
5168 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
5169 skipped when using openssl x509 multiple times on a single input file,
5170 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
5173 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
5174 set string type: to handle setting ASN1_TIME structures. Fix ca
5175 utility to correctly initialize revocation date of CRLs.
5178 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
5179 the clients preferred ciphersuites and rather use its own preferences.
5180 Should help to work around M$ SGC (Server Gated Cryptography) bug in
5181 Internet Explorer by ensuring unchanged hash method during stepup.
5182 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
5185 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
5186 to aes and add a new 'exist' option to print out symbols that don't
5190 *) Additional options to ocsp utility to allow flags to be set and
5191 additional certificates supplied.
5194 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
5195 OCSP client a number of certificate to only verify the response
5199 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
5200 handle the new API. Currently only ECB, CBC modes supported. Add new
5203 Add TLS AES ciphersuites as described in RFC3268, "Advanced
5204 Encryption Standard (AES) Ciphersuites for Transport Layer
5205 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
5206 not enabled by default and were not part of the "ALL" ciphersuite
5207 alias because they were not yet official; they could be
5208 explicitly requested by specifying the "AESdraft" ciphersuite
5209 group alias. In the final release of OpenSSL 0.9.7, the group
5210 alias is called "AES" and is part of "ALL".)
5211 [Ben Laurie, Steve Henson, Bodo Moeller]
5213 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
5214 request to response.
5217 *) Functions for OCSP responders. OCSP_request_onereq_count(),
5218 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
5219 extract information from a certificate request. OCSP_response_create()
5220 creates a response and optionally adds a basic response structure.
5221 OCSP_basic_add1_status() adds a complete single response to a basic
5222 response and returns the OCSP_SINGLERESP structure just added (to allow
5223 extensions to be included for example). OCSP_basic_add1_cert() adds a
5224 certificate to a basic response and OCSP_basic_sign() signs a basic
5225 response with various flags. New helper functions ASN1_TIME_check()
5226 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
5227 (converts ASN1_TIME to GeneralizedTime).
5230 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
5231 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
5232 structure from a certificate. X509_pubkey_digest() digests the public_key
5233 contents: this is used in various key identifiers.
5236 *) Make sk_sort() tolerate a NULL argument.
5237 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
5239 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
5240 passed by the function are trusted implicitly. If any of them signed the
5241 response then it is assumed to be valid and is not verified.
5244 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
5245 to data. This was previously part of the PKCS7 ASN1 code. This
5246 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
5247 [Steve Henson, reported by Kenneth R. Robinette
5248 <support@securenetterm.com>]
5250 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
5251 routines: without these tracing memory leaks is very painful.
5252 Fix leaks in PKCS12 and PKCS7 routines.
5255 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
5256 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
5257 effectively meant GeneralizedTime would never be used. Now it
5258 is initialised to -1 but X509_time_adj() now has to check the value
5259 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
5260 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
5261 [Steve Henson, reported by Kenneth R. Robinette
5262 <support@securenetterm.com>]
5264 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
5265 result in a zero length in the ASN1_INTEGER structure which was
5266 not consistent with the structure when d2i_ASN1_INTEGER() was used
5267 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
5268 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
5269 where it did not print out a minus for negative ASN1_INTEGER.
5272 *) Add summary printout to ocsp utility. The various functions which
5273 convert status values to strings have been renamed to:
5274 OCSP_response_status_str(), OCSP_cert_status_str() and
5275 OCSP_crl_reason_str() and are no longer static. New options
5276 to verify nonce values and to disable verification. OCSP response
5277 printout format cleaned up.
5280 *) Add additional OCSP certificate checks. These are those specified
5281 in RFC2560. This consists of two separate checks: the CA of the
5282 certificate being checked must either be the OCSP signer certificate
5283 or the issuer of the OCSP signer certificate. In the latter case the
5284 OCSP signer certificate must contain the OCSP signing extended key
5285 usage. This check is performed by attempting to match the OCSP
5286 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
5287 in the OCSP_CERTID structures of the response.
5290 *) Initial OCSP certificate verification added to OCSP_basic_verify()
5291 and related routines. This uses the standard OpenSSL certificate
5292 verify routines to perform initial checks (just CA validity) and
5293 to obtain the certificate chain. Then additional checks will be
5294 performed on the chain. Currently the root CA is checked to see
5295 if it is explicitly trusted for OCSP signing. This is used to set
5296 a root CA as a global signing root: that is any certificate that
5297 chains to that CA is an acceptable OCSP signing certificate.
5300 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5301 extensions from a separate configuration file.
5302 As when reading extensions from the main configuration file,
5303 the '-extensions ...' option may be used for specifying the
5305 [Massimiliano Pala <madwolf@comune.modena.it>]
5307 *) New OCSP utility. Allows OCSP requests to be generated or
5308 read. The request can be sent to a responder and the output
5309 parsed, outputed or printed in text form. Not complete yet:
5310 still needs to check the OCSP response validity.
5313 *) New subcommands for 'openssl ca':
5314 'openssl ca -status <serial>' prints the status of the cert with
5315 the given serial number (according to the index file).
5316 'openssl ca -updatedb' updates the expiry status of certificates
5318 [Massimiliano Pala <madwolf@comune.modena.it>]
5320 *) New '-newreq-nodes' command option to CA.pl. This is like
5321 '-newreq', but calls 'openssl req' with the '-nodes' option
5322 so that the resulting key is not encrypted.
5323 [Damien Miller <djm@mindrot.org>]
5325 *) New configuration for the GNU Hurd.
5326 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5328 *) Initial code to implement OCSP basic response verify. This
5329 is currently incomplete. Currently just finds the signer's
5330 certificate and verifies the signature on the response.
5333 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5334 value of OPENSSLDIR. This is available via the new '-d' option
5335 to 'openssl version', and is also included in 'openssl version -a'.
5338 *) Allowing defining memory allocation callbacks that will be given
5339 file name and line number information in additional arguments
5340 (a const char* and an int). The basic functionality remains, as
5341 well as the original possibility to just replace malloc(),
5342 realloc() and free() by functions that do not know about these
5343 additional arguments. To register and find out the current
5344 settings for extended allocation functions, the following
5345 functions are provided:
5347 CRYPTO_set_mem_ex_functions
5348 CRYPTO_set_locked_mem_ex_functions
5349 CRYPTO_get_mem_ex_functions
5350 CRYPTO_get_locked_mem_ex_functions
5352 These work the same way as CRYPTO_set_mem_functions and friends.
5353 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5354 extended allocation function is enabled.
5355 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5356 a conventional allocation function is enabled.
5357 [Richard Levitte, Bodo Moeller]
5359 *) Finish off removing the remaining LHASH function pointer casts.
5360 There should no longer be any prototype-casting required when using
5361 the LHASH abstraction, and any casts that remain are "bugs". See
5362 the callback types and macros at the head of lhash.h for details
5363 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5366 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5367 If /dev/[u]random devices are not available or do not return enough
5368 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5370 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5371 /etc/entropy will be queried once each in this sequence, quering stops
5372 when enough entropy was collected without querying more sockets.
5375 *) Change the Unix RAND_poll() variant to be able to poll several
5376 random devices, as specified by DEVRANDOM, until a sufficient amount
5377 of data has been collected. We spend at most 10 ms on each file
5378 (select timeout) and read in non-blocking mode. DEVRANDOM now
5379 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5380 (previously it was just the string "/dev/urandom"), so on typical
5381 platforms the 10 ms delay will never occur.
5382 Also separate out the Unix variant to its own file, rand_unix.c.
5383 For VMS, there's a currently-empty rand_vms.c.
5386 *) Move OCSP client related routines to ocsp_cl.c. These
5387 provide utility functions which an application needing
5388 to issue a request to an OCSP responder and analyse the
5389 response will typically need: as opposed to those which an
5390 OCSP responder itself would need which will be added later.
5392 OCSP_request_sign() signs an OCSP request with an API similar
5393 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5394 response. OCSP_response_get1_basic() extracts basic response
5395 from response. OCSP_resp_find_status(): finds and extracts status
5396 information from an OCSP_CERTID structure (which will be created
5397 when the request structure is built). These are built from lower
5398 level functions which work on OCSP_SINGLERESP structures but
5399 wont normally be used unless the application wishes to examine
5400 extensions in the OCSP response for example.
5402 Replace nonce routines with a pair of functions.
5403 OCSP_request_add1_nonce() adds a nonce value and optionally
5404 generates a random value. OCSP_check_nonce() checks the
5405 validity of the nonce in an OCSP response.
5408 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5409 This doesn't copy the supplied OCSP_CERTID and avoids the
5410 need to free up the newly created id. Change return type
5411 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5412 This can then be used to add extensions to the request.
5413 Deleted OCSP_request_new(), since most of its functionality
5414 is now in OCSP_REQUEST_new() (and the case insensitive name
5415 clash) apart from the ability to set the request name which
5416 will be added elsewhere.
5419 *) Update OCSP API. Remove obsolete extensions argument from
5420 various functions. Extensions are now handled using the new
5421 OCSP extension code. New simple OCSP HTTP function which
5422 can be used to send requests and parse the response.
5425 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5426 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5427 uses the special reorder version of SET OF to sort the attributes
5428 and reorder them to match the encoded order. This resolves a long
5429 standing problem: a verify on a PKCS7 structure just after signing
5430 it used to fail because the attribute order did not match the
5431 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5432 it uses the received order. This is necessary to tolerate some broken
5433 software that does not order SET OF. This is handled by encoding
5434 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5435 to produce the required SET OF.
5438 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5439 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5440 files to get correct declarations of the ASN.1 item variables.
5443 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5444 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5445 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5446 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5447 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5448 ASN1_ITEM and no wrapper functions.
5451 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5452 replace the old function pointer based I/O routines. Change most of
5453 the *_d2i_bio() and *_d2i_fp() functions to use these.
5456 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5457 lines, recognice more "algorithms" that can be deselected, and make
5458 it complain about algorithm deselection that isn't recognised.
5461 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5462 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5463 to use new functions. Add NO_ASN1_OLD which can be set to remove
5464 some old style ASN1 functions: this can be used to determine if old
5465 code will still work when these eventually go away.
5468 *) New extension functions for OCSP structures, these follow the
5469 same conventions as certificates and CRLs.
5472 *) New function X509V3_add1_i2d(). This automatically encodes and
5473 adds an extension. Its behaviour can be customised with various
5474 flags to append, replace or delete. Various wrappers added for
5475 certifcates and CRLs.
5478 *) Fix to avoid calling the underlying ASN1 print routine when
5479 an extension cannot be parsed. Correct a typo in the
5480 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5483 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5484 entries for variables.
5487 *) Add functionality to apps/openssl.c for detecting locking
5488 problems: As the program is single-threaded, all we have
5489 to do is register a locking callback using an array for
5490 storing which locks are currently held by the program.
5493 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5494 SSL_get_ex_data_X509_STORE_idx(), which is used in
5495 ssl_verify_cert_chain() and thus can be called at any time
5496 during TLS/SSL handshakes so that thread-safety is essential.
5497 Unfortunately, the ex_data design is not at all suited
5498 for multi-threaded use, so it probably should be abolished.
5501 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5502 [Broadcom, tweaked and integrated by Geoff Thorpe]
5504 *) Move common extension printing code to new function
5505 X509V3_print_extensions(). Reorganise OCSP print routines and
5506 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5509 *) New function X509_signature_print() to remove duplication in some
5513 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5514 set (this was treated exactly the same as SET OF previously). This
5515 is used to reorder the STACK representing the structure to match the
5516 encoding. This will be used to get round a problem where a PKCS7
5517 structure which was signed could not be verified because the STACK
5518 order did not reflect the encoded order.
5521 *) Reimplement the OCSP ASN1 module using the new code.
5524 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5525 for its ASN1 operations. The old style function pointers still exist
5526 for now but they will eventually go away.
5529 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5530 completely replaces the old ASN1 functionality with a table driven
5531 encoder and decoder which interprets an ASN1_ITEM structure describing
5532 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5533 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5534 has also been converted to the new form.
5537 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5538 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5539 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5540 for negative moduli.
5543 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5544 of not touching the result's sign bit.
5547 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5551 *) Changed the LHASH code to use prototypes for callbacks, and created
5552 macros to declare and implement thin (optionally static) functions
5553 that provide type-safety and avoid function pointer casting for the
5554 type-specific callbacks.
5557 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5559 [Veers Staats <staatsvr@asc.hpc.mil>,
5560 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5562 *) Reformat the FAQ so the different questions and answers can be divided
5563 in sections depending on the subject.
5566 *) Have the zlib compression code load ZLIB.DLL dynamically under
5570 *) New function BN_mod_sqrt for computing square roots modulo a prime
5571 (using the probabilistic Tonelli-Shanks algorithm unless
5572 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5573 be handled deterministically).
5574 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5576 *) Make BN_mod_inverse faster by explicitly handling small quotients
5577 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5578 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5581 *) New function BN_kronecker.
5584 *) Fix BN_gcd so that it works on negative inputs; the result is
5585 positive unless both parameters are zero.
5586 Previously something reasonably close to an infinite loop was
5587 possible because numbers could be growing instead of shrinking
5588 in the implementation of Euclid's algorithm.
5591 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5592 sign of the number in question.
5594 Fix BN_is_word(a,w) to work correctly for w == 0.
5596 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5597 because its test if the absolute value of 'a' equals 'w'.
5598 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5599 it exists mostly for use in the implementations of BN_is_zero(),
5600 BN_is_one(), and BN_is_word().
5603 *) New function BN_swap.
5606 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5607 the exponentiation functions are more likely to produce reasonable
5608 results on negative inputs.
5611 *) Change BN_mod_mul so that the result is always non-negative.
5612 Previously, it could be negative if one of the factors was negative;
5613 I don't think anyone really wanted that behaviour.
5616 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5617 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5618 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5619 and add new functions:
5628 BN_mod_lshift1_quick
5632 These functions always generate non-negative results.
5634 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5635 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5637 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5638 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5639 be reduced modulo m.
5640 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5643 The following entry accidentily appeared in the CHANGES file
5644 distributed with OpenSSL 0.9.7. The modifications described in
5645 it do *not* apply to OpenSSL 0.9.7.
5647 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5648 was actually never needed) and in BN_mul(). The removal in BN_mul()
5649 required a small change in bn_mul_part_recursive() and the addition
5650 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5651 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5652 bn_sub_words() and bn_add_words() except they take arrays with
5657 *) In 'openssl passwd', verify passwords read from the terminal
5658 unless the '-salt' option is used (which usually means that
5659 verification would just waste user's time since the resulting
5660 hash is going to be compared with some given password hash)
5661 or the new '-noverify' option is used.
5663 This is an incompatible change, but it does not affect
5664 non-interactive use of 'openssl passwd' (passwords on the command
5665 line, '-stdin' option, '-in ...' option) and thus should not
5669 *) Remove all references to RSAref, since there's no more need for it.
5672 *) Make DSO load along a path given through an environment variable
5673 (SHLIB_PATH) with shl_load().
5676 *) Constify the ENGINE code as a result of BIGNUM constification.
5677 Also constify the RSA code and most things related to it. In a
5678 few places, most notable in the depth of the ASN.1 code, ugly
5679 casts back to non-const were required (to be solved at a later
5683 *) Make it so the openssl application has all engines loaded by default.
5686 *) Constify the BIGNUM routines a little more.
5689 *) Add the following functions:
5691 ENGINE_load_cswift()
5693 ENGINE_load_atalla()
5695 ENGINE_load_builtin_engines()
5697 That way, an application can itself choose if external engines that
5698 are built-in in OpenSSL shall ever be used or not. The benefit is
5699 that applications won't have to be linked with libdl or other dso
5700 libraries unless it's really needed.
5702 Changed 'openssl engine' to load all engines on demand.
5703 Changed the engine header files to avoid the duplication of some
5704 declarations (they differed!).
5707 *) 'openssl engine' can now list capabilities.
5710 *) Better error reporting in 'openssl engine'.
5713 *) Never call load_dh_param(NULL) in s_server.
5716 *) Add engine application. It can currently list engines by name and
5717 identity, and test if they are actually available.
5720 *) Improve RPM specification file by forcing symbolic linking and making
5721 sure the installed documentation is also owned by root.root.
5722 [Damien Miller <djm@mindrot.org>]
5724 *) Give the OpenSSL applications more possibilities to make use of
5725 keys (public as well as private) handled by engines.
5728 *) Add OCSP code that comes from CertCo.
5731 *) Add VMS support for the Rijndael code.
5734 *) Added untested support for Nuron crypto accelerator.
5737 *) Add support for external cryptographic devices. This code was
5738 previously distributed separately as the "engine" branch.
5739 [Geoff Thorpe, Richard Levitte]
5741 *) Rework the filename-translation in the DSO code. It is now possible to
5742 have far greater control over how a "name" is turned into a filename
5743 depending on the operating environment and any oddities about the
5744 different shared library filenames on each system.
5747 *) Support threads on FreeBSD-elf in Configure.
5750 *) Fix for SHA1 assembly problem with MASM: it produces
5751 warnings about corrupt line number information when assembling
5752 with debugging information. This is caused by the overlapping
5754 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5757 NCONF_get_number() has no error checking at all. As a replacement,
5758 NCONF_get_number_e() is defined (_e for "error checking") and is
5759 promoted strongly. The old NCONF_get_number is kept around for
5760 binary backward compatibility.
5761 Make it possible for methods to load from something other than a BIO,
5762 by providing a function pointer that is given a name instead of a BIO.
5763 For example, this could be used to load configuration data from an
5767 *) Fix for non blocking accept BIOs. Added new I/O special reason
5768 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5769 with non blocking I/O was not possible because no retry code was
5770 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5774 *) Added the beginnings of Rijndael support.
5777 *) Fix for bug in DirectoryString mask setting. Add support for
5778 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5779 to allow certificate printing to more controllable, additional
5780 'certopt' option to 'x509' to allow new printing options to be
5784 *) Clean old EAY MD5 hack from e_os.h.
5787 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5789 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5790 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5791 [Joe Orton, Steve Henson]
5793 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5795 *) Fix additional bug revealed by the NISCC test suite:
5797 Stop bug triggering large recursion when presented with
5798 certain ASN.1 tags (CVE-2003-0851)
5801 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5803 *) Fix various bugs revealed by running the NISCC test suite:
5805 Stop out of bounds reads in the ASN1 code when presented with
5806 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5808 If verify callback ignores invalid public key errors don't try to check
5809 certificate signature with the NULL public key.
5813 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5814 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5818 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5819 extra data after the compression methods not only for TLS 1.0
5820 but also for SSL 3.0 (as required by the specification).
5821 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5823 *) Change X509_certificate_type() to mark the key as exported/exportable
5824 when it's 512 *bits* long, not 512 bytes.
5827 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5829 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5830 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5831 a protocol version number mismatch like a decryption error
5832 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5835 *) Turn on RSA blinding by default in the default implementation
5836 to avoid a timing attack. Applications that don't want it can call
5837 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5838 They would be ill-advised to do so in most cases.
5839 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5841 *) Change RSA blinding code so that it works when the PRNG is not
5842 seeded (in this case, the secret RSA exponent is abused as
5843 an unpredictable seed -- if it is not unpredictable, there
5844 is no point in blinding anyway). Make RSA blinding thread-safe
5845 by remembering the creator's thread ID in rsa->blinding and
5846 having all other threads use local one-time blinding factors
5847 (this requires more computation than sharing rsa->blinding, but
5848 avoids excessive locking; and if an RSA object is not shared
5849 between threads, blinding will still be very fast).
5852 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5854 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5855 via timing by performing a MAC computation even if incorrrect
5856 block cipher padding has been found. This is a countermeasure
5857 against active attacks where the attacker has to distinguish
5858 between bad padding and a MAC verification error. (CVE-2003-0078)
5860 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5861 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5862 Martin Vuagnoux (EPFL, Ilion)]
5864 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5866 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5867 memory from it's contents. This is done with a counter that will
5868 place alternating values in each byte. This can be used to solve
5869 two issues: 1) the removal of calls to memset() by highly optimizing
5870 compilers, and 2) cleansing with other values than 0, since those can
5871 be read through on certain media, for example a swap space on disk.
5874 *) Bugfix: client side session caching did not work with external caching,
5875 because the session->cipher setting was not restored when reloading
5876 from the external cache. This problem was masked, when
5877 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
5878 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
5881 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
5882 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
5883 [Zeev Lieber <zeev-l@yahoo.com>]
5885 *) Undo an undocumented change introduced in 0.9.6e which caused
5886 repeated calls to OpenSSL_add_all_ciphers() and
5887 OpenSSL_add_all_digests() to be ignored, even after calling
5891 *) Change the default configuration reader to deal with last line not
5892 being properly terminated.
5895 *) Change X509_NAME_cmp() so it applies the special rules on handling
5896 DN values that are of type PrintableString, as well as RDNs of type
5897 emailAddress where the value has the type ia5String.
5898 [stefank@valicert.com via Richard Levitte]
5900 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
5901 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
5902 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
5903 the bitwise-OR of the two for use by the majority of applications
5904 wanting this behaviour, and update the docs. The documented
5905 behaviour and actual behaviour were inconsistent and had been
5906 changing anyway, so this is more a bug-fix than a behavioural
5908 [Geoff Thorpe, diagnosed by Nadav Har'El]
5910 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
5911 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
5914 *) Fix initialization code race conditions in
5915 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
5916 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
5917 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
5918 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
5919 ssl2_get_cipher_by_char(),
5920 ssl3_get_cipher_by_char().
5921 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
5923 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
5924 the cached sessions are flushed, as the remove_cb() might use ex_data
5925 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
5926 (see [openssl.org #212]).
5927 [Geoff Thorpe, Lutz Jaenicke]
5929 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
5930 length, instead of the encoding length to d2i_ASN1_OBJECT.
5933 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
5935 *) [In 0.9.6g-engine release:]
5936 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
5937 [Lynn Gazis <lgazis@rainbow.com>]
5939 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
5941 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
5942 and get fix the header length calculation.
5943 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
5944 Alon Kantor <alonk@checkpoint.com> (and others),
5947 *) Use proper error handling instead of 'assertions' in buffer
5948 overflow checks added in 0.9.6e. This prevents DoS (the
5949 assertions could call abort()).
5950 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
5952 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
5954 *) Add various sanity checks to asn1_get_length() to reject
5955 the ASN1 length bytes if they exceed sizeof(long), will appear
5956 negative or the content length exceeds the length of the
5958 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5960 *) Fix cipher selection routines: ciphers without encryption had no flags
5961 for the cipher strength set and where therefore not handled correctly
5962 by the selection routines (PR #130).
5965 *) Fix EVP_dsa_sha macro.
5969 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
5970 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
5971 that was added in OpenSSL 0.9.6d.
5973 As the countermeasure turned out to be incompatible with some
5974 broken SSL implementations, the new option is part of SSL_OP_ALL.
5975 SSL_OP_ALL is usually employed when compatibility with weird SSL
5976 implementations is desired (e.g. '-bugs' option to 's_client' and
5977 's_server'), so the new option is automatically set in many
5981 *) Changes in security patch:
5983 Changes marked "(CHATS)" were sponsored by the Defense Advanced
5984 Research Projects Agency (DARPA) and Air Force Research Laboratory,
5985 Air Force Materiel Command, USAF, under agreement number
5988 *) Add various sanity checks to asn1_get_length() to reject
5989 the ASN1 length bytes if they exceed sizeof(long), will appear
5990 negative or the content length exceeds the length of the
5991 supplied buffer. (CVE-2002-0659)
5992 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5994 *) Assertions for various potential buffer overflows, not known to
5996 [Ben Laurie (CHATS)]
5998 *) Various temporary buffers to hold ASCII versions of integers were
5999 too small for 64 bit platforms. (CVE-2002-0655)
6000 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6002 *) Remote buffer overflow in SSL3 protocol - an attacker could
6003 supply an oversized session ID to a client. (CVE-2002-0656)
6004 [Ben Laurie (CHATS)]
6006 *) Remote buffer overflow in SSL2 protocol - an attacker could
6007 supply an oversized client master key. (CVE-2002-0656)
6008 [Ben Laurie (CHATS)]
6010 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6012 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6013 encoded as NULL) with id-dsa-with-sha1.
6014 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6016 *) Check various X509_...() return values in apps/req.c.
6017 [Nils Larsch <nla@trustcenter.de>]
6019 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6020 an end-of-file condition would erronously be flagged, when the CRLF
6021 was just at the end of a processed block. The bug was discovered when
6022 processing data through a buffering memory BIO handing the data to a
6023 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6024 <ptsekov@syntrex.com> and Nedelcho Stanev.
6027 *) Implement a countermeasure against a vulnerability recently found
6028 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6029 before application data chunks to avoid the use of known IVs
6030 with data potentially chosen by the attacker.
6033 *) Fix length checks in ssl3_get_client_hello().
6036 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6037 to prevent ssl3_read_internal() from incorrectly assuming that
6038 ssl3_read_bytes() found application data while handshake
6039 processing was enabled when in fact s->s3->in_read_app_data was
6040 merely automatically cleared during the initial handshake.
6041 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6043 *) Fix object definitions for Private and Enterprise: they were not
6044 recognized in their shortname (=lowercase) representation. Extend
6045 obj_dat.pl to issue an error when using undefined keywords instead
6046 of silently ignoring the problem (Svenning Sorensen
6047 <sss@sss.dnsalias.net>).
6050 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6051 generators, i.e. generators other than 2 and 5. (Previously, the
6052 code did not properly initialise the 'add' and 'rem' values to
6053 BN_generate_prime().)
6055 In the new general case, we do not insist that 'generator' is
6056 actually a primitive root: This requirement is rather pointless;
6057 a generator of the order-q subgroup is just as good, if not
6061 *) Map new X509 verification errors to alerts. Discovered and submitted by
6062 Tom Wu <tom@arcot.com>.
6065 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6066 returning non-zero before the data has been completely received
6067 when using non-blocking I/O.
6068 [Bodo Moeller; problem pointed out by John Hughes]
6070 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6071 [Ben Laurie, Lutz Jaenicke]
6073 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6074 Yoram Zahavi <YoramZ@gilian.com>).
6077 *) Add information about CygWin 1.3 and on, and preserve proper
6078 configuration for the versions before that.
6079 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6081 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6082 check whether we deal with a copy of a session and do not delete from
6083 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6084 <izhar@checkpoint.com>.
6087 *) Do not store session data into the internal session cache, if it
6088 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6089 flag is set). Proposed by Aslam <aslam@funk.com>.
6092 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6096 *) [In 0.9.6d-engine release:]
6097 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6098 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6100 *) Add the configuration target linux-s390x.
6101 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6103 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6104 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6105 variable as an indication that a ClientHello message has been
6106 received. As the flag value will be lost between multiple
6107 invocations of ssl3_accept when using non-blocking I/O, the
6108 function may not be aware that a handshake has actually taken
6109 place, thus preventing a new session from being added to the
6112 To avoid this problem, we now set s->new_session to 2 instead of
6113 using a local variable.
6114 [Lutz Jaenicke, Bodo Moeller]
6116 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6117 if the SSL_R_LENGTH_MISMATCH error is detected.
6118 [Geoff Thorpe, Bodo Moeller]
6120 *) New 'shared_ldflag' column in Configure platform table.
6123 *) Fix EVP_CIPHER_mode macro.
6124 ["Dan S. Camper" <dan@bti.net>]
6126 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
6127 type, we must throw them away by setting rr->length to 0.
6128 [D P Chang <dpc@qualys.com>]
6130 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
6132 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
6133 <Dominikus.Scherkl@biodata.com>. (The previous implementation
6134 worked incorrectly for those cases where range = 10..._2 and
6135 3*range is two bits longer than range.)
6138 *) Only add signing time to PKCS7 structures if it is not already
6142 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
6143 OBJ_ld_ce should be OBJ_id_ce.
6144 Also some ip-pda OIDs in crypto/objects/objects.txt were
6145 incorrect (cf. RFC 3039).
6146 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
6148 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
6149 returns early because it has nothing to do.
6150 [Andy Schneider <andy.schneider@bjss.co.uk>]
6152 *) [In 0.9.6c-engine release:]
6153 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
6154 [Andy Schneider <andy.schneider@bjss.co.uk>]
6156 *) [In 0.9.6c-engine release:]
6157 Add support for Cryptographic Appliance's keyserver technology.
6158 (Use engine 'keyclient')
6159 [Cryptographic Appliances and Geoff Thorpe]
6161 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
6162 is called via tools/c89.sh because arguments have to be
6163 rearranged (all '-L' options must appear before the first object
6165 [Richard Shapiro <rshapiro@abinitio.com>]
6167 *) [In 0.9.6c-engine release:]
6168 Add support for Broadcom crypto accelerator cards, backported
6170 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
6172 *) [In 0.9.6c-engine release:]
6173 Add support for SureWare crypto accelerator cards from
6174 Baltimore Technologies. (Use engine 'sureware')
6175 [Baltimore Technologies and Mark Cox]
6177 *) [In 0.9.6c-engine release:]
6178 Add support for crypto accelerator cards from Accelerated
6179 Encryption Processing, www.aep.ie. (Use engine 'aep')
6180 [AEP Inc. and Mark Cox]
6182 *) Add a configuration entry for gcc on UnixWare.
6183 [Gary Benson <gbenson@redhat.com>]
6185 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
6186 messages are stored in a single piece (fixed-length part and
6187 variable-length part combined) and fix various bugs found on the way.
6190 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
6191 instead. BIO_gethostbyname() does not know what timeouts are
6192 appropriate, so entries would stay in cache even when they have
6194 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
6196 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
6197 faced with a pathologically small ClientHello fragment that does
6198 not contain client_version: Instead of aborting with an error,
6199 simply choose the highest available protocol version (i.e.,
6200 TLS 1.0 unless it is disabled). In practice, ClientHello
6201 messages are never sent like this, but this change gives us
6202 strictly correct behaviour at least for TLS.
6205 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
6206 never resets s->method to s->ctx->method when called from within
6207 one of the SSL handshake functions.
6208 [Bodo Moeller; problem pointed out by Niko Baric]
6210 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
6211 (sent using the client's version number) if client_version is
6212 smaller than the protocol version in use. Also change
6213 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
6214 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
6215 the client will at least see that alert.
6218 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
6222 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
6223 client receives HelloRequest while in a handshake.
6224 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
6226 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
6227 should end in 'break', not 'goto end' which circuments various
6228 cleanups done in state SSL_ST_OK. But session related stuff
6229 must be disabled for SSL_ST_OK in the case that we just sent a
6232 Also avoid some overhead by not calling ssl_init_wbio_buffer()
6233 before just sending a HelloRequest.
6234 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
6236 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
6237 reveal whether illegal block cipher padding was found or a MAC
6238 verification error occured. (Neither SSLerr() codes nor alerts
6239 are directly visible to potential attackers, but the information
6240 may leak via logfiles.)
6242 Similar changes are not required for the SSL 2.0 implementation
6243 because the number of padding bytes is sent in clear for SSL 2.0,
6244 and the extra bytes are just ignored. However ssl/s2_pkt.c
6245 failed to verify that the purported number of padding bytes is in
6249 *) Add OpenUNIX-8 support including shared libraries
6250 (Boyd Lynn Gerber <gerberb@zenez.com>).
6253 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
6254 'wristwatch attack' using huge encoding parameters (cf.
6255 James H. Manger's CRYPTO 2001 paper). Note that the
6256 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
6257 encoding parameters and hence was not vulnerable.
6260 *) BN_sqr() bug fix.
6261 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
6263 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
6264 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
6265 followed by modular reduction.
6266 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
6268 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
6269 equivalent based on BN_pseudo_rand() instead of BN_rand().
6272 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
6273 This function was broken, as the check for a new client hello message
6274 to handle SGC did not allow these large messages.
6275 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
6278 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
6281 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
6282 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
6285 *) Rework the configuration and shared library support for Tru64 Unix.
6286 The configuration part makes use of modern compiler features and
6287 still retains old compiler behavior for those that run older versions
6288 of the OS. The shared library support part includes a variant that
6289 uses the RPATH feature, and is available through the special
6290 configuration target "alpha-cc-rpath", which will never be selected
6292 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
6294 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
6295 with the same message size as in ssl3_get_certificate_request().
6296 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
6297 messages might inadvertently be reject as too long.
6298 [Petr Lampa <lampa@fee.vutbr.cz>]
6300 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6303 *) Modified SSL library such that the verify_callback that has been set
6304 specificly for an SSL object with SSL_set_verify() is actually being
6305 used. Before the change, a verify_callback set with this function was
6306 ignored and the verify_callback() set in the SSL_CTX at the time of
6307 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6308 to allow the necessary settings.
6311 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6312 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6313 done automatically (in contradiction to the requirements of the C
6314 standard). This made problems when used from OpenSSH.
6317 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6318 dh->length and always used
6320 BN_rand_range(priv_key, dh->p).
6322 BN_rand_range() is not necessary for Diffie-Hellman, and this
6323 specific range makes Diffie-Hellman unnecessarily inefficient if
6324 dh->length (recommended exponent length) is much smaller than the
6325 length of dh->p. We could use BN_rand_range() if the order of
6326 the subgroup was stored in the DH structure, but we only have
6331 BN_rand(priv_key, l, ...)
6333 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6339 RSA_eay_public_encrypt
6340 RSA_eay_private_decrypt
6341 RSA_eay_private_encrypt (signing)
6342 RSA_eay_public_decrypt (signature verification)
6344 (default implementations for RSA_public_encrypt,
6345 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6346 always reject numbers >= n.
6349 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6350 to synchronize access to 'locking_thread'. This is necessary on
6351 systems where access to 'locking_thread' (an 'unsigned long'
6352 variable) is not atomic.
6355 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6356 *before* setting the 'crypto_lock_rand' flag. The previous code had
6357 a race condition if 0 is a valid thread ID.
6358 [Travis Vitek <vitek@roguewave.com>]
6360 *) Add support for shared libraries under Irix.
6361 [Albert Chin-A-Young <china@thewrittenword.com>]
6363 *) Add configuration option to build on Linux on both big-endian and
6365 [Ralf Baechle <ralf@uni-koblenz.de>]
6367 *) Add the possibility to create shared libraries on HP-UX.
6370 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6372 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6373 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6374 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6375 PRNG state recovery was possible based on the output of
6376 one PRNG request appropriately sized to gain knowledge on
6377 'md' followed by enough consecutive 1-byte PRNG requests
6378 to traverse all of 'state'.
6380 1. When updating 'md_local' (the current thread's copy of 'md')
6381 during PRNG output generation, hash all of the previous
6382 'md_local' value, not just the half used for PRNG output.
6384 2. Make the number of bytes from 'state' included into the hash
6385 independent from the number of PRNG bytes requested.
6387 The first measure alone would be sufficient to avoid
6388 Markku-Juhani's attack. (Actually it had never occurred
6389 to me that the half of 'md_local' used for chaining was the
6390 half from which PRNG output bytes were taken -- I had always
6391 assumed that the secret half would be used.) The second
6392 measure makes sure that additional data from 'state' is never
6393 mixed into 'md_local' in small portions; this heuristically
6394 further strengthens the PRNG.
6397 *) Fix crypto/bn/asm/mips3.s.
6400 *) When only the key is given to "enc", the IV is undefined. Print out
6401 an error message in this case.
6404 *) Handle special case when X509_NAME is empty in X509 printing routines.
6407 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6408 positive and less than q.
6411 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6412 used: it isn't thread safe and the add_lock_callback should handle
6414 [Paul Rose <Paul.Rose@bridge.com>]
6416 *) Verify that incoming data obeys the block size in
6417 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6421 [Ulf Möller, Bodo Möller]
6423 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6424 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6425 when fixing the server behaviour for backwards-compatible 'client
6426 hello' messages. (Note that the attack is impractical against
6427 SSL 3.0 and TLS 1.0 anyway because length and version checking
6428 means that the probability of guessing a valid ciphertext is
6429 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6432 Before 0.9.5, the countermeasure (hide the error by generating a
6433 random 'decryption result') did not work properly because
6434 ERR_clear_error() was missing, meaning that SSL_get_error() would
6435 detect the supposedly ignored error.
6437 Both problems are now fixed.
6440 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6441 (previously it was 1024).
6444 *) Fix for compatibility mode trust settings: ignore trust settings
6445 unless some valid trust or reject settings are present.
6448 *) Fix for blowfish EVP: its a variable length cipher.
6451 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6452 parameters in DSA public key structures and return an error in the
6453 DSA routines if parameters are absent.
6456 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6457 in the current directory if neither $RANDFILE nor $HOME was set.
6458 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6459 caused some confusion to Windows users who haven't defined $HOME.
6460 Thus RAND_file_name() is changed again: e_os.h can define a
6461 DEFAULT_HOME, which will be used if $HOME is not set.
6462 For Windows, we use "C:"; on other platforms, we still require
6463 environment variables.
6465 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6466 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6467 having multiple threads call RAND_poll() concurrently.
6470 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6471 combination of a flag and a thread ID variable.
6472 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6473 flag), *other* threads can enter ssleay_add_bytes without obeying
6474 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6475 that they do not hold after the first thread unsets add_do_not_lock).
6478 *) Change bctest again: '-x' expressions are not available in all
6482 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6484 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6485 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6487 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6488 the default extension for executables, if any. Also, make the perl
6489 scripts that use symlink() to test if it really exists and use "cp"
6490 if it doesn't. All this made OpenSSL compilable and installable in
6494 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6495 If SEQUENCE is length is indefinite just set c->slen to the total
6496 amount of data available.
6497 [Steve Henson, reported by shige@FreeBSD.org]
6498 [This change does not apply to 0.9.7.]
6500 *) Change bctest to avoid here-documents inside command substitution
6501 (workaround for FreeBSD /bin/sh bug).
6502 For compatibility with Ultrix, avoid shell functions (introduced
6503 in the bctest version that searches along $PATH).
6506 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6507 with des_encrypt() defined on some operating systems, like Solaris
6511 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6512 On the Importance of Eliminating Errors in Cryptographic
6513 Computations, J. Cryptology 14 (2001) 2, 101-119,
6514 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6517 *) MIPS assembler BIGNUM division bug fix.
6520 *) Disabled incorrect Alpha assembler code.
6523 *) Fix PKCS#7 decode routines so they correctly update the length
6524 after reading an EOC for the EXPLICIT tag.
6526 [This change does not apply to 0.9.7.]
6528 *) Fix bug in PKCS#12 key generation routines. This was triggered
6529 if a 3DES key was generated with a 0 initial byte. Include
6530 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6531 (but broken) behaviour.
6534 *) Enhance bctest to search for a working bc along $PATH and print
6536 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6538 *) Fix memory leaks in err.c: free err_data string if necessary;
6539 don't write to the wrong index in ERR_set_error_data.
6542 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6546 *) Replace rdtsc with _emit statements for VC++ version 5.
6547 [Jeremy Cooper <jeremy@baymoo.org>]
6549 *) Make it possible to reuse SSLv2 sessions.
6552 *) In copy_email() check for >= 0 as a return value for
6553 X509_NAME_get_index_by_NID() since 0 is a valid index.
6554 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6556 *) Avoid coredump with unsupported or invalid public keys by checking if
6557 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6558 PKCS7_verify() fails with non detached data.
6561 *) Don't use getenv in library functions when run as setuid/setgid.
6562 New function OPENSSL_issetugid().
6565 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6566 due to incorrect handling of multi-threading:
6568 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6570 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6572 3. Count how many times MemCheck_off() has been called so that
6573 nested use can be treated correctly. This also avoids
6574 inband-signalling in the previous code (which relied on the
6575 assumption that thread ID 0 is impossible).
6578 *) Add "-rand" option also to s_client and s_server.
6581 *) Fix CPU detection on Irix 6.x.
6582 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6583 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6585 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6588 [This change does not apply to 0.9.7.]
6590 *) Use the cached encoding of an X509_NAME structure rather than
6591 copying it. This is apparently the reason for the libsafe "errors"
6592 but the code is actually correct.
6595 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6596 Bleichenbacher's DSA attack.
6597 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6598 to be set and top=0 forces the highest bit to be set; top=-1 is new
6599 and leaves the highest bit random.
6600 [Ulf Moeller, Bodo Moeller]
6602 *) In the NCONF_...-based implementations for CONF_... queries
6603 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6604 a temporary CONF structure with the data component set to NULL
6605 (which gives segmentation faults in lh_retrieve).
6606 Instead, use NULL for the CONF pointer in CONF_get_string and
6607 CONF_get_number (which may use environment variables) and directly
6608 return NULL from CONF_get_section.
6611 *) Fix potential buffer overrun for EBCDIC.
6614 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6615 keyUsage if basicConstraints absent for a CA.
6618 *) Make SMIME_write_PKCS7() write mail header values with a format that
6619 is more generally accepted (no spaces before the semicolon), since
6620 some programs can't parse those values properly otherwise. Also make
6621 sure BIO's that break lines after each write do not create invalid
6625 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6626 macros previously used would not encode an empty SEQUENCE OF
6627 and break the signature.
6629 [This change does not apply to 0.9.7.]
6631 *) Zero the premaster secret after deriving the master secret in
6635 *) Add some EVP_add_digest_alias registrations (as found in
6636 OpenSSL_add_all_digests()) to SSL_library_init()
6637 aka OpenSSL_add_ssl_algorithms(). This provides improved
6638 compatibility with peers using X.509 certificates
6639 with unconventional AlgorithmIdentifier OIDs.
6642 *) Fix for Irix with NO_ASM.
6643 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6645 *) ./config script fixes.
6646 [Ulf Moeller, Richard Levitte]
6648 *) Fix 'openssl passwd -1'.
6651 *) Change PKCS12_key_gen_asc() so it can cope with non null
6652 terminated strings whose length is passed in the passlen
6653 parameter, for example from PEM callbacks. This was done
6654 by adding an extra length parameter to asc2uni().
6655 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6657 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6658 call failed, free the DSA structure.
6661 *) Fix to uni2asc() to cope with zero length Unicode strings.
6662 These are present in some PKCS#12 files.
6665 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6666 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6667 when writing a 32767 byte record.
6668 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6670 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6671 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6673 (RSA objects have a reference count access to which is protected
6674 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6675 so they are meant to be shared between threads.)
6676 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6677 "Reddie, Steven" <Steven.Reddie@ca.com>]
6679 *) Fix a deadlock in CRYPTO_mem_leaks().
6682 *) Use better test patterns in bntest.
6685 *) rand_win.c fix for Borland C.
6688 *) BN_rshift bugfix for n == 0.
6691 *) Add a 'bctest' script that checks for some known 'bc' bugs
6692 so that 'make test' does not abort just because 'bc' is broken.
6695 *) Store verify_result within SSL_SESSION also for client side to
6696 avoid potential security hole. (Re-used sessions on the client side
6697 always resulted in verify_result==X509_V_OK, not using the original
6698 result of the server certificate verification.)
6701 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6702 SSL3_RT_APPLICATION_DATA, return 0.
6703 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6707 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6708 releases, have been re-implemented by renaming the previous
6709 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6710 and ssl3_read_internal, respectively, and adding 'peek' parameters
6711 to them. The new ssl[23]_{read,peek} functions are calls to
6712 ssl[23]_read_internal with the 'peek' flag set appropriately.
6713 A 'peek' parameter has also been added to ssl3_read_bytes, which
6714 does the actual work for ssl3_read_internal.
6717 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6718 the method-specific "init()" handler. Also clean up ex_data after
6719 calling the method-specific "finish()" handler. Previously, this was
6720 happening the other way round.
6723 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6724 The previous value, 12, was not always sufficient for BN_mod_exp().
6727 *) Make sure that shared libraries get the internal name engine with
6728 the full version number and not just 0. This should mark the
6729 shared libraries as not backward compatible. Of course, this should
6730 be changed again when we can guarantee backward binary compatibility.
6733 *) Fix typo in get_cert_by_subject() in by_dir.c
6734 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6736 *) Rework the system to generate shared libraries:
6738 - Make note of the expected extension for the shared libraries and
6739 if there is a need for symbolic links from for example libcrypto.so.0
6740 to libcrypto.so.0.9.7. There is extended info in Configure for
6743 - Make as few rebuilds of the shared libraries as possible.
6745 - Still avoid linking the OpenSSL programs with the shared libraries.
6747 - When installing, install the shared libraries separately from the
6751 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6753 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6754 and not in SSL_clear because the latter is also used by the
6755 accept/connect functions; previously, the settings made by
6756 SSL_set_read_ahead would be lost during the handshake.
6757 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6759 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6760 Previously, it would create entries for disableed algorithms no
6764 *) Added several new manual pages for SSL_* function.
6767 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6769 *) In ssl23_get_client_hello, generate an error message when faced
6770 with an initial SSL 3.0/TLS record that is too small to contain the
6771 first two bytes of the ClientHello message, i.e. client_version.
6772 (Note that this is a pathologic case that probably has never happened
6773 in real life.) The previous approach was to use the version number
6774 from the record header as a substitute; but our protocol choice
6775 should not depend on that one because it is not authenticated
6776 by the Finished messages.
6779 *) More robust randomness gathering functions for Windows.
6780 [Jeffrey Altman <jaltman@columbia.edu>]
6782 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6783 not set then we don't setup the error code for issuer check errors
6784 to avoid possibly overwriting other errors which the callback does
6785 handle. If an application does set the flag then we assume it knows
6786 what it is doing and can handle the new informational codes
6790 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6791 a general "ANY" type, as such it should be able to decode anything
6792 including tagged types. However it didn't check the class so it would
6793 wrongly interpret tagged types in the same way as their universal
6794 counterpart and unknown types were just rejected. Changed so that the
6795 tagged and unknown types are handled in the same way as a SEQUENCE:
6796 that is the encoding is stored intact. There is also a new type
6797 "V_ASN1_OTHER" which is used when the class is not universal, in this
6798 case we have no idea what the actual type is so we just lump them all
6802 *) On VMS, stdout may very well lead to a file that is written to
6803 in a record-oriented fashion. That means that every write() will
6804 write a separate record, which will be read separately by the
6805 programs trying to read from it. This can be very confusing.
6807 The solution is to put a BIO filter in the way that will buffer
6808 text until a linefeed is reached, and then write everything a
6809 line at a time, so every record written will be an actual line,
6810 not chunks of lines and not (usually doesn't happen, but I've
6811 seen it once) several lines in one record. BIO_f_linebuffer() is
6814 Currently, it's a VMS-only method, because that's where it has
6815 been tested well enough.
6818 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6819 it can return incorrect results.
6820 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6821 but it was in 0.9.6-beta[12].)
6824 *) Disable the check for content being present when verifying detached
6825 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6826 include zero length content when signing messages.
6829 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6830 BIO_ctrl (for BIO pairs).
6833 *) Add DSO method for VMS.
6836 *) Bug fix: Montgomery multiplication could produce results with the
6840 *) Add RPM specification openssl.spec and modify it to build three
6841 packages. The default package contains applications, application
6842 documentation and run-time libraries. The devel package contains
6843 include files, static libraries and function documentation. The
6844 doc package contains the contents of the doc directory. The original
6845 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6848 *) Add a large number of documentation files for many SSL routines.
6849 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6851 *) Add a configuration entry for Sony News 4.
6852 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6854 *) Don't set the two most significant bits to one when generating a
6855 random number < q in the DSA library.
6858 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6859 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6860 the underlying transport is blocking) if a handshake took place.
6861 (The default behaviour is needed by applications such as s_client
6862 and s_server that use select() to determine when to use SSL_read;
6863 but for applications that know in advance when to expect data, it
6864 just makes things more complicated.)
6867 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
6871 *) Add a few more EBCDIC conditionals that make `req' and `x509'
6872 work better on such systems.
6873 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6875 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
6876 Update PKCS12_parse() so it copies the friendlyName and the
6877 keyid to the certificates aux info.
6880 *) Fix bug in PKCS7_verify() which caused an infinite loop
6881 if there was more than one signature.
6882 [Sven Uszpelkat <su@celocom.de>]
6884 *) Major change in util/mkdef.pl to include extra information
6885 about each symbol, as well as presentig variables as well
6886 as functions. This change means that there's n more need
6887 to rebuild the .num files when some algorithms are excluded.
6890 *) Allow the verify time to be set by an application,
6891 rather than always using the current time.
6894 *) Phase 2 verify code reorganisation. The certificate
6895 verify code now looks up an issuer certificate by a
6896 number of criteria: subject name, authority key id
6897 and key usage. It also verifies self signed certificates
6898 by the same criteria. The main comparison function is
6899 X509_check_issued() which performs these checks.
6901 Lot of changes were necessary in order to support this
6902 without completely rewriting the lookup code.
6904 Authority and subject key identifier are now cached.
6906 The LHASH 'certs' is X509_STORE has now been replaced
6907 by a STACK_OF(X509_OBJECT). This is mainly because an
6908 LHASH can't store or retrieve multiple objects with
6909 the same hash value.
6911 As a result various functions (which were all internal
6912 use only) have changed to handle the new X509_STORE
6913 structure. This will break anything that messed round
6914 with X509_STORE internally.
6916 The functions X509_STORE_add_cert() now checks for an
6917 exact match, rather than just subject name.
6919 The X509_STORE API doesn't directly support the retrieval
6920 of multiple certificates matching a given criteria, however
6921 this can be worked round by performing a lookup first
6922 (which will fill the cache with candidate certificates)
6923 and then examining the cache for matches. This is probably
6924 the best we can do without throwing out X509_LOOKUP
6925 entirely (maybe later...).
6927 The X509_VERIFY_CTX structure has been enhanced considerably.
6929 All certificate lookup operations now go via a get_issuer()
6930 callback. Although this currently uses an X509_STORE it
6931 can be replaced by custom lookups. This is a simple way
6932 to bypass the X509_STORE hackery necessary to make this
6933 work and makes it possible to use more efficient techniques
6934 in future. A very simple version which uses a simple
6935 STACK for its trusted certificate store is also provided
6936 using X509_STORE_CTX_trusted_stack().
6938 The verify_cb() and verify() callbacks now have equivalents
6939 in the X509_STORE_CTX structure.
6941 X509_STORE_CTX also has a 'flags' field which can be used
6942 to customise the verify behaviour.
6945 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
6946 excludes S/MIME capabilities.
6949 *) When a certificate request is read in keep a copy of the
6950 original encoding of the signed data and use it when outputing
6951 again. Signatures then use the original encoding rather than
6952 a decoded, encoded version which may cause problems if the
6953 request is improperly encoded.
6956 *) For consistency with other BIO_puts implementations, call
6957 buffer_write(b, ...) directly in buffer_puts instead of calling
6960 In BIO_puts, increment b->num_write as in BIO_write.
6961 [Peter.Sylvester@EdelWeb.fr]
6963 *) Fix BN_mul_word for the case where the word is 0. (We have to use
6964 BN_zero, we may not return a BIGNUM with an array consisting of
6968 *) Avoid calling abort() from within the library when problems are
6969 detected, except if preprocessor symbols have been defined
6970 (such as REF_CHECK, BN_DEBUG etc.).
6973 *) New openssl application 'rsautl'. This utility can be
6974 used for low level RSA operations. DER public key
6975 BIO/fp routines also added.
6978 *) New Configure entry and patches for compiling on QNX 4.
6979 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
6981 *) A demo state-machine implementation was sponsored by
6982 Nuron (http://www.nuron.com/) and is now available in
6983 demos/state_machine.
6986 *) New options added to the 'dgst' utility for signature
6987 generation and verification.
6990 *) Unrecognized PKCS#7 content types are now handled via a
6991 catch all ASN1_TYPE structure. This allows unsupported
6992 types to be stored as a "blob" and an application can
6993 encode and decode it manually.
6996 *) Fix various signed/unsigned issues to make a_strex.c
6998 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7000 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7001 length if passed a buffer. ASN1_INTEGER_to_BN failed
7002 if passed a NULL BN and its argument was negative.
7003 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7005 *) Modification to PKCS#7 encoding routines to output definite
7006 length encoding. Since currently the whole structures are in
7007 memory there's not real point in using indefinite length
7008 constructed encoding. However if OpenSSL is compiled with
7009 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7012 *) Added BIO_vprintf() and BIO_vsnprintf().
7015 *) Added more prefixes to parse for in the the strings written
7016 through a logging bio, to cover all the levels that are available
7017 through syslog. The prefixes are now:
7019 PANIC, EMERG, EMR => LOG_EMERG
7020 ALERT, ALR => LOG_ALERT
7021 CRIT, CRI => LOG_CRIT
7022 ERROR, ERR => LOG_ERR
7023 WARNING, WARN, WAR => LOG_WARNING
7024 NOTICE, NOTE, NOT => LOG_NOTICE
7025 INFO, INF => LOG_INFO
7026 DEBUG, DBG => LOG_DEBUG
7028 and as before, if none of those prefixes are present at the
7029 beginning of the string, LOG_ERR is chosen.
7031 On Win32, the LOG_* levels are mapped according to this:
7033 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7034 LOG_WARNING => EVENTLOG_WARNING_TYPE
7035 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7039 *) Made it possible to reconfigure with just the configuration
7040 argument "reconf" or "reconfigure". The command line arguments
7041 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7042 and are retrieved from there when reconfiguring.
7046 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7048 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7051 *) The obj_dat.pl script was messing up the sorting of object
7052 names. The reason was that it compared the quoted version
7053 of strings as a result "OCSP" > "OCSP Signing" because
7054 " > SPACE. Changed script to store unquoted versions of
7055 names and add quotes on output. It was also omitting some
7056 names from the lookup table if they were given a default
7057 value (that is if SN is missing it is given the same
7058 value as LN and vice versa), these are now added on the
7059 grounds that if an object has a name we should be able to
7060 look it up. Finally added warning output when duplicate
7061 short or long names are found.
7064 *) Changes needed for Tandem NSK.
7065 [Scott Uroff <scott@xypro.com>]
7067 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7068 RSA_padding_check_SSLv23(), special padding was never detected
7069 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7070 version rollback attacks was not effective.
7072 In s23_clnt.c, don't use special rollback-attack detection padding
7073 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7074 client; similarly, in s23_srvr.c, don't do the rollback check if
7075 SSL 2.0 is the only protocol enabled in the server.
7078 *) Make it possible to get hexdumps of unprintable data with 'openssl
7079 asn1parse'. By implication, the functions ASN1_parse_dump() and
7080 BIO_dump_indent() are added.
7083 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7084 these print out strings and name structures based on various
7085 flags including RFC2253 support and proper handling of
7086 multibyte characters. Added options to the 'x509' utility
7087 to allow the various flags to be set.
7090 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7091 Also change the functions X509_cmp_current_time() and
7092 X509_gmtime_adj() work with an ASN1_TIME structure,
7093 this will enable certificates using GeneralizedTime in validity
7094 dates to be checked.
7097 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7098 negative public key encodings) on by default,
7099 NO_NEG_PUBKEY_BUG can be set to disable it.
7102 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7103 content octets. An i2c_ASN1_OBJECT is unnecessary because
7104 the encoding can be trivially obtained from the structure.
7107 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7108 not read locks (CRYPTO_r_[un]lock).
7111 *) A first attempt at creating official support for shared
7112 libraries through configuration. I've kept it so the
7113 default is static libraries only, and the OpenSSL programs
7114 are always statically linked for now, but there are
7115 preparations for dynamic linking in place.
7116 This has been tested on Linux and Tru64.
7119 *) Randomness polling function for Win9x, as described in:
7120 Peter Gutmann, Software Generation of Practically Strong
7124 *) Fix so PRNG is seeded in req if using an already existing
7128 *) New options to smime application. -inform and -outform
7129 allow alternative formats for the S/MIME message including
7130 PEM and DER. The -content option allows the content to be
7131 specified separately. This should allow things like Netscape
7132 form signing output easier to verify.
7135 *) Fix the ASN1 encoding of tags using the 'long form'.
7138 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
7139 STRING types. These convert content octets to and from the
7140 underlying type. The actual tag and length octets are
7141 already assumed to have been read in and checked. These
7142 are needed because all other string types have virtually
7143 identical handling apart from the tag. By having versions
7144 of the ASN1 functions that just operate on content octets
7145 IMPLICIT tagging can be handled properly. It also allows
7146 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
7147 and ASN1_INTEGER are identical apart from the tag.
7150 *) Change the handling of OID objects as follows:
7152 - New object identifiers are inserted in objects.txt, following
7153 the syntax given in objects.README.
7154 - objects.pl is used to process obj_mac.num and create a new
7156 - obj_dat.pl is used to create a new obj_dat.h, using the data in
7159 This is currently kind of a hack, and the perl code in objects.pl
7160 isn't very elegant, but it works as I intended. The simplest way
7161 to check that it worked correctly is to look in obj_dat.h and
7162 check the array nid_objs and make sure the objects haven't moved
7163 around (this is important!). Additions are OK, as well as
7164 consistent name changes.
7167 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
7170 *) Addition of the command line parameter '-rand file' to 'openssl req'.
7171 The given file adds to whatever has already been seeded into the
7172 random pool through the RANDFILE configuration file option or
7173 environment variable, or the default random state file.
7176 *) mkstack.pl now sorts each macro group into lexical order.
7177 Previously the output order depended on the order the files
7178 appeared in the directory, resulting in needless rewriting
7182 *) Patches to make OpenSSL compile under Win32 again. Mostly
7183 work arounds for the VC++ problem that it treats func() as
7184 func(void). Also stripped out the parts of mkdef.pl that
7185 added extra typesafe functions: these no longer exist.
7188 *) Reorganisation of the stack code. The macros are now all
7189 collected in safestack.h . Each macro is defined in terms of
7190 a "stack macro" of the form SKM_<name>(type, a, b). The
7191 DEBUG_SAFESTACK is now handled in terms of function casts,
7192 this has the advantage of retaining type safety without the
7193 use of additional functions. If DEBUG_SAFESTACK is not defined
7194 then the non typesafe macros are used instead. Also modified the
7195 mkstack.pl script to handle the new form. Needs testing to see
7196 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
7197 the default if no major problems. Similar behaviour for ASN1_SET_OF
7198 and PKCS12_STACK_OF.
7201 *) When some versions of IIS use the 'NET' form of private key the
7202 key derivation algorithm is different. Normally MD5(password) is
7203 used as a 128 bit RC4 key. In the modified case
7204 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
7205 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
7206 as the old Netscape_RSA functions except they have an additional
7207 'sgckey' parameter which uses the modified algorithm. Also added
7208 an -sgckey command line option to the rsa utility. Thanks to
7209 Adrian Peck <bertie@ncipher.com> for posting details of the modified
7210 algorithm to openssl-dev.
7213 *) The evp_local.h macros were using 'c.##kname' which resulted in
7214 invalid expansion on some systems (SCO 5.0.5 for example).
7215 Corrected to 'c.kname'.
7216 [Phillip Porch <root@theporch.com>]
7218 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
7219 a STACK of email addresses from a certificate or request, these look
7220 in the subject name and the subject alternative name extensions and
7221 omit any duplicate addresses.
7224 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
7225 This makes DSA verification about 2 % faster.
7228 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
7229 (meaning that now 2^5 values will be precomputed, which is only 4 KB
7230 plus overhead for 1024 bit moduli).
7231 This makes exponentiations about 0.5 % faster for 1024 bit
7232 exponents (as measured by "openssl speed rsa2048").
7235 *) Rename memory handling macros to avoid conflicts with other
7237 Malloc => OPENSSL_malloc
7238 Malloc_locked => OPENSSL_malloc_locked
7239 Realloc => OPENSSL_realloc
7240 Free => OPENSSL_free
7243 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
7244 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
7247 *) CygWin32 support.
7248 [John Jarvie <jjarvie@newsguy.com>]
7250 *) The type-safe stack code has been rejigged. It is now only compiled
7251 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
7252 by default all type-specific stack functions are "#define"d back to
7253 standard stack functions. This results in more streamlined output
7254 but retains the type-safety checking possibilities of the original
7258 *) The STACK code has been cleaned up, and certain type declarations
7259 that didn't make a lot of sense have been brought in line. This has
7260 also involved a cleanup of sorts in safestack.h to more correctly
7261 map type-safe stack functions onto their plain stack counterparts.
7262 This work has also resulted in a variety of "const"ifications of
7263 lots of the code, especially "_cmp" operations which should normally
7264 be prototyped with "const" parameters anyway.
7267 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
7268 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
7269 (The PRNG state consists of two parts, the large pool 'state' and 'md',
7270 where all of 'md' is used each time the PRNG is used, but 'state'
7271 is used only indexed by a cyclic counter. As entropy may not be
7272 well distributed from the beginning, 'md' is important as a
7273 chaining variable. However, the output function chains only half
7274 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
7275 all of 'md', and seeding with STATE_SIZE dummy bytes will result
7276 in all of 'state' being rewritten, with the new values depending
7277 on virtually all of 'md'. This overcomes the 80 bit limitation.)
7280 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
7281 the handshake is continued after ssl_verify_cert_chain();
7282 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
7283 can lead to 'unexplainable' connection aborts later.
7284 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
7286 *) Major EVP API cipher revision.
7287 Add hooks for extra EVP features. This allows various cipher
7288 parameters to be set in the EVP interface. Support added for variable
7289 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
7290 setting of RC2 and RC5 parameters.
7292 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
7295 Remove lots of duplicated code from the EVP library. For example *every*
7296 cipher init() function handles the 'iv' in the same way according to the
7297 cipher mode. They also all do nothing if the 'key' parameter is NULL and
7298 for CFB and OFB modes they zero ctx->num.
7300 New functionality allows removal of S/MIME code RC2 hack.
7302 Most of the routines have the same form and so can be declared in terms
7305 By shifting this to the top level EVP_CipherInit() it can be removed from
7306 all individual ciphers. If the cipher wants to handle IVs or keys
7307 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7310 Change lots of functions like EVP_EncryptUpdate() to now return a
7311 value: although software versions of the algorithms cannot fail
7312 any installed hardware versions can.
7315 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7316 this option is set, tolerate broken clients that send the negotiated
7317 protocol version number instead of the requested protocol version
7321 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7322 i.e. non-zero for export ciphersuites, zero otherwise.
7323 Previous versions had this flag inverted, inconsistent with
7324 rsa_tmp_cb (..._TMP_RSA_CB).
7325 [Bodo Moeller; problem reported by Amit Chopra]
7327 *) Add missing DSA library text string. Work around for some IIS
7328 key files with invalid SEQUENCE encoding.
7331 *) Add a document (doc/standards.txt) that list all kinds of standards
7332 and so on that are implemented in OpenSSL.
7335 *) Enhance c_rehash script. Old version would mishandle certificates
7336 with the same subject name hash and wouldn't handle CRLs at all.
7337 Added -fingerprint option to crl utility, to support new c_rehash
7341 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7344 *) Fix for SSL server purpose checking. Server checking was
7345 rejecting certificates which had extended key usage present
7346 but no ssl client purpose.
7347 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7349 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7350 is a little unclear about how a blank password is handled.
7351 Since the password in encoded as a BMPString with terminating
7352 double NULL a zero length password would end up as just the
7353 double NULL. However no password at all is different and is
7354 handled differently in the PKCS#12 key generation code. NS
7355 treats a blank password as zero length. MSIE treats it as no
7356 password on export: but it will try both on import. We now do
7357 the same: PKCS12_parse() tries zero length and no password if
7358 the password is set to "" or NULL (NULL is now a valid password:
7359 it wasn't before) as does the pkcs12 application.
7362 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7363 perror when PEM_read_bio_X509_REQ fails, the error message must
7364 be obtained from the error queue.
7367 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7368 it in ERR_remove_state if appropriate, and change ERR_get_state
7369 accordingly to avoid race conditions (this is necessary because
7370 thread_hash is no longer constant once set).
7373 *) Bugfix for linux-elf makefile.one.
7376 *) RSA_get_default_method() will now cause a default
7377 RSA_METHOD to be chosen if one doesn't exist already.
7378 Previously this was only set during a call to RSA_new()
7379 or RSA_new_method(NULL) meaning it was possible for
7380 RSA_get_default_method() to return NULL.
7383 *) Added native name translation to the existing DSO code
7384 that will convert (if the flag to do so is set) filenames
7385 that are sufficiently small and have no path information
7386 into a canonical native form. Eg. "blah" converted to
7387 "libblah.so" or "blah.dll" etc.
7390 *) New function ERR_error_string_n(e, buf, len) which is like
7391 ERR_error_string(e, buf), but writes at most 'len' bytes
7392 including the 0 terminator. For ERR_error_string_n, 'buf'
7394 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7396 *) CONF library reworked to become more general. A new CONF
7397 configuration file reader "class" is implemented as well as a
7398 new functions (NCONF_*, for "New CONF") to handle it. The now
7399 old CONF_* functions are still there, but are reimplemented to
7400 work in terms of the new functions. Also, a set of functions
7401 to handle the internal storage of the configuration data is
7402 provided to make it easier to write new configuration file
7403 reader "classes" (I can definitely see something reading a
7404 configuration file in XML format, for example), called _CONF_*,
7405 or "the configuration storage API"...
7407 The new configuration file reading functions are:
7409 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7410 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7412 NCONF_default, NCONF_WIN32
7414 NCONF_dump_fp, NCONF_dump_bio
7416 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7417 NCONF_new creates a new CONF object. This works in the same way
7418 as other interfaces in OpenSSL, like the BIO interface.
7419 NCONF_dump_* dump the internal storage of the configuration file,
7420 which is useful for debugging. All other functions take the same
7421 arguments as the old CONF_* functions wth the exception of the
7422 first that must be a `CONF *' instead of a `LHASH *'.
7424 To make it easer to use the new classes with the old CONF_* functions,
7425 the function CONF_set_default_method is provided.
7428 *) Add '-tls1' option to 'openssl ciphers', which was already
7429 mentioned in the documentation but had not been implemented.
7430 (This option is not yet really useful because even the additional
7431 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7434 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7435 OpenSSL-based applications) load shared libraries and bind to
7436 them in a portable way.
7437 [Geoff Thorpe, with contributions from Richard Levitte]
7439 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7441 *) Make sure _lrotl and _lrotr are only used with MSVC.
7443 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7444 (the default implementation of RAND_status).
7446 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7447 to '-clrext' (= clear extensions), as intended and documented.
7448 [Bodo Moeller; inconsistency pointed out by Michael Attili
7451 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7452 was larger than the MD block size.
7453 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7455 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7456 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7457 using the passed key: if the passed key was a private key the result
7458 of X509_print(), for example, would be to print out all the private key
7462 *) des_quad_cksum() byte order bug fix.
7463 [Ulf Möller, using the problem description in krb4-0.9.7, where
7464 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7466 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7468 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7470 *) For easily testing in shell scripts whether some command
7471 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7472 returns with exit code 0 iff no command of the given name is available.
7473 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7474 the output goes to stdout and nothing is printed to stderr.
7475 Additional arguments are always ignored.
7477 Since for each cipher there is a command of the same name,
7478 the 'no-cipher' compilation switches can be tested this way.
7480 ('openssl no-XXX' is not able to detect pseudo-commands such
7481 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7484 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7487 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7488 is set; it will be thrown away anyway because each handshake creates
7490 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7491 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7492 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7493 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7496 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7497 'Q' and 'R' lose their special meanings (quit/renegotiate).
7498 This is part of what -quiet does; unlike -quiet, -ign_eof
7499 does not suppress any output.
7502 *) Add compatibility options to the purpose and trust code. The
7503 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7504 accepts a certificate or CA, this was the previous behaviour,
7505 with all the associated security issues.
7507 X509_TRUST_COMPAT is the old trust behaviour: only and
7508 automatically trust self signed roots in certificate store. A
7509 new trust setting X509_TRUST_DEFAULT is used to specify that
7510 a purpose has no associated trust setting and it should instead
7511 use the value in the default purpose.
7514 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7515 and fix a memory leak.
7518 *) In util/mkerr.pl (which implements 'make errors'), preserve
7519 reason strings from the previous version of the .c file, as
7520 the default to have only downcase letters (and digits) in
7521 automatically generated reasons codes is not always appropriate.
7524 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7525 using strerror. Previously, ERR_reason_error_string() returned
7526 library names as reason strings for SYSerr; but SYSerr is a special
7527 case where small numbers are errno values, not library numbers.
7530 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7531 converts DSA parameters into DH parameters. (When creating parameters,
7532 DSA_generate_parameters is used.)
7535 *) Include 'length' (recommended exponent length) in C code generated
7536 by 'openssl dhparam -C'.
7539 *) The second argument to set_label in perlasm was already being used
7540 so couldn't be used as a "file scope" flag. Moved to third argument
7544 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7545 instead of RAND_bytes for encryption IVs and salts.
7548 *) Include RAND_status() into RAND_METHOD instead of implementing
7549 it only for md_rand.c Otherwise replacing the PRNG by calling
7550 RAND_set_rand_method would be impossible.
7553 *) Don't let DSA_generate_key() enter an infinite loop if the random
7554 number generation fails.
7557 *) New 'rand' application for creating pseudo-random output.
7560 *) Added configuration support for Linux/IA64
7561 [Rolf Haberrecker <rolf@suse.de>]
7563 *) Assembler module support for Mingw32.
7566 *) Shared library support for HPUX (in shlib/).
7567 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7569 *) Shared library support for Solaris gcc.
7570 [Lutz Behnke <behnke@trustcenter.de>]
7572 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7574 *) PKCS7_encrypt() was adding text MIME headers twice because they
7575 were added manually and by SMIME_crlf_copy().
7578 *) In bntest.c don't call BN_rand with zero bits argument.
7579 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7581 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7582 case was implemented. This caused BN_div_recp() to fail occasionally.
7585 *) Add an optional second argument to the set_label() in the perl
7586 assembly language builder. If this argument exists and is set
7587 to 1 it signals that the assembler should use a symbol whose
7588 scope is the entire file, not just the current function. This
7589 is needed with MASM which uses the format label:: for this scope.
7590 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7592 *) Change the ASN1 types so they are typedefs by default. Before
7593 almost all types were #define'd to ASN1_STRING which was causing
7594 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7598 *) Change names of new functions to the new get1/get0 naming
7599 convention: After 'get1', the caller owns a reference count
7600 and has to call ..._free; 'get0' returns a pointer to some
7601 data structure without incrementing reference counters.
7602 (Some of the existing 'get' functions increment a reference
7603 counter, some don't.)
7604 Similarly, 'set1' and 'add1' functions increase reference
7605 counters or duplicate objects.
7608 *) Allow for the possibility of temp RSA key generation failure:
7609 the code used to assume it always worked and crashed on failure.
7612 *) Fix potential buffer overrun problem in BIO_printf().
7613 [Ulf Möller, using public domain code by Patrick Powell; problem
7614 pointed out by David Sacerdote <das33@cornell.edu>]
7616 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7617 RAND_egd() and RAND_status(). In the command line application,
7618 the EGD socket can be specified like a seed file using RANDFILE
7622 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7623 Some CAs (e.g. Verisign) distribute certificates in this form.
7626 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7627 list to exclude them. This means that no special compilation option
7628 is needed to use anonymous DH: it just needs to be included in the
7632 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7633 EVP_MD_type. The old functionality is available in a new macro called
7634 EVP_MD_md(). Change code that uses it and update docs.
7637 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7638 where the 'void *' argument is replaced by a function pointer argument.
7639 Previously 'void *' was abused to point to functions, which works on
7640 many platforms, but is not correct. As these functions are usually
7641 called by macros defined in OpenSSL header files, most source code
7642 should work without changes.
7645 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7646 sections with information on -D... compiler switches used for
7647 compiling the library so that applications can see them. To enable
7648 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7649 must be defined. E.g.,
7650 #define OPENSSL_ALGORITHM_DEFINES
7651 #include <openssl/opensslconf.h>
7652 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7653 [Richard Levitte, Ulf and Bodo Möller]
7655 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7659 *) Change the 'other' type in certificate aux info to a STACK_OF
7660 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7661 the required ASN1 format: arbitrary types determined by an OID.
7664 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7665 argument to 'req'. This is not because the function is newer or
7666 better than others it just uses the work 'NEW' in the certificate
7667 request header lines. Some software needs this.
7670 *) Reorganise password command line arguments: now passwords can be
7671 obtained from various sources. Delete the PEM_cb function and make
7672 it the default behaviour: i.e. if the callback is NULL and the
7673 usrdata argument is not NULL interpret it as a null terminated pass
7674 phrase. If usrdata and the callback are NULL then the pass phrase
7675 is prompted for as usual.
7678 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7679 the support is automatically enabled. The resulting binaries will
7680 autodetect the card and use it if present.
7681 [Ben Laurie and Compaq Inc.]
7683 *) Work around for Netscape hang bug. This sends certificate request
7684 and server done in one record. Since this is perfectly legal in the
7685 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7686 the bugs/SSLv3 entry for more info.
7689 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7692 *) Add -rand argument to smime and pkcs12 applications and read/write
7696 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7699 *) Add command line password options to the remaining applications.
7702 *) Bug fix for BN_div_recp() for numerators with an even number of
7706 *) More tests in bntest.c, and changed test_bn output.
7709 *) ./config recognizes MacOS X now.
7712 *) Bug fix for BN_div() when the first words of num and divsor are
7713 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7716 *) Add support for various broken PKCS#8 formats, and command line
7717 options to produce them.
7720 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7721 get temporary BIGNUMs from a BN_CTX.
7724 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7728 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7729 include a #define from the old name to the new. The original intent
7730 was that statically linked binaries could for example just call
7731 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7732 link with digests. This never worked becayse SSLeay_add_all_digests()
7733 and SSLeay_add_all_ciphers() were in the same source file so calling
7734 one would link with the other. They are now in separate source files.
7737 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7740 *) Use a less unusual form of the Miller-Rabin primality test (it used
7741 a binary algorithm for exponentiation integrated into the Miller-Rabin
7742 loop, our standard modexp algorithms are faster).
7745 *) Support for the EBCDIC character set completed.
7746 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7748 *) Source code cleanups: use const where appropriate, eliminate casts,
7749 use void * instead of char * in lhash.
7752 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7753 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7754 this the server could overwrite ephemeral keys that the client
7758 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7759 using 50 iterations of the Rabin-Miller test.
7761 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7762 iterations of the Rabin-Miller test as required by the appendix
7763 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7764 As BN_is_prime_fasttest includes trial division, DSA parameter
7765 generation becomes much faster.
7767 This implies a change for the callback functions in DSA_is_prime
7768 and DSA_generate_parameters: The callback function is called once
7769 for each positive witness in the Rabin-Miller test, not just
7770 occasionally in the inner loop; and the parameters to the
7771 callback function now provide an iteration count for the outer
7772 loop rather than for the current invocation of the inner loop.
7773 DSA_generate_parameters additionally can call the callback
7774 function with an 'iteration count' of -1, meaning that a
7775 candidate has passed the trial division test (when q is generated
7776 from an application-provided seed, trial division is skipped).
7779 *) New function BN_is_prime_fasttest that optionally does trial
7780 division before starting the Rabin-Miller test and has
7781 an additional BN_CTX * argument (whereas BN_is_prime always
7782 has to allocate at least one BN_CTX).
7783 'callback(1, -1, cb_arg)' is called when a number has passed the
7784 trial division stage.
7787 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7791 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7794 *) New function BN_pseudo_rand().
7797 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7798 bignum version of BN_from_montgomery() with the working code from
7799 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7803 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7804 made it impossible to use the same SSL_SESSION data structure in
7805 SSL2 clients in multiple threads.
7808 *) The return value of RAND_load_file() no longer counts bytes obtained
7809 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7810 to seed the PRNG (previously an explicit byte count was required).
7811 [Ulf Möller, Bodo Möller]
7813 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7814 used (char *) instead of (void *) and had casts all over the place.
7817 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7820 *) Retain source code compatibility for BN_prime_checks macro:
7821 BN_is_prime(..., BN_prime_checks, ...) now uses
7822 BN_prime_checks_for_size to determine the appropriate number of
7823 Rabin-Miller iterations.
7826 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7827 DH_CHECK_P_NOT_SAFE_PRIME.
7828 (Check if this is true? OpenPGP calls them "strong".)
7831 *) Merge the functionality of "dh" and "gendh" programs into a new program
7832 "dhparam". The old programs are retained for now but will handle DH keys
7833 (instead of parameters) in future.
7836 *) Make the ciphers, s_server and s_client programs check the return values
7837 when a new cipher list is set.
7840 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7841 ciphers. Before when the 56bit ciphers were enabled the sorting was
7844 The syntax for the cipher sorting has been extended to support sorting by
7845 cipher-strength (using the strength_bits hard coded in the tables).
7846 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7848 Fix a bug in the cipher-command parser: when supplying a cipher command
7849 string with an "undefined" symbol (neither command nor alphanumeric
7850 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7851 an error is flagged.
7853 Due to the strength-sorting extension, the code of the
7854 ssl_create_cipher_list() function was completely rearranged. I hope that
7855 the readability was also increased :-)
7856 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7858 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7859 for the first serial number and places 2 in the serial number file. This
7860 avoids problems when the root CA is created with serial number zero and
7861 the first user certificate has the same issuer name and serial number
7865 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7866 the new code. Add documentation for this stuff.
7869 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
7870 X509_*() to X509at_*() on the grounds that they don't handle X509
7871 structures and behave in an analagous way to the X509v3 functions:
7872 they shouldn't be called directly but wrapper functions should be used
7875 So we also now have some wrapper functions that call the X509at functions
7876 when passed certificate requests. (TO DO: similar things can be done with
7877 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
7878 things. Some of these need some d2i or i2d and print functionality
7879 because they handle more complex structures.)
7882 *) Add missing #ifndefs that caused missing symbols when building libssl
7883 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
7884 NO_RSA in ssl/s2*.c.
7885 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
7887 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
7888 has a return value which indicates the quality of the random data
7889 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
7890 error queue. New function RAND_pseudo_bytes() generates output that is
7891 guaranteed to be unique but not unpredictable. RAND_add is like
7892 RAND_seed, but takes an extra argument for an entropy estimate
7893 (RAND_seed always assumes full entropy).
7896 *) Do more iterations of Rabin-Miller probable prime test (specifically,
7897 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
7898 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
7899 in crypto/bn/bn_prime.c for the complete table). This guarantees a
7900 false-positive rate of at most 2^-80 for random input.
7903 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
7906 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
7907 in the 0.9.5 release), this returns the chain
7908 from an X509_CTX structure with a dup of the stack and all
7909 the X509 reference counts upped: so the stack will exist
7910 after X509_CTX_cleanup() has been called. Modify pkcs12.c
7913 Also make SSL_SESSION_print() print out the verify return
7917 *) Add manpage for the pkcs12 command. Also change the default
7918 behaviour so MAC iteration counts are used unless the new
7919 -nomaciter option is used. This improves file security and
7920 only older versions of MSIE (4.0 for example) need it.
7923 *) Honor the no-xxx Configure options when creating .DEF files.
7926 *) Add PKCS#10 attributes to field table: challengePassword,
7927 unstructuredName and unstructuredAddress. These are taken from
7928 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
7929 international characters are used.
7931 More changes to X509_ATTRIBUTE code: allow the setting of types
7932 based on strings. Remove the 'loc' parameter when adding
7933 attributes because these will be a SET OF encoding which is sorted
7937 *) Initial changes to the 'req' utility to allow request generation
7938 automation. This will allow an application to just generate a template
7939 file containing all the field values and have req construct the
7942 Initial support for X509_ATTRIBUTE handling. Stacks of these are
7943 used all over the place including certificate requests and PKCS#7
7944 structures. They are currently handled manually where necessary with
7945 some primitive wrappers for PKCS#7. The new functions behave in a
7946 manner analogous to the X509 extension functions: they allow
7947 attributes to be looked up by NID and added.
7949 Later something similar to the X509V3 code would be desirable to
7950 automatically handle the encoding, decoding and printing of the
7951 more complex types. The string types like challengePassword can
7952 be handled by the string table functions.
7954 Also modified the multi byte string table handling. Now there is
7955 a 'global mask' which masks out certain types. The table itself
7956 can use the flag STABLE_NO_MASK to ignore the mask setting: this
7957 is useful when for example there is only one permissible type
7958 (as in countryName) and using the mask might result in no valid
7962 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
7963 SSL_get_peer_finished to allow applications to obtain the latest
7964 Finished messages sent to the peer or expected from the peer,
7965 respectively. (SSL_get_peer_finished is usually the Finished message
7966 actually received from the peer, otherwise the protocol will be aborted.)
7968 As the Finished message are message digests of the complete handshake
7969 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
7970 be used for external authentication procedures when the authentication
7971 provided by SSL/TLS is not desired or is not enough.
7974 *) Enhanced support for Alpha Linux is added. Now ./config checks if
7975 the host supports BWX extension and if Compaq C is present on the
7976 $PATH. Just exploiting of the BWX extension results in 20-30%
7977 performance kick for some algorithms, e.g. DES and RC4 to mention
7978 a couple. Compaq C in turn generates ~20% faster code for MD5 and
7982 *) Add support for MS "fast SGC". This is arguably a violation of the
7983 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
7984 weak crypto and after checking the certificate is SGC a second one
7985 with strong crypto. MS SGC stops the first handshake after receiving
7986 the server certificate message and sends a second client hello. Since
7987 a server will typically do all the time consuming operations before
7988 expecting any further messages from the client (server key exchange
7989 is the most expensive) there is little difference between the two.
7991 To get OpenSSL to support MS SGC we have to permit a second client
7992 hello message after we have sent server done. In addition we have to
7993 reset the MAC if we do get this second client hello.
7996 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
7997 if a DER encoded private key is RSA or DSA traditional format. Changed
7998 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
7999 format DER encoded private key. Newer code should use PKCS#8 format which
8000 has the key type encoded in the ASN1 structure. Added DER private key
8001 support to pkcs8 application.
8004 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8005 ciphersuites has been selected (as required by the SSL 3/TLS 1
8006 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8007 is set, we interpret this as a request to violate the specification
8008 (the worst that can happen is a handshake failure, and 'correct'
8009 behaviour would result in a handshake failure anyway).
8012 *) In SSL_CTX_add_session, take into account that there might be multiple
8013 SSL_SESSION structures with the same session ID (e.g. when two threads
8014 concurrently obtain them from an external cache).
8015 The internal cache can handle only one SSL_SESSION with a given ID,
8016 so if there's a conflict, we now throw out the old one to achieve
8020 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8021 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8022 some routines that use cipher OIDs: some ciphers do not have OIDs
8023 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8027 *) Simplify the trust setting structure and code. Now we just have
8028 two sequences of OIDs for trusted and rejected settings. These will
8029 typically have values the same as the extended key usage extension
8030 and any application specific purposes.
8032 The trust checking code now has a default behaviour: it will just
8033 check for an object with the same NID as the passed id. Functions can
8034 be provided to override either the default behaviour or the behaviour
8035 for a given id. SSL client, server and email already have functions
8036 in place for compatibility: they check the NID and also return "trusted"
8037 if the certificate is self signed.
8040 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8041 traditional format into an EVP_PKEY structure.
8044 *) Add a password callback function PEM_cb() which either prompts for
8045 a password if usr_data is NULL or otherwise assumes it is a null
8046 terminated password. Allow passwords to be passed on command line
8047 environment or config files in a few more utilities.
8050 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8051 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8052 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8053 Update documentation.
8056 *) Support for ASN1 "NULL" type. This could be handled before by using
8057 ASN1_TYPE but there wasn't any function that would try to read a NULL
8058 and produce an error if it couldn't. For compatibility we also have
8059 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8060 don't allocate anything because they don't need to.
8063 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8065 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8067 *) Rebuild of the memory allocation routines used by OpenSSL code and
8068 possibly others as well. The purpose is to make an interface that
8069 provide hooks so anyone can build a separate set of allocation and
8070 deallocation routines to be used by OpenSSL, for example memory
8071 pool implementations, or something else, which was previously hard
8072 since Malloc(), Realloc() and Free() were defined as macros having
8073 the values malloc, realloc and free, respectively (except for Win32
8074 compilations). The same is provided for memory debugging code.
8075 OpenSSL already comes with functionality to find memory leaks, but
8076 this gives people a chance to debug other memory problems.
8078 With these changes, a new set of functions and macros have appeared:
8080 CRYPTO_set_mem_debug_functions() [F]
8081 CRYPTO_get_mem_debug_functions() [F]
8082 CRYPTO_dbg_set_options() [F]
8083 CRYPTO_dbg_get_options() [F]
8084 CRYPTO_malloc_debug_init() [M]
8086 The memory debug functions are NULL by default, unless the library
8087 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8088 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8089 gives the standard debugging functions that come with OpenSSL) or
8090 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8091 provided by the library user) must be used. When the standard
8092 debugging functions are used, CRYPTO_dbg_set_options can be used to
8093 request additional information:
8094 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8095 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8097 Also, things like CRYPTO_set_mem_functions will always give the
8098 expected result (the new set of functions is used for allocation
8099 and deallocation) at all times, regardless of platform and compiler
8102 To finish it up, some functions that were never use in any other
8103 way than through macros have a new API and new semantic:
8106 CRYPTO_dbg_realloc()
8109 All macros of value have retained their old syntax.
8110 [Richard Levitte and Bodo Moeller]
8112 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8113 ordering of SMIMECapabilities wasn't in "strength order" and there
8114 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8118 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8119 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8120 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
8122 *) Merge in my S/MIME library for OpenSSL. This provides a simple
8123 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
8124 functionality to handle multipart/signed properly) and a utility
8125 called 'smime' to call all this stuff. This is based on code I
8126 originally wrote for Celo who have kindly allowed it to be
8127 included in OpenSSL.
8130 *) Add variants des_set_key_checked and des_set_key_unchecked of
8131 des_set_key (aka des_key_sched). Global variable des_check_key
8132 decides which of these is called by des_set_key; this way
8133 des_check_key behaves as it always did, but applications and
8134 the library itself, which was buggy for des_check_key == 1,
8135 have a cleaner way to pick the version they need.
8138 *) New function PKCS12_newpass() which changes the password of a
8142 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
8143 dynamic mix. In both cases the ids can be used as an index into the
8144 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
8145 functions so they accept a list of the field values and the
8146 application doesn't need to directly manipulate the X509_TRUST
8150 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
8154 *) Modify the way the V3 extension code looks up extensions. This now
8155 works in a similar way to the object code: we have some "standard"
8156 extensions in a static table which is searched with OBJ_bsearch()
8157 and the application can add dynamic ones if needed. The file
8158 crypto/x509v3/ext_dat.h now has the info: this file needs to be
8159 updated whenever a new extension is added to the core code and kept
8160 in ext_nid order. There is a simple program 'tabtest.c' which checks
8161 this. New extensions are not added too often so this file can readily
8162 be maintained manually.
8164 There are two big advantages in doing things this way. The extensions
8165 can be looked up immediately and no longer need to be "added" using
8166 X509V3_add_standard_extensions(): this function now does nothing.
8167 [Side note: I get *lots* of email saying the extension code doesn't
8168 work because people forget to call this function]
8169 Also no dynamic allocation is done unless new extensions are added:
8170 so if we don't add custom extensions there is no need to call
8171 X509V3_EXT_cleanup().
8174 *) Modify enc utility's salting as follows: make salting the default. Add a
8175 magic header, so unsalted files fail gracefully instead of just decrypting
8176 to garbage. This is because not salting is a big security hole, so people
8177 should be discouraged from doing it.
8180 *) Fixes and enhancements to the 'x509' utility. It allowed a message
8181 digest to be passed on the command line but it only used this
8182 parameter when signing a certificate. Modified so all relevant
8183 operations are affected by the digest parameter including the
8184 -fingerprint and -x509toreq options. Also -x509toreq choked if a
8185 DSA key was used because it didn't fix the digest.
8188 *) Initial certificate chain verify code. Currently tests the untrusted
8189 certificates for consistency with the verify purpose (which is set
8190 when the X509_STORE_CTX structure is set up) and checks the pathlength.
8192 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
8193 this is because it will reject chains with invalid extensions whereas
8194 every previous version of OpenSSL and SSLeay made no checks at all.
8196 Trust code: checks the root CA for the relevant trust settings. Trust
8197 settings have an initial value consistent with the verify purpose: e.g.
8198 if the verify purpose is for SSL client use it expects the CA to be
8199 trusted for SSL client use. However the default value can be changed to
8200 permit custom trust settings: one example of this would be to only trust
8201 certificates from a specific "secure" set of CAs.
8203 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
8204 which should be used for version portability: especially since the
8205 verify structure is likely to change more often now.
8207 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
8208 to set them. If not set then assume SSL clients will verify SSL servers
8211 Two new options to the verify program: -untrusted allows a set of
8212 untrusted certificates to be passed in and -purpose which sets the
8213 intended purpose of the certificate. If a purpose is set then the
8214 new chain verify code is used to check extension consistency.
8217 *) Support for the authority information access extension.
8220 *) Modify RSA and DSA PEM read routines to transparently handle
8221 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
8222 public keys in a format compatible with certificate
8223 SubjectPublicKeyInfo structures. Unfortunately there were already
8224 functions called *_PublicKey_* which used various odd formats so
8225 these are retained for compatibility: however the DSA variants were
8226 never in a public release so they have been deleted. Changed dsa/rsa
8227 utilities to handle the new format: note no releases ever handled public
8228 keys so we should be OK.
8230 The primary motivation for this change is to avoid the same fiasco
8231 that dogs private keys: there are several incompatible private key
8232 formats some of which are standard and some OpenSSL specific and
8233 require various evil hacks to allow partial transparent handling and
8234 even then it doesn't work with DER formats. Given the option anything
8235 other than PKCS#8 should be dumped: but the other formats have to
8236 stay in the name of compatibility.
8238 With public keys and the benefit of hindsight one standard format
8239 is used which works with EVP_PKEY, RSA or DSA structures: though
8240 it clearly returns an error if you try to read the wrong kind of key.
8242 Added a -pubkey option to the 'x509' utility to output the public key.
8243 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
8244 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
8245 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
8246 that do the same as the EVP_PKEY_assign_*() except they up the
8247 reference count of the added key (they don't "swallow" the
8251 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
8252 CRLs would fail if the file contained no certificates or no CRLs:
8253 added a new function to read in both types and return the number
8254 read: this means that if none are read it will be an error. The
8255 DER versions of the certificate and CRL reader would always fail
8256 because it isn't possible to mix certificates and CRLs in DER format
8257 without choking one or the other routine. Changed this to just read
8258 a certificate: this is the best we can do. Also modified the code
8259 in apps/verify.c to take notice of return codes: it was previously
8260 attempting to read in certificates from NULL pointers and ignoring
8261 any errors: this is one reason why the cert and CRL reader seemed
8262 to work. It doesn't check return codes from the default certificate
8263 routines: these may well fail if the certificates aren't installed.
8266 *) Code to support otherName option in GeneralName.
8269 *) First update to verify code. Change the verify utility
8270 so it warns if it is passed a self signed certificate:
8271 for consistency with the normal behaviour. X509_verify
8272 has been modified to it will now verify a self signed
8273 certificate if *exactly* the same certificate appears
8274 in the store: it was previously impossible to trust a
8275 single self signed certificate. This means that:
8276 openssl verify ss.pem
8277 now gives a warning about a self signed certificate but
8278 openssl verify -CAfile ss.pem ss.pem
8282 *) For servers, store verify_result in SSL_SESSION data structure
8283 (and add it to external session representation).
8284 This is needed when client certificate verifications fails,
8285 but an application-provided verification callback (set by
8286 SSL_CTX_set_cert_verify_callback) allows accepting the session
8287 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
8288 but returns 1): When the session is reused, we have to set
8289 ssl->verify_result to the appropriate error code to avoid
8291 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
8293 *) Fix a bug in the new PKCS#7 code: it didn't consider the
8294 case in PKCS7_dataInit() where the signed PKCS7 structure
8295 didn't contain any existing data because it was being created.
8296 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
8298 *) Add a salt to the key derivation routines in enc.c. This
8299 forms the first 8 bytes of the encrypted file. Also add a
8300 -S option to allow a salt to be input on the command line.
8303 *) New function X509_cmp(). Oddly enough there wasn't a function
8304 to compare two certificates. We do this by working out the SHA1
8305 hash and comparing that. X509_cmp() will be needed by the trust
8309 *) SSL_get1_session() is like SSL_get_session(), but increments
8310 the reference count in the SSL_SESSION returned.
8311 [Geoff Thorpe <geoff@eu.c2.net>]
8313 *) Fix for 'req': it was adding a null to request attributes.
8314 Also change the X509_LOOKUP and X509_INFO code to handle
8315 certificate auxiliary information.
8318 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8322 *) Add the possibility to add extra information to the memory leak
8323 detecting output, to form tracebacks, showing from where each
8324 allocation was originated: CRYPTO_push_info("constant string") adds
8325 the string plus current file name and line number to a per-thread
8326 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8327 is like calling CYRPTO_pop_info() until the stack is empty.
8328 Also updated memory leak detection code to be multi-thread-safe.
8331 *) Add options -text and -noout to pkcs7 utility and delete the
8332 encryption options which never did anything. Update docs.
8335 *) Add options to some of the utilities to allow the pass phrase
8336 to be included on either the command line (not recommended on
8337 OSes like Unix) or read from the environment. Update the
8338 manpages and fix a few bugs.
8341 *) Add a few manpages for some of the openssl commands.
8344 *) Fix the -revoke option in ca. It was freeing up memory twice,
8345 leaking and not finding already revoked certificates.
8348 *) Extensive changes to support certificate auxiliary information.
8349 This involves the use of X509_CERT_AUX structure and X509_AUX
8350 functions. An X509_AUX function such as PEM_read_X509_AUX()
8351 can still read in a certificate file in the usual way but it
8352 will also read in any additional "auxiliary information". By
8353 doing things this way a fair degree of compatibility can be
8354 retained: existing certificates can have this information added
8355 using the new 'x509' options.
8357 Current auxiliary information includes an "alias" and some trust
8358 settings. The trust settings will ultimately be used in enhanced
8359 certificate chain verification routines: currently a certificate
8360 can only be trusted if it is self signed and then it is trusted
8364 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8365 The problem was that one of the replacement routines had not been working
8366 since SSLeay releases. For now the offending routine has been replaced
8367 with non-optimised assembler. Even so, this now gives around 95%
8368 performance improvement for 1024 bit RSA signs.
8371 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8372 handling. Most clients have the effective key size in bits equal to
8373 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8374 A few however don't do this and instead use the size of the decrypted key
8375 to determine the RC2 key length and the AlgorithmIdentifier to determine
8376 the effective key length. In this case the effective key length can still
8377 be 40 bits but the key length can be 168 bits for example. This is fixed
8378 by manually forcing an RC2 key into the EVP_PKEY structure because the
8379 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8380 the key length and effective key length are equal.
8383 *) Add a bunch of functions that should simplify the creation of
8384 X509_NAME structures. Now you should be able to do:
8385 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8386 and have it automatically work out the correct field type and fill in
8387 the structures. The more adventurous can try:
8388 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8389 and it will (hopefully) work out the correct multibyte encoding.
8392 *) Change the 'req' utility to use the new field handling and multibyte
8393 copy routines. Before the DN field creation was handled in an ad hoc
8394 way in req, ca, and x509 which was rather broken and didn't support
8395 BMPStrings or UTF8Strings. Since some software doesn't implement
8396 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8397 using the dirstring_type option. See the new comment in the default
8398 openssl.cnf for more info.
8401 *) Make crypto/rand/md_rand.c more robust:
8402 - Assure unique random numbers after fork().
8403 - Make sure that concurrent threads access the global counter and
8404 md serializably so that we never lose entropy in them
8405 or use exactly the same state in multiple threads.
8406 Access to the large state is not always serializable because
8407 the additional locking could be a performance killer, and
8408 md should be large enough anyway.
8411 *) New file apps/app_rand.c with commonly needed functionality
8412 for handling the random seed file.
8414 Use the random seed file in some applications that previously did not:
8416 dsaparam -genkey (which also ignored its '-rand' option),
8419 x509 (when signing).
8420 Except on systems with /dev/urandom, it is crucial to have a random
8421 seed file at least for key creation, DSA signing, and for DH exchanges;
8422 for RSA signatures we could do without one.
8424 gendh and gendsa (unlike genrsa) used to read only the first byte
8425 of each file listed in the '-rand' option. The function as previously
8426 found in genrsa is now in app_rand.c and is used by all programs
8427 that support '-rand'.
8430 *) In RAND_write_file, use mode 0600 for creating files;
8431 don't just chmod when it may be too late.
8434 *) Report an error from X509_STORE_load_locations
8435 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8438 *) New function ASN1_mbstring_copy() this copies a string in either
8439 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8440 into an ASN1_STRING type. A mask of permissible types is passed
8441 and it chooses the "minimal" type to use or an error if not type
8445 *) Add function equivalents to the various macros in asn1.h. The old
8446 macros are retained with an M_ prefix. Code inside the library can
8447 use the M_ macros. External code (including the openssl utility)
8448 should *NOT* in order to be "shared library friendly".
8451 *) Add various functions that can check a certificate's extensions
8452 to see if it usable for various purposes such as SSL client,
8453 server or S/MIME and CAs of these types. This is currently
8454 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8455 verification. Also added a -purpose flag to x509 utility to
8456 print out all the purposes.
8459 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8463 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8464 for, obtain and decode and extension and obtain its critical flag.
8465 This allows all the necessary extension code to be handled in a
8466 single function call.
8469 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8470 platforms. See crypto/rc4/rc4_enc.c for further details.
8473 *) New -noout option to asn1parse. This causes no output to be produced
8474 its main use is when combined with -strparse and -out to extract data
8475 from a file (which may not be in ASN.1 format).
8478 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8479 when producing the local key id.
8480 [Richard Levitte <levitte@stacken.kth.se>]
8482 *) New option -dhparam in s_server. This allows a DH parameter file to be
8483 stated explicitly. If it is not stated then it tries the first server
8484 certificate file. The previous behaviour hard coded the filename
8488 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8489 a public key to be input or output. For example:
8490 openssl rsa -in key.pem -pubout -out pubkey.pem
8491 Also added necessary DSA public key functions to handle this.
8494 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8495 in the message. This was handled by allowing
8496 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8497 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8499 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8500 to the end of the strings whereas this didn't. This would cause problems
8501 if strings read with d2i_ASN1_bytes() were later modified.
8502 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8504 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8505 data and it contains EOF it will end up returning an error. This is
8506 caused by input 46 bytes long. The cause is due to the way base64
8507 BIOs find the start of base64 encoded data. They do this by trying a
8508 trial decode on each line until they find one that works. When they
8509 do a flag is set and it starts again knowing it can pass all the
8510 data directly through the decoder. Unfortunately it doesn't reset
8511 the context it uses. This means that if EOF is reached an attempt
8512 is made to pass two EOFs through the context and this causes the
8513 resulting error. This can also cause other problems as well. As is
8514 usual with these problems it takes *ages* to find and the fix is
8515 trivial: move one line.
8516 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8518 *) Ugly workaround to get s_client and s_server working under Windows. The
8519 old code wouldn't work because it needed to select() on sockets and the
8520 tty (for keypresses and to see if data could be written). Win32 only
8521 supports select() on sockets so we select() with a 1s timeout on the
8522 sockets and then see if any characters are waiting to be read, if none
8523 are present then we retry, we also assume we can always write data to
8524 the tty. This isn't nice because the code then blocks until we've
8525 received a complete line of data and it is effectively polling the
8526 keyboard at 1s intervals: however it's quite a bit better than not
8527 working at all :-) A dedicated Windows application might handle this
8528 with an event loop for example.
8531 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8532 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8533 will be called when RSA_sign() and RSA_verify() are used. This is useful
8534 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8535 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8536 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8537 This necessitated the support of an extra signature type NID_md5_sha1
8538 for SSL signatures and modifications to the SSL library to use it instead
8539 of calling RSA_public_decrypt() and RSA_private_encrypt().
8542 *) Add new -verify -CAfile and -CApath options to the crl program, these
8543 will lookup a CRL issuers certificate and verify the signature in a
8544 similar way to the verify program. Tidy up the crl program so it
8545 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8546 less strict. It will now permit CRL extensions even if it is not
8547 a V2 CRL: this will allow it to tolerate some broken CRLs.
8550 *) Initialize all non-automatic variables each time one of the openssl
8551 sub-programs is started (this is necessary as they may be started
8552 multiple times from the "OpenSSL>" prompt).
8553 [Lennart Bang, Bodo Moeller]
8555 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8556 removing all other RSA functionality (this is what NO_RSA does). This
8557 is so (for example) those in the US can disable those operations covered
8558 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8562 *) Non-copying interface to BIO pairs.
8563 (still largely untested)
8566 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8567 ASCII string. This was handled independently in various places before.
8570 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8571 UTF8 strings a character at a time.
8574 *) Use client_version from client hello to select the protocol
8575 (s23_srvr.c) and for RSA client key exchange verification
8576 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8579 *) Add various utility functions to handle SPKACs, these were previously
8580 handled by poking round in the structure internals. Added new function
8581 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8582 print, verify and generate SPKACs. Based on an original idea from
8583 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8586 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8589 *) Allow the config file extension section to be overwritten on the
8590 command line. Based on an original idea from Massimiliano Pala
8591 <madwolf@comune.modena.it>. The new option is called -extensions
8592 and can be applied to ca, req and x509. Also -reqexts to override
8593 the request extensions in req and -crlexts to override the crl extensions
8597 *) Add new feature to the SPKAC handling in ca. Now you can include
8598 the same field multiple times by preceding it by "XXXX." for example:
8601 this is the same syntax as used in the req config file.
8604 *) Allow certificate extensions to be added to certificate requests. These
8605 are specified in a 'req_extensions' option of the req section of the
8606 config file. They can be printed out with the -text option to req but
8607 are otherwise ignored at present.
8610 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8611 data read consists of only the final block it would not decrypted because
8612 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8613 A misplaced 'break' also meant the decrypted final block might not be
8614 copied until the next read.
8617 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8618 a few extra parameters to the DH structure: these will be useful if
8619 for example we want the value of 'q' or implement X9.42 DH.
8622 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8623 provides hooks that allow the default DSA functions or functions on a
8624 "per key" basis to be replaced. This allows hardware acceleration and
8625 hardware key storage to be handled without major modification to the
8626 library. Also added low level modexp hooks and CRYPTO_EX structure and
8627 associated functions.
8630 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8631 as "read only": it can't be written to and the buffer it points to will
8632 not be freed. Reading from a read only BIO is much more efficient than
8633 a normal memory BIO. This was added because there are several times when
8634 an area of memory needs to be read from a BIO. The previous method was
8635 to create a memory BIO and write the data to it, this results in two
8636 copies of the data and an O(n^2) reading algorithm. There is a new
8637 function BIO_new_mem_buf() which creates a read only memory BIO from
8638 an area of memory. Also modified the PKCS#7 routines to use read only
8642 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8643 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8644 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8645 but a retry condition occured while trying to read the rest.
8648 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8649 NID_pkcs7_encrypted by default: this was wrong since this should almost
8650 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8651 the encrypted data type: this is a more sensible place to put it and it
8652 allows the PKCS#12 code to be tidied up that duplicated this
8656 *) Changed obj_dat.pl script so it takes its input and output files on
8657 the command line. This should avoid shell escape redirection problems
8661 *) Initial support for certificate extension requests, these are included
8662 in things like Xenroll certificate requests. Included functions to allow
8663 extensions to be obtained and added.
8666 *) -crlf option to s_client and s_server for sending newlines as
8667 CRLF (as required by many protocols).
8670 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8672 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8673 [Ralf S. Engelschall]
8675 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8676 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8678 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8682 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8683 DH parameters/keys (q is lost during that conversion, but the resulting
8684 DH parameters contain its length).
8686 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8687 much faster than DH_generate_parameters (which creates parameters
8688 where p = 2*q + 1), and also the smaller q makes DH computations
8689 much more efficient (160-bit exponentiation instead of 1024-bit
8690 exponentiation); so this provides a convenient way to support DHE
8691 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8692 utter importance to use
8693 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8695 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8696 when such DH parameters are used, because otherwise small subgroup
8697 attacks may become possible!
8700 *) Avoid memory leak in i2d_DHparams.
8703 *) Allow the -k option to be used more than once in the enc program:
8704 this allows the same encrypted message to be read by multiple recipients.
8707 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8708 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8709 it will always use the numerical form of the OID, even if it has a short
8713 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8714 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8715 otherwise bn_mod_exp was called. In the case of hardware keys for example
8716 no private key components need be present and it might store extra data
8717 in the RSA structure, which cannot be accessed from bn_mod_exp.
8718 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8719 private key operations.
8722 *) Added support for SPARC Linux.
8725 *) pem_password_cb function type incompatibly changed from
8726 typedef int pem_password_cb(char *buf, int size, int rwflag);
8728 ....(char *buf, int size, int rwflag, void *userdata);
8729 so that applications can pass data to their callbacks:
8730 The PEM[_ASN1]_{read,write}... functions and macros now take an
8731 additional void * argument, which is just handed through whenever
8732 the password callback is called.
8733 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8735 New function SSL_CTX_set_default_passwd_cb_userdata.
8737 Compatibility note: As many C implementations push function arguments
8738 onto the stack in reverse order, the new library version is likely to
8739 interoperate with programs that have been compiled with the old
8740 pem_password_cb definition (PEM_whatever takes some data that
8741 happens to be on the stack as its last argument, and the callback
8742 just ignores this garbage); but there is no guarantee whatsoever that
8745 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8746 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8747 problems not only on Windows, but also on some Unix platforms.
8748 To avoid problematic command lines, these definitions are now in an
8749 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8750 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8753 *) MIPS III/IV assembler module is reimplemented.
8756 *) More DES library cleanups: remove references to srand/rand and
8757 delete an unused file.
8760 *) Add support for the the free Netwide assembler (NASM) under Win32,
8761 since not many people have MASM (ml) and it can be hard to obtain.
8762 This is currently experimental but it seems to work OK and pass all
8763 the tests. Check out INSTALL.W32 for info.
8766 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8767 without temporary keys kept an extra copy of the server key,
8768 and connections with temporary keys did not free everything in case
8772 *) New function RSA_check_key and new openssl rsa option -check
8773 for verifying the consistency of RSA keys.
8774 [Ulf Moeller, Bodo Moeller]
8776 *) Various changes to make Win32 compile work:
8777 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8778 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8779 comparison" warnings.
8780 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8783 *) Add a debugging option to PKCS#5 v2 key generation function: when
8784 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8785 derived keys are printed to stderr.
8788 *) Copy the flags in ASN1_STRING_dup().
8789 [Roman E. Pavlov <pre@mo.msk.ru>]
8791 *) The x509 application mishandled signing requests containing DSA
8792 keys when the signing key was also DSA and the parameters didn't match.
8794 It was supposed to omit the parameters when they matched the signing key:
8795 the verifying software was then supposed to automatically use the CA's
8796 parameters if they were absent from the end user certificate.
8798 Omitting parameters is no longer recommended. The test was also
8799 the wrong way round! This was probably due to unusual behaviour in
8800 EVP_cmp_parameters() which returns 1 if the parameters match.
8801 This meant that parameters were omitted when they *didn't* match and
8802 the certificate was useless. Certificates signed with 'ca' didn't have
8804 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8806 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8807 The interface is as follows:
8808 Applications can use
8809 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8810 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8811 "off" is now the default.
8812 The library internally uses
8813 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8814 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8815 to disable memory-checking temporarily.
8817 Some inconsistent states that previously were possible (and were
8818 even the default) are now avoided.
8820 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8821 with each memory chunk allocated; this is occasionally more helpful
8822 than just having a counter.
8824 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8826 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8830 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8831 which largely parallels "options", but is for changing API behaviour,
8832 whereas "options" are about protocol behaviour.
8833 Initial "mode" flags are:
8835 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8836 a single record has been written.
8837 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8838 retries use the same buffer location.
8839 (But all of the contents must be
8843 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8846 *) Fix problems with no-hmac etc.
8847 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8849 *) New functions RSA_get_default_method(), RSA_set_method() and
8850 RSA_get_method(). These allows replacement of RSA_METHODs without having
8851 to mess around with the internals of an RSA structure.
8854 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8855 Also really enable memory leak checks in openssl.c and in some
8857 [Chad C. Mulligan, Bodo Moeller]
8859 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8860 up the length of negative integers. This has now been simplified to just
8861 store the length when it is first determined and use it later, rather
8862 than trying to keep track of where data is copied and updating it to
8864 [Steve Henson, reported by Brien Wheeler
8865 <bwheeler@authentica-security.com>]
8867 *) Add a new function PKCS7_signatureVerify. This allows the verification
8868 of a PKCS#7 signature but with the signing certificate passed to the
8869 function itself. This contrasts with PKCS7_dataVerify which assumes the
8870 certificate is present in the PKCS#7 structure. This isn't always the
8871 case: certificates can be omitted from a PKCS#7 structure and be
8872 distributed by "out of band" means (such as a certificate database).
8875 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
8876 function prototypes in pem.h, also change util/mkdef.pl to add the
8877 necessary function names.
8880 *) mk1mf.pl (used by Windows builds) did not properly read the
8881 options set by Configure in the top level Makefile, and Configure
8882 was not even able to write more than one option correctly.
8883 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
8886 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
8887 file to be loaded from a BIO or FILE pointer. The BIO version will
8888 for example allow memory BIOs to contain config info.
8891 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
8892 Whoever hopes to achieve shared-library compatibility across versions
8893 must use this, not the compile-time macro.
8894 (Exercise 0.9.4: Which is the minimum library version required by
8896 Note: All this applies only to multi-threaded programs, others don't
8900 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
8901 through a BIO pair triggered the default case, i.e.
8902 SSLerr(...,SSL_R_UNKNOWN_STATE).
8905 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
8906 can use the SSL library even if none of the specific BIOs is
8910 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
8911 for the encoded length.
8912 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
8914 *) Add initial documentation of the X509V3 functions.
8917 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
8918 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
8919 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
8920 secure PKCS#8 private key format with a high iteration count.
8923 *) Fix determination of Perl interpreter: A perl or perl5
8924 _directory_ in $PATH was also accepted as the interpreter.
8925 [Ralf S. Engelschall]
8927 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
8928 wrong with it but it was very old and did things like calling
8929 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
8933 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
8934 to use the new extension code.
8937 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
8938 with macros. This should make it easier to change their form, add extra
8939 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
8943 *) Add to configuration table a new entry that can specify an alternative
8944 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
8945 according to Mark Crispin <MRC@Panda.COM>.
8949 *) DES CBC did not update the IV. Weird.
8952 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
8953 Changing the behaviour of the former might break existing programs --
8954 where IV updating is needed, des_ncbc_encrypt can be used.
8957 *) When bntest is run from "make test" it drives bc to check its
8958 calculations, as well as internally checking them. If an internal check
8959 fails, it needs to cause bc to give a non-zero result or make test carries
8960 on without noticing the failure. Fixed.
8963 *) DES library cleanups.
8966 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
8967 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
8968 ciphers. NOTE: although the key derivation function has been verified
8969 against some published test vectors it has not been extensively tested
8970 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
8974 *) Instead of "mkdir -p", which is not fully portable, use new
8975 Perl script "util/mkdir-p.pl".
8978 *) Rewrite the way password based encryption (PBE) is handled. It used to
8979 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
8980 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
8981 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
8982 the 'parameter' field of the AlgorithmIdentifier is passed to the
8983 underlying key generation function so it must do its own ASN1 parsing.
8984 This has also changed the EVP_PBE_CipherInit() function which now has a
8985 'parameter' argument instead of literal salt and iteration count values
8986 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
8989 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
8990 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
8991 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
8992 KEY" because this clashed with PKCS#8 unencrypted string. Since this
8993 value was just used as a "magic string" and not used directly its
8994 value doesn't matter.
8997 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9001 *) "linux-sparc64" configuration (ultrapenguin).
9002 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9003 "linux-sparc" configuration.
9004 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9006 *) config now generates no-xxx options for missing ciphers.
9009 *) Support the EBCDIC character set (work in progress).
9010 File ebcdic.c not yet included because it has a different license.
9011 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9013 *) Support BS2000/OSD-POSIX.
9014 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9016 *) Make callbacks for key generation use void * instead of char *.
9019 *) Make S/MIME samples compile (not yet tested).
9022 *) Additional typesafe stacks.
9025 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9029 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9031 *) New configuration variant "sco5-gcc".
9033 *) Updated some demos.
9034 [Sean O Riordain, Wade Scholine]
9036 *) Add missing BIO_free at exit of pkcs12 application.
9039 *) Fix memory leak in conf.c.
9042 *) Updates for Win32 to assembler version of MD5.
9045 *) Set #! path to perl in apps/der_chop to where we found it
9046 instead of using a fixed path.
9049 *) SHA library changes for irix64-mips4-cc.
9052 *) Improvements for VMS support.
9056 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9058 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9059 This also avoids the problems with SC4.2 and unpatched SC5.
9060 [Andy Polyakov <appro@fy.chalmers.se>]
9062 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9063 These are required because of the typesafe stack would otherwise break
9064 existing code. If old code used a structure member which used to be STACK
9065 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9066 sk_num or sk_value it would produce an error because the num, data members
9067 are not present in STACK_OF. Now it just produces a warning. sk_set
9068 replaces the old method of assigning a value to sk_value
9069 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9070 that does this will no longer work (and should use sk_set instead) but
9071 this could be regarded as a "questionable" behaviour anyway.
9074 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9075 correctly handle encrypted S/MIME data.
9078 *) Change type of various DES function arguments from des_cblock
9079 (which means, in function argument declarations, pointer to char)
9080 to des_cblock * (meaning pointer to array with 8 char elements),
9081 which allows the compiler to do more typechecking; it was like
9082 that back in SSLeay, but with lots of ugly casts.
9084 Introduce new type const_des_cblock.
9087 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9088 problems: find RecipientInfo structure that matches recipient certificate
9089 and initialise the ASN1 structures properly based on passed cipher.
9092 *) Belatedly make the BN tests actually check the results.
9095 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9096 to and from BNs: it was completely broken. New compilation option
9097 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9098 key elements as negative integers.
9101 *) Reorganize and speed up MD5.
9102 [Andy Polyakov <appro@fy.chalmers.se>]
9105 [Richard Levitte <richard@levitte.org>]
9107 *) New option -out to asn1parse to allow the parsed structure to be
9108 output to a file. This is most useful when combined with the -strparse
9109 option to examine the output of things like OCTET STRINGS.
9112 *) Make SSL library a little more fool-proof by not requiring any longer
9113 that SSL_set_{accept,connect}_state be called before
9114 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9115 in many applications because usually everything *appeared* to work as
9116 intended anyway -- now it really works as intended).
9119 *) Move openssl.cnf out of lib/.
9122 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
9123 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
9124 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
9125 [Ralf S. Engelschall]
9127 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
9128 handle PKCS#7 enveloped data properly.
9129 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
9131 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
9132 copying pointers. The cert_st handling is changed by this in
9133 various ways (and thus what used to be known as ctx->default_cert
9134 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
9135 any longer when s->cert does not give us what we need).
9136 ssl_cert_instantiate becomes obsolete by this change.
9137 As soon as we've got the new code right (possibly it already is?),
9138 we have solved a couple of bugs of the earlier code where s->cert
9139 was used as if it could not have been shared with other SSL structures.
9141 Note that using the SSL API in certain dirty ways now will result
9142 in different behaviour than observed with earlier library versions:
9143 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
9144 does not influence s as it used to.
9146 In order to clean up things more thoroughly, inside SSL_SESSION
9147 we don't use CERT any longer, but a new structure SESS_CERT
9148 that holds per-session data (if available); currently, this is
9149 the peer's certificate chain and, for clients, the server's certificate
9150 and temporary key. CERT holds only those values that can have
9151 meaningful defaults in an SSL_CTX.
9154 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
9155 from the internal representation. Various PKCS#7 fixes: remove some
9156 evil casts and set the enc_dig_alg field properly based on the signing
9160 *) Allow PKCS#12 password to be set from the command line or the
9161 environment. Let 'ca' get its config file name from the environment
9162 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
9166 *) Allow certificate policies extension to use an IA5STRING for the
9167 organization field. This is contrary to the PKIX definition but
9168 VeriSign uses it and IE5 only recognises this form. Document 'x509'
9172 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
9173 without disallowing inline assembler and the like for non-pedantic builds.
9176 *) Support Borland C++ builder.
9177 [Janez Jere <jj@void.si>, modified by Ulf Möller]
9182 *) SHA-1 cleanups and performance enhancements.
9183 [Andy Polyakov <appro@fy.chalmers.se>]
9185 *) Sparc v8plus assembler for the bignum library.
9186 [Andy Polyakov <appro@fy.chalmers.se>]
9188 *) Accept any -xxx and +xxx compiler options in Configure.
9191 *) Update HPUX configuration.
9194 *) Add missing sk_<type>_unshift() function to safestack.h
9195 [Ralf S. Engelschall]
9197 *) New function SSL_CTX_use_certificate_chain_file that sets the
9198 "extra_cert"s in addition to the certificate. (This makes sense
9199 only for "PEM" format files, as chains as a whole are not
9203 *) Support verify_depth from the SSL API.
9204 x509_vfy.c had what can be considered an off-by-one-error:
9205 Its depth (which was not part of the external interface)
9206 was actually counting the number of certificates in a chain;
9207 now it really counts the depth.
9210 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
9211 instead of X509err, which often resulted in confusing error
9212 messages since the error codes are not globally unique
9213 (e.g. an alleged error in ssl3_accept when a certificate
9214 didn't match the private key).
9216 *) New function SSL_CTX_set_session_id_context that allows to set a default
9217 value (so that you don't need SSL_set_session_id_context for each
9218 connection using the SSL_CTX).
9221 *) OAEP decoding bug fix.
9224 *) Support INSTALL_PREFIX for package builders, as proposed by
9228 *) New Configure options "threads" and "no-threads". For systems
9229 where the proper compiler options are known (currently Solaris
9230 and Linux), "threads" is the default.
9233 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
9236 *) Install various scripts to $(OPENSSLDIR)/misc, not to
9237 $(INSTALLTOP)/bin -- they shouldn't clutter directories
9238 such as /usr/local/bin.
9241 *) "make linux-shared" to build shared libraries.
9242 [Niels Poppe <niels@netbox.org>]
9244 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
9247 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
9248 extension adding in x509 utility.
9251 *) Remove NOPROTO sections and error code comments.
9254 *) Partial rewrite of the DEF file generator to now parse the ANSI
9258 *) New Configure options --prefix=DIR and --openssldir=DIR.
9261 *) Complete rewrite of the error code script(s). It is all now handled
9262 by one script at the top level which handles error code gathering,
9263 header rewriting and C source file generation. It should be much better
9264 than the old method: it now uses a modified version of Ulf's parser to
9265 read the ANSI prototypes in all header files (thus the old K&R definitions
9266 aren't needed for error creation any more) and do a better job of
9267 translating function codes into names. The old 'ASN1 error code imbedded
9268 in a comment' is no longer necessary and it doesn't use .err files which
9269 have now been deleted. Also the error code call doesn't have to appear all
9270 on one line (which resulted in some large lines...).
9273 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
9276 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
9277 0 (which usually indicates a closed connection), but continue reading.
9280 *) Fix some race conditions.
9283 *) Add support for CRL distribution points extension. Add Certificate
9284 Policies and CRL distribution points documentation.
9287 *) Move the autogenerated header file parts to crypto/opensslconf.h.
9290 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
9291 8 of keying material. Merlin has also confirmed interop with this fix
9292 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
9293 [Merlin Hughes <merlin@baltimore.ie>]
9295 *) Fix lots of warnings.
9296 [Richard Levitte <levitte@stacken.kth.se>]
9298 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
9299 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9300 [Richard Levitte <levitte@stacken.kth.se>]
9302 *) Fix problems with sizeof(long) == 8.
9303 [Andy Polyakov <appro@fy.chalmers.se>]
9305 *) Change functions to ANSI C.
9308 *) Fix typos in error codes.
9309 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9311 *) Remove defunct assembler files from Configure.
9314 *) SPARC v8 assembler BIGNUM implementation.
9315 [Andy Polyakov <appro@fy.chalmers.se>]
9317 *) Support for Certificate Policies extension: both print and set.
9318 Various additions to support the r2i method this uses.
9321 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9322 return a const string when you are expecting an allocated buffer.
9325 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9326 types DirectoryString and DisplayText.
9329 *) Add code to allow r2i extensions to access the configuration database,
9330 add an LHASH database driver and add several ctx helper functions.
9333 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9334 fail when they extended the size of a BIGNUM.
9337 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9338 support typesafe stack.
9341 *) Fix typo in SSL_[gs]et_options().
9342 [Nils Frostberg <nils@medcom.se>]
9344 *) Delete various functions and files that belonged to the (now obsolete)
9345 old X509V3 handling code.
9348 *) New Configure option "rsaref".
9351 *) Don't auto-generate pem.h.
9354 *) Introduce type-safe ASN.1 SETs.
9357 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9358 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9360 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9361 that links with OpenSSL (well at least cause lots of warnings), but fear
9362 not: the conversion is trivial, and it eliminates loads of evil casts. A
9363 few STACKed things have been converted already. Feel free to convert more.
9364 In the fullness of time, I'll do away with the STACK type altogether.
9367 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9368 specified in <certfile> by updating the entry in the index.txt file.
9369 This way one no longer has to edit the index.txt file manually for
9370 revoking a certificate. The -revoke option does the gory details now.
9371 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9373 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9374 `-text' option at all and this way the `-noout -text' combination was
9375 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9376 [Ralf S. Engelschall]
9378 *) Make sure a corresponding plain text error message exists for the
9379 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9380 verify callback function determined that a certificate was revoked.
9381 [Ralf S. Engelschall]
9383 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9384 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9385 all available cipers including rc5, which was forgotten until now.
9386 In order to let the testing shell script know which algorithms
9387 are available, a new (up to now undocumented) command
9388 "openssl list-cipher-commands" is used.
9391 *) Bugfix: s_client occasionally would sleep in select() when
9392 it should have checked SSL_pending() first.
9395 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9396 the raw DSA values prior to ASN.1 encoding.
9399 *) Tweaks to Configure
9400 [Niels Poppe <niels@netbox.org>]
9402 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9406 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9409 *) New config option to avoid instructions that are illegal on the 80386.
9410 The default code is faster, but requires at least a 486.
9413 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9414 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9415 same as SSL2_VERSION anyway.
9418 *) New "-showcerts" option for s_client.
9421 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9422 application. Various cleanups and fixes.
9425 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9426 modify error routines to work internally. Add error codes and PBE init
9427 to library startup routines.
9430 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9431 packing functions to asn1 and evp. Changed function names and error
9432 codes along the way.
9435 *) PKCS12 integration: and so it begins... First of several patches to
9436 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9437 objects to objects.h
9440 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9441 and display support for Thawte strong extranet extension.
9444 *) Add LinuxPPC support.
9445 [Jeff Dubrule <igor@pobox.org>]
9447 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9448 bn_div_words in alpha.s.
9449 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9451 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9452 OAEP isn't supported when OpenSSL is built with RSAref.
9453 [Ulf Moeller <ulf@fitug.de>]
9455 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9456 so they no longer are missing under -DNOPROTO.
9457 [Soren S. Jorvang <soren@t.dk>]
9460 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9462 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9463 doesn't work when the session is reused. Coming soon!
9466 *) Fix a security hole, that allows sessions to be reused in the wrong
9467 context thus bypassing client cert protection! All software that uses
9468 client certs and session caches in multiple contexts NEEDS PATCHING to
9469 allow session reuse! A fuller solution is in the works.
9470 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9472 *) Some more source tree cleanups (removed obsolete files
9473 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9474 permission on "config" script to be executable) and a fix for the INSTALL
9476 [Ulf Moeller <ulf@fitug.de>]
9478 *) Remove some legacy and erroneous uses of malloc, free instead of
9480 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9482 *) Make rsa_oaep_test return non-zero on error.
9483 [Ulf Moeller <ulf@fitug.de>]
9485 *) Add support for native Solaris shared libraries. Configure
9486 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9487 if someone would make that last step automatic.
9488 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9490 *) ctx_size was not built with the right compiler during "make links". Fixed.
9493 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9494 except NULL ciphers". This means the default cipher list will no longer
9495 enable NULL ciphers. They need to be specifically enabled e.g. with
9496 the string "DEFAULT:eNULL".
9499 *) Fix to RSA private encryption routines: if p < q then it would
9500 occasionally produce an invalid result. This will only happen with
9501 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9504 *) Be less restrictive and allow also `perl util/perlpath.pl
9505 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9506 because this way one can also use an interpreter named `perl5' (which is
9507 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9508 installed as `perl').
9509 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9511 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9512 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9514 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9515 advapi32.lib to Win32 build and change the pem test comparision
9516 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9517 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9518 and crypto/des/ede_cbcm_enc.c.
9521 *) DES quad checksum was broken on big-endian architectures. Fixed.
9524 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9525 Win32 test batch file so it (might) work again. The Win32 test batch file
9526 is horrible: I feel ill....
9529 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9530 in e_os.h. Audit of header files to check ANSI and non ANSI
9531 sections: 10 functions were absent from non ANSI section and not exported
9532 from Windows DLLs. Fixed up libeay.num for new functions.
9535 *) Make `openssl version' output lines consistent.
9536 [Ralf S. Engelschall]
9538 *) Fix Win32 symbol export lists for BIO functions: Added
9539 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9540 to ms/libeay{16,32}.def.
9541 [Ralf S. Engelschall]
9543 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9544 fine under Unix and passes some trivial tests I've now added. But the
9545 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9546 added to make sure no one expects that this stuff really works in the
9547 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9548 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9550 [Ralf S. Engelschall]
9552 *) Fix the generation of two part addresses in perl.
9553 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9555 *) Add config entry for Linux on MIPS.
9556 [John Tobey <jtobey@channel1.com>]
9558 *) Make links whenever Configure is run, unless we are on Windoze.
9561 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9562 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9566 *) Add a useful kludge to allow package maintainers to specify compiler and
9567 other platforms details on the command line without having to patch the
9568 Configure script everytime: One now can use ``perl Configure
9569 <id>:<details>'', i.e. platform ids are allowed to have details appended
9570 to them (seperated by colons). This is treated as there would be a static
9571 pre-configured entry in Configure's %table under key <id> with value
9572 <details> and ``perl Configure <id>'' is called. So, when you want to
9573 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9574 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9575 now, which overrides the FreeBSD-elf entry on-the-fly.
9576 [Ralf S. Engelschall]
9578 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9581 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9582 on the `perl Configure ...' command line. This way one can compile
9583 OpenSSL libraries with Position Independent Code (PIC) which is needed
9584 for linking it into DSOs.
9585 [Ralf S. Engelschall]
9587 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9591 *) Cleaned up the LICENSE document: The official contact for any license
9592 questions now is the OpenSSL core team under openssl-core@openssl.org.
9593 And add a paragraph about the dual-license situation to make sure people
9594 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9595 to the OpenSSL toolkit.
9596 [Ralf S. Engelschall]
9598 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9599 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9600 Additonally cleaned up the `make links' target: Remove unnecessary
9601 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9602 to speed processing and no longer clutter the display with confusing
9603 stuff. Instead only the actually done links are displayed.
9604 [Ralf S. Engelschall]
9606 *) Permit null encryption ciphersuites, used for authentication only. It used
9607 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9608 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9612 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9613 signed attributes when verifying signatures (this would break them),
9614 the detached data encoding was wrong and public keys obtained using
9615 X509_get_pubkey() weren't freed.
9618 *) Add text documentation for the BUFFER functions. Also added a work around
9619 to a Win95 console bug. This was triggered by the password read stuff: the
9620 last character typed gets carried over to the next fread(). If you were
9621 generating a new cert request using 'req' for example then the last
9622 character of the passphrase would be CR which would then enter the first
9626 *) Added the new `Includes OpenSSL Cryptography Software' button as
9627 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9628 button and can be used by applications based on OpenSSL to show the
9629 relationship to the OpenSSL project.
9630 [Ralf S. Engelschall]
9632 *) Remove confusing variables in function signatures in files
9633 ssl/ssl_lib.c and ssl/ssl.h.
9634 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9636 *) Don't install bss_file.c under PREFIX/include/
9637 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9639 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9640 functions that return function pointers and has support for NT specific
9641 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9642 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9643 unsigned to signed types: this was killing the Win32 compile.
9646 *) Add new certificate file to stack functions,
9647 SSL_add_dir_cert_subjects_to_stack() and
9648 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9649 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9650 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9651 This means that Apache-SSL and similar packages don't have to mess around
9652 to add as many CAs as they want to the preferred list.
9655 *) Experiment with doxygen documentation. Currently only partially applied to
9657 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9658 openssl.doxy as the configuration file.
9661 *) Get rid of remaining C++-style comments which strict C compilers hate.
9662 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9664 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9665 compiled in by default: it has problems with large keys.
9668 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9669 DH private keys and/or callback functions which directly correspond to
9670 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9671 is needed for applications which have to configure certificates on a
9672 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9674 For the RSA certificate situation is makes no difference, but
9675 for the DSA certificate situation this fixes the "no shared cipher"
9676 problem where the OpenSSL cipher selection procedure failed because the
9677 temporary keys were not overtaken from the context and the API provided
9678 no way to reconfigure them.
9679 The new functions now let applications reconfigure the stuff and they
9680 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9681 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9682 non-public-API function ssl_cert_instantiate() is used as a helper
9683 function and also to reduce code redundancy inside ssl_rsa.c.
9684 [Ralf S. Engelschall]
9686 *) Move s_server -dcert and -dkey options out of the undocumented feature
9687 area because they are useful for the DSA situation and should be
9688 recognized by the users.
9689 [Ralf S. Engelschall]
9691 *) Fix the cipher decision scheme for export ciphers: the export bits are
9692 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9693 SSL_EXP_MASK. So, the original variable has to be used instead of the
9694 already masked variable.
9695 [Richard Levitte <levitte@stacken.kth.se>]
9697 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9698 [Richard Levitte <levitte@stacken.kth.se>]
9700 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9701 from `int' to `unsigned int' because it's a length and initialized by
9702 EVP_DigestFinal() which expects an `unsigned int *'.
9703 [Richard Levitte <levitte@stacken.kth.se>]
9705 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9706 script. Instead use the usual Shell->Perl transition trick.
9707 [Ralf S. Engelschall]
9709 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9710 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9711 -noout -modulus' as it's already the case for `openssl rsa -noout
9712 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9713 currently the public key is printed (a decision which was already done by
9714 `openssl dsa -modulus' in the past) which serves a similar purpose.
9715 Additionally the NO_RSA no longer completely removes the whole -modulus
9716 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9718 [Ralf S. Engelschall]
9720 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9721 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9722 [Arne Ansper <arne@ats.cyber.ee>]
9724 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9725 to be added. Now both 'req' and 'ca' can use new objects defined in the
9729 *) Add cool BIO that does syslog (or event log on NT).
9730 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9732 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9733 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9734 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9735 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9738 *) Add preliminary config info for new extension code.
9741 *) Make RSA_NO_PADDING really use no padding.
9742 [Ulf Moeller <ulf@fitug.de>]
9744 *) Generate errors when private/public key check is done.
9747 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9748 for some CRL extensions and new objects added.
9751 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9752 key usage extension and fuller support for authority key id.
9755 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9756 padding method for RSA, which is recommended for new applications in PKCS
9757 #1 v2.0 (RFC 2437, October 1998).
9758 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9759 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9760 against Bleichbacher's attack on RSA.
9761 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9764 *) Updates to the new SSL compression code
9765 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9767 *) Fix so that the version number in the master secret, when passed
9768 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9769 (because the server will not accept higher), that the version number
9770 is 0x03,0x01, not 0x03,0x00
9771 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9773 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9774 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9775 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9778 *) Support for RAW extensions where an arbitrary extension can be
9779 created by including its DER encoding. See apps/openssl.cnf for
9783 *) Make sure latest Perl versions don't interpret some generated C array
9784 code as Perl array code in the crypto/err/err_genc.pl script.
9785 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9787 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9788 not many people have the assembler. Various Win32 compilation fixes and
9789 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9793 *) Modify configure script 'Configure' to automatically create crypto/date.h
9794 file under Win32 and also build pem.h from pem.org. New script
9795 util/mkfiles.pl to create the MINFO file on environments that can't do a
9796 'make files': perl util/mkfiles.pl >MINFO should work.
9799 *) Major rework of DES function declarations, in the pursuit of correctness
9800 and purity. As a result, many evil casts evaporated, and some weirdness,
9801 too. You may find this causes warnings in your code. Zapping your evil
9802 casts will probably fix them. Mostly.
9805 *) Fix for a typo in asn1.h. Bug fix to object creation script
9806 obj_dat.pl. It considered a zero in an object definition to mean
9807 "end of object": none of the objects in objects.h have any zeros
9808 so it wasn't spotted.
9809 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9811 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9812 Masking (CBCM). In the absence of test vectors, the best I have been able
9813 to do is check that the decrypt undoes the encrypt, so far. Send me test
9814 vectors if you have them.
9817 *) Correct calculation of key length for export ciphers (too much space was
9818 allocated for null ciphers). This has not been tested!
9821 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9822 message is now correct (it understands "crypto" and "ssl" on its
9823 command line). There is also now an "update" option. This will update
9824 the util/ssleay.num and util/libeay.num files with any new functions.
9826 perl util/mkdef.pl crypto ssl update
9827 it will update them.
9830 *) Overhauled the Perl interface (perl/*):
9831 - ported BN stuff to OpenSSL's different BN library
9832 - made the perl/ source tree CVS-aware
9833 - renamed the package from SSLeay to OpenSSL (the files still contain
9834 their history because I've copied them in the repository)
9835 - removed obsolete files (the test scripts will be replaced
9836 by better Test::Harness variants in the future)
9837 [Ralf S. Engelschall]
9839 *) First cut for a very conservative source tree cleanup:
9840 1. merge various obsolete readme texts into doc/ssleay.txt
9841 where we collect the old documents and readme texts.
9842 2. remove the first part of files where I'm already sure that we no
9843 longer need them because of three reasons: either they are just temporary
9844 files which were left by Eric or they are preserved original files where
9845 I've verified that the diff is also available in the CVS via "cvs diff
9846 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9847 the crypto/md/ stuff).
9848 [Ralf S. Engelschall]
9850 *) More extension code. Incomplete support for subject and issuer alt
9851 name, issuer and authority key id. Change the i2v function parameters
9852 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9853 what that's for :-) Fix to ASN1 macro which messed up
9854 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9857 *) Preliminary support for ENUMERATED type. This is largely copied from the
9861 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9862 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9864 *) Make sure `make rehash' target really finds the `openssl' program.
9865 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9867 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9868 like to hear about it if this slows down other processors.
9871 *) Add CygWin32 platform information to Configure script.
9872 [Alan Batie <batie@aahz.jf.intel.com>]
9874 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
9875 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
9877 *) New program nseq to manipulate netscape certificate sequences
9880 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
9884 *) Fixes to BN code. Previously the default was to define BN_RECURSION
9885 but the BN code had some problems that would cause failures when
9886 doing certificate verification and some other functions.
9887 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9889 *) Add ASN1 and PEM code to support netscape certificate sequences.
9892 *) Add ASN1 and PEM code to support netscape certificate sequences.
9895 *) Add several PKIX and private extended key usage OIDs.
9898 *) Modify the 'ca' program to handle the new extension code. Modify
9899 openssl.cnf for new extension format, add comments.
9902 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
9903 and add a sample to openssl.cnf so req -x509 now adds appropriate
9907 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
9908 error code, add initial support to X509_print() and x509 application.
9911 *) Takes a deep breath and start addding X509 V3 extension support code. Add
9912 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
9913 stuff is currently isolated and isn't even compiled yet.
9916 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
9917 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
9918 Removed the versions check from X509 routines when loading extensions:
9919 this allows certain broken certificates that don't set the version
9920 properly to be processed.
9923 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
9924 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
9925 can still be regenerated with "make depend".
9928 *) Spelling mistake in C version of CAST-128.
9929 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
9931 *) Changes to the error generation code. The perl script err-code.pl
9932 now reads in the old error codes and retains the old numbers, only
9933 adding new ones if necessary. It also only changes the .err files if new
9934 codes are added. The makefiles have been modified to only insert errors
9935 when needed (to avoid needlessly modifying header files). This is done
9936 by only inserting errors if the .err file is newer than the auto generated
9937 C file. To rebuild all the error codes from scratch (the old behaviour)
9938 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
9939 or delete all the .err files.
9942 *) CAST-128 was incorrectly implemented for short keys. The C version has
9943 been fixed, but is untested. The assembler versions are also fixed, but
9944 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
9945 to regenerate it if needed.
9946 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
9947 Hagino <itojun@kame.net>]
9949 *) File was opened incorrectly in randfile.c.
9950 [Ulf Möller <ulf@fitug.de>]
9952 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
9953 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
9954 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
9955 al: it's just almost always a UTCTime. Note this patch adds new error
9956 codes so do a "make errors" if there are problems.
9959 *) Correct Linux 1 recognition in config.
9960 [Ulf Möller <ulf@fitug.de>]
9962 *) Remove pointless MD5 hash when using DSA keys in ca.
9963 [Anonymous <nobody@replay.com>]
9965 *) Generate an error if given an empty string as a cert directory. Also
9966 generate an error if handed NULL (previously returned 0 to indicate an
9967 error, but didn't set one).
9968 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
9970 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
9973 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
9974 parameters. This was causing a warning which killed off the Win32 compile.
9977 *) Remove C++ style comments from crypto/bn/bn_local.h.
9978 [Neil Costigan <neil.costigan@celocom.com>]
9980 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
9981 based on a text string, looking up short and long names and finally
9982 "dot" format. The "dot" format stuff didn't work. Added new function
9983 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
9984 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
9985 OID is not part of the table.
9988 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
9989 X509_LOOKUP_by_alias().
9992 *) Sort openssl functions by name.
9995 *) Get the gendsa program working (hopefully) and add it to app list. Remove
9996 encryption from sample DSA keys (in case anyone is interested the password
10000 *) Make _all_ *_free functions accept a NULL pointer.
10001 [Frans Heymans <fheymans@isaserver.be>]
10003 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10005 [Anonymous <nobody@replay.com>]
10007 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10008 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10010 *) Don't blow it for numeric -newkey arguments to apps/req.
10011 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10013 *) Temp key "for export" tests were wrong in s3_srvr.c.
10014 [Anonymous <nobody@replay.com>]
10016 *) Add prototype for temp key callback functions
10017 SSL_CTX_set_tmp_{rsa,dh}_callback().
10020 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10021 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10024 *) X509_name_add_entry() freed the wrong thing after an error.
10025 [Arne Ansper <arne@ats.cyber.ee>]
10027 *) rsa_eay.c would attempt to free a NULL context.
10028 [Arne Ansper <arne@ats.cyber.ee>]
10030 *) BIO_s_socket() had a broken should_retry() on Windoze.
10031 [Arne Ansper <arne@ats.cyber.ee>]
10033 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10034 [Arne Ansper <arne@ats.cyber.ee>]
10036 *) Make sure the already existing X509_STORE->depth variable is initialized
10037 in X509_STORE_new(), but document the fact that this variable is still
10038 unused in the certificate verification process.
10039 [Ralf S. Engelschall]
10041 *) Fix the various library and apps files to free up pkeys obtained from
10042 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10045 *) Fix reference counting in X509_PUBKEY_get(). This makes
10046 demos/maurice/example2.c work, amongst others, probably.
10047 [Steve Henson and Ben Laurie]
10049 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10050 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10051 are no longer created. This way we have a single and consistent command
10052 line interface `openssl <command>', similar to `cvs <command>'.
10053 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10055 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10056 BIT STRING wrapper always have zero unused bits.
10059 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10062 *) Make the top-level INSTALL documentation easier to understand.
10065 *) Makefiles updated to exit if an error occurs in a sub-directory
10066 make (including if user presses ^C) [Paul Sutton]
10068 *) Make Montgomery context stuff explicit in RSA data structure.
10071 *) Fix build order of pem and err to allow for generated pem.h.
10074 *) Fix renumbering bug in X509_NAME_delete_entry().
10077 *) Enhanced the err-ins.pl script so it makes the error library number
10078 global and can add a library name. This is needed for external ASN1 and
10079 other error libraries.
10082 *) Fixed sk_insert which never worked properly.
10085 *) Fix ASN1 macros so they can handle indefinite length construted
10086 EXPLICIT tags. Some non standard certificates use these: they can now
10090 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10091 into a single doc/ssleay.txt bundle. This way the information is still
10092 preserved but no longer messes up this directory. Now it's new room for
10093 the new set of documenation files.
10094 [Ralf S. Engelschall]
10096 *) SETs were incorrectly DER encoded. This was a major pain, because they
10097 shared code with SEQUENCEs, which aren't coded the same. This means that
10098 almost everything to do with SETs or SEQUENCEs has either changed name or
10099 number of arguments.
10100 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10102 *) Fix test data to work with the above.
10105 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10106 was already fixed by Eric for 0.9.1 it seems.
10107 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10109 *) Autodetect FreeBSD3.
10112 *) Fix various bugs in Configure. This affects the following platforms:
10116 unixware-2.0-pentium
10120 *) Eliminate generated files from CVS. Reorder tests to regenerate files
10121 before they are needed.
10124 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
10128 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
10130 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
10131 changed SSLeay to OpenSSL in version strings.
10132 [Ralf S. Engelschall]
10134 *) Some fixups to the top-level documents.
10137 *) Fixed the nasty bug where rsaref.h was not found under compile-time
10138 because the symlink to include/ was missing.
10139 [Ralf S. Engelschall]
10141 *) Incorporated the popular no-RSA/DSA-only patches
10142 which allow to compile a RSA-free SSLeay.
10143 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
10145 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
10146 when "ssleay" is still not found.
10147 [Ralf S. Engelschall]
10149 *) Added more platforms to Configure: Cray T3E, HPUX 11,
10150 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
10152 *) Updated the README file.
10153 [Ralf S. Engelschall]
10155 *) Added various .cvsignore files in the CVS repository subdirs
10156 to make a "cvs update" really silent.
10157 [Ralf S. Engelschall]
10159 *) Recompiled the error-definition header files and added
10160 missing symbols to the Win32 linker tables.
10161 [Ralf S. Engelschall]
10163 *) Cleaned up the top-level documents;
10164 o new files: CHANGES and LICENSE
10165 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
10166 o merged COPYRIGHT into LICENSE
10167 o removed obsolete TODO file
10168 o renamed MICROSOFT to INSTALL.W32
10169 [Ralf S. Engelschall]
10171 *) Removed dummy files from the 0.9.1b source tree:
10172 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
10173 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
10174 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
10175 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
10176 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
10177 [Ralf S. Engelschall]
10179 *) Added various platform portability fixes.
10182 *) The Genesis of the OpenSSL rpject:
10183 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
10184 Young and Tim J. Hudson created while they were working for C2Net until
10186 [The OpenSSL Project]
10189 Changes between 0.9.0b and 0.9.1b [not released]
10191 *) Updated a few CA certificates under certs/
10194 *) Changed some BIGNUM api stuff.
10197 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
10198 DGUX x86, Linux Alpha, etc.
10201 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
10202 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
10206 *) Add -strparse option to asn1pars program which parses nested
10208 [Dr Stephen Henson <shenson@bigfoot.com>]
10210 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
10213 *) DSA fix for "ca" program.
10216 *) Added "-genkey" option to "dsaparam" program.
10219 *) Added RIPE MD160 (rmd160) message digest.
10222 *) Added -a (all) option to "ssleay version" command.
10225 *) Added PLATFORM define which is the id given to Configure.
10228 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
10231 *) Extended the ASN.1 parser routines.
10234 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
10237 *) Added a BN_CTX to the BN library.
10240 *) Fixed the weak key values in DES library
10243 *) Changed API in EVP library for cipher aliases.
10246 *) Added support for RC2/64bit cipher.
10249 *) Converted the lhash library to the crypto/mem.c functions.
10252 *) Added more recognized ASN.1 object ids.
10255 *) Added more RSA padding checks for SSL/TLS.
10258 *) Added BIO proxy/filter functionality.
10261 *) Added extra_certs to SSL_CTX which can be used
10262 send extra CA certificates to the client in the CA cert chain sending
10263 process. It can be configured with SSL_CTX_add_extra_chain_cert().
10266 *) Now Fortezza is denied in the authentication phase because
10267 this is key exchange mechanism is not supported by SSLeay at all.
10270 *) Additional PKCS1 checks.
10273 *) Support the string "TLSv1" for all TLS v1 ciphers.
10276 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
10277 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
10280 *) Fixed a few memory leaks.
10283 *) Fixed various code and comment typos.
10286 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
10287 bytes sent in the client random.
10288 [Edward Bishop <ebishop@spyglass.com>]
projects / openssl.git / blob