5 Changes between 1.0.1 and 1.1.0 [xx XXX xxxx]
7 *) Update and tidy signature algorithm extension processing. Work out
8 shared signature algorithms based on preferences and peer algorithms
9 and print them out in s_client and s_server. Abort handshake if no
10 shared signature algorithms.
13 *) Add new functions to allow customised supported signature algorithms
14 for SSL and SSL_CTX structures. Add options to s_client and s_server
18 *) New function SSL_certs_clear() to delete all references to certificates
19 from an SSL structure. Before this once a certificate had been added
20 it couldn't be removed.
23 *) Initial SSL tracing code. This parses out SSL/TLS records using the
24 message callback and prints the results. Needs compile time option
25 "enable-ssl-trace". New options to s_client and s_server to enable
29 *) New functions to retrieve certificate signature and signature
33 *) Print out deprecated issuer and subject unique ID fields in
38 [Emilia Kasper, Adam Langley, Ben Laurie (Google)]
40 *) Support for automatic EC temporary key parameter selection. If enabled
41 the most preferred EC parameters are automatically used instead of
42 hardcoded fixed parameters. Now a server just has to call:
43 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
44 support ECDH and use the most appropriate parameters.
47 *) Enhance and tidy EC curve and point format TLS extension code. Use
48 static structures instead of allocation if default values are used.
49 New ctrls to set curves we wish to support and to retrieve shared curves.
50 Print out shared curves in s_server. New options to s_server and s_client
51 to set list of supported curves.
54 *) New ctrls to retrieve supported signature algorithms and
55 supported curve values as an array of NIDs. Extend openssl utility
56 to print out received values.
59 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
60 between NIDs and the more common NIST names such as "P-256". Enhance
61 ecparam utility and ECC method to recognise the NIST names for curves.
64 *) Enhance SSL/TLS certificate chain handling to support different
65 chains for each certificate instead of one chain in the parent SSL_CTX.
68 *) Support for fixed DH ciphersuite client authentication: where both
69 server and client use DH certificates with common parameters.
72 *) Support for fixed DH ciphersuites: those requiring DH server
76 *) Transparently support X9.42 DH parameters when calling
77 PEM_read_bio_DHparameters. This means existing applications can handle
78 the new parameter format automatically.
81 *) Initial experimental support for X9.42 DH parameter format: mainly
82 to support use of 'q' parameter for RFC5114 parameters.
85 *) Add DH parameters from RFC5114 including test data to dhtest.
88 *) Update fips_test_suite to support multiple command line options. New
89 test to induce all self test errors in sequence and check expected
93 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
94 sign or verify all in one operation.
97 *) Add fips_algvs: a multicall fips utility incorporaing all the algorithm
98 test programs and fips_test_suite. Includes functionality to parse
99 the minimal script output of fipsalgest.pl directly.
102 *) Add authorisation parameter to FIPS_module_mode_set().
105 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
108 *) Use separate DRBG fields for internal and external flags. New function
109 FIPS_drbg_health_check() to perform on demand health checking. Add
110 generation tests to fips_test_suite with reduced health check interval to
111 demonstrate periodic health checking. Add "nodh" option to
112 fips_test_suite to skip very slow DH test.
115 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
119 *) More extensive health check for DRBG checking many more failure modes.
120 New function FIPS_selftest_drbg_all() to handle every possible DRBG
121 combination: call this in fips_test_suite.
124 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
125 and POST to handle Dual EC cases.
128 *) Add support for canonical generation of DSA parameter 'g'. See
131 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
132 POST to handle HMAC cases.
135 *) Add functions FIPS_module_version() and FIPS_module_version_text()
136 to return numberical and string versions of the FIPS module number.
139 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
140 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implmeneted
141 outside the validated module in the FIPS capable OpenSSL.
144 *) Minor change to DRBG entropy callback semantics. In some cases
145 there is no mutiple of the block length between min_len and
146 max_len. Allow the callback to return more than max_len bytes
147 of entropy but discard any extra: it is the callback's responsibility
148 to ensure that the extra data discarded does not impact the
149 requested amount of entropy.
152 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
153 information in FIPS186-3, SP800-57 and SP800-131A.
156 *) CCM support via EVP. Interface is very similar to GCM case except we
157 must supply all data in one chunk (i.e. no update, final) and the
158 message length must be supplied if AAD is used. Add algorithm test
162 *) Initial version of POST overhaul. Add POST callback to allow the status
163 of POST to be monitored and/or failures induced. Modify fips_test_suite
164 to use callback. Always run all selftests even if one fails.
167 *) XTS support including algorithm test driver in the fips_gcmtest program.
168 Note: this does increase the maximum key length from 32 to 64 bytes but
169 there should be no binary compatibility issues as existing applications
170 will never use XTS mode.
173 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
174 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
175 performs algorithm blocking for unapproved PRNG types. Also do not
176 set PRNG type in FIPS_mode_set(): leave this to the application.
177 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
178 the standard OpenSSL PRNG: set additional data to a date time vector.
181 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
182 This shouldn't present any incompatibility problems because applications
183 shouldn't be using these directly and any that are will need to rethink
184 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
187 *) Extensive self tests and health checking required by SP800-90 DRBG.
188 Remove strength parameter from FIPS_drbg_instantiate and always
189 instantiate at maximum supported strength.
192 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
195 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
198 *) New function DH_compute_key_padded() to compute a DH key and pad with
199 leading zeroes if needed: this complies with SP800-56A et al.
202 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
203 anything, incomplete, subject to change and largely untested at present.
206 *) Modify fipscanisteronly build option to only build the necessary object
207 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
210 *) Add experimental option FIPSSYMS to give all symbols in
211 fipscanister.o and FIPS or fips prefix. This will avoid
212 conflicts with future versions of OpenSSL. Add perl script
213 util/fipsas.pl to preprocess assembly language source files
214 and rename any affected symbols.
217 *) Add selftest checks and algorithm block of non-fips algorithms in
218 FIPS mode. Remove DES2 from selftests.
221 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
222 return internal method without any ENGINE dependencies. Add new
223 tiny fips sign and verify functions.
226 *) New build option no-ec2m to disable characteristic 2 code.
229 *) New build option "fipscanisteronly". This only builds fipscanister.o
230 and (currently) associated fips utilities. Uses the file Makefile.fips
231 instead of Makefile.org as the prototype.
234 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
235 Update fips_gcmtest to use IV generator.
238 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
239 setting output buffer to NULL. The *Final function must be
240 called although it will not retrieve any additional data. The tag
241 can be set or retrieved with a ctrl. The IV length is by default 12
242 bytes (96 bits) but can be set to an alternative value. If the IV
243 length exceeds the maximum IV length (currently 16 bytes) it cannot be
247 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
248 underlying do_cipher function handles all cipher semantics itself
249 including padding and finalisation. This is useful if (for example)
250 an ENGINE cipher handles block padding itself. The behaviour of
251 do_cipher is subtly changed if this flag is set: the return value
252 is the number of characters written to the output buffer (zero is
253 no longer an error code) or a negative error code. Also if the
254 input buffer is NULL and length 0 finalisation should be performed.
257 *) If a candidate issuer certificate is already part of the constructed
258 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
261 *) Improve forward-security support: add functions
263 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
264 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
266 for use by SSL/TLS servers; the callback function will be called whenever a
267 new session is created, and gets to decide whether the session may be
268 cached to make it resumable (return 0) or not (return 1). (As by the
269 SSL/TLS protocol specifications, the session_id sent by the server will be
270 empty to indicate that the session is not resumable; also, the server will
271 not generate RFC 4507 (RFC 5077) session tickets.)
273 A simple reasonable callback implementation is to return is_forward_secure.
274 This parameter will be set to 1 or 0 depending on the ciphersuite selected
275 by the SSL/TLS server library, indicating whether it can provide forward
277 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
279 *) New function OPENSSL_gmtime_diff to find the difference in days
280 and seconds between two tm structures. This will be used to provide
281 additional functionality for ASN1_TIME.
284 *) Add -trusted_first option which attempts to find certificates in the
285 trusted store even if an untrusted chain is also supplied.
288 *) Initial experimental support for explicitly trusted non-root CAs.
289 OpenSSL still tries to build a complete chain to a root but if an
290 intermediate CA has a trust setting included that is used. The first
291 setting is used: whether to trust or reject.
294 *) New -verify_name option in command line utilities to set verification
298 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
299 Add CMAC pkey methods.
302 *) Experimental regnegotiation in s_server -www mode. If the client
303 browses /reneg connection is renegotiated. If /renegcert it is
304 renegotiated requesting a certificate.
307 *) Add an "external" session cache for debugging purposes to s_server. This
308 should help trace issues which normally are only apparent in deployed
309 multi-process servers.
312 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
313 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
314 BIO_set_cipher() and some obscure PEM functions were changed so they
315 can now return an error. The RAND changes required a change to the
316 RAND_METHOD structure.
319 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
320 a gcc attribute to warn if the result of a function is ignored. This
321 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
322 whose return value is often ignored.
325 Changes between 1.0.1c and 1.0.1d [xx XXX xxxx]
327 *) Fix possible deadlock when decoding public keys.
330 *) Don't use TLS 1.0 record version number in initial client hello
334 Changes between 1.0.1b and 1.0.1c [10 May 2012]
336 *) Sanity check record length before skipping explicit IV in TLS
337 1.2, 1.1 and DTLS to avoid DoS attack.
339 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
340 fuzzing as a service testing platform.
344 *) Initialise tkeylen properly when encrypting CMS messages.
345 Thanks to Solar Designer of Openwall for reporting this issue.
348 *) In FIPS mode don't try to use composite ciphers as they are not
352 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
354 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
355 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
356 mean any application compiled against OpenSSL 1.0.0 headers setting
357 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
358 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
359 0x10000000L Any application which was previously compiled against
360 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
361 will need to be recompiled as a result. Letting be results in
362 inability to disable specifically TLS 1.1 and in client context,
363 in unlike event, limit maximum offered version to TLS 1.0 [see below].
366 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
367 disable just protocol X, but all protocols above X *if* there are
368 protocols *below* X still enabled. In more practical terms it means
369 that if application wants to disable TLS1.0 in favor of TLS1.1 and
370 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
371 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
375 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
377 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
378 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
379 in CRYPTO_realloc_clean.
381 Thanks to Tavis Ormandy, Google Security Team, for discovering this
382 issue and to Adam Langley <agl@chromium.org> for fixing it.
384 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
386 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
389 *) Workarounds for some broken servers that "hang" if a client hello
390 record length exceeds 255 bytes:
392 1. Do not use record version number > TLS 1.0 in initial client
393 hello: some (but not all) hanging servers will now work.
394 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
395 the number of ciphers sent in the client hello. This should be
396 set to an even number, such as 50, for example by passing:
397 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
398 Most broken servers should now work.
399 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
400 TLS 1.2 client support entirely.
403 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
406 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
408 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
409 STRING form instead of a DigestInfo.
412 *) The format used for MDC2 RSA signatures is inconsistent between EVP
413 and the RSA_sign/RSA_verify functions. This was made more apparent when
414 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
415 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
416 the correct format in RSA_verify so both forms transparently work.
419 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
420 STRING form instead of a DigestInfo.
423 *) Some servers which support TLS 1.0 can choke if we initially indicate
424 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
425 encrypted premaster secret. As a workaround use the maximum pemitted
426 client version in client hello, this should keep such servers happy
427 and still work with previous versions of OpenSSL.
430 *) Add support for TLS/DTLS heartbeats.
431 [Robin Seggelmann <seggelmann@fh-muenster.de>]
433 *) Add support for SCTP.
434 [Robin Seggelmann <seggelmann@fh-muenster.de>]
436 *) Improved PRNG seeding for VOS.
437 [Paul Green <Paul.Green@stratus.com>]
439 *) Extensive assembler packs updates, most notably:
441 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
442 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
443 - x86_64: bit-sliced AES implementation;
444 - ARM: NEON support, contemporary platforms optimizations;
445 - s390x: z196 support;
446 - *: GHASH and GF(2^m) multiplication implementations;
450 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
451 (removal of unnecessary code)
452 [Peter Sylvester <peter.sylvester@edelweb.fr>]
454 *) Add TLS key material exporter from RFC 5705.
457 *) Add DTLS-SRTP negotiation from RFC 5764.
460 *) Add Next Protocol Negotiation,
461 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
462 disabled with a no-npn flag to config or Configure. Code donated
464 [Adam Langley <agl@google.com> and Ben Laurie]
466 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
467 NIST-P256, NIST-P521, with constant-time single point multiplication on
468 typical inputs. Compiler support for the nonstandard type __uint128_t is
469 required to use this (present in gcc 4.4 and later, for 64-bit builds).
470 Code made available under Apache License version 2.0.
472 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
473 line to include this in your build of OpenSSL, and run "make depend" (or
474 "make update"). This enables the following EC_METHODs:
476 EC_GFp_nistp224_method()
477 EC_GFp_nistp256_method()
478 EC_GFp_nistp521_method()
480 EC_GROUP_new_by_curve_name() will automatically use these (while
481 EC_GROUP_new_curve_GFp() currently prefers the more flexible
483 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
485 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
486 all platforms. Move ssize_t definition from e_os.h to the public
487 header file e_os2.h as it now appears in public header file cms.h
490 *) New -sigopt option to the ca, req and x509 utilities. Additional
491 signature parameters can be passed using this option and in
495 *) Add RSA PSS signing function. This will generate and set the
496 appropriate AlgorithmIdentifiers for PSS based on those in the
497 corresponding EVP_MD_CTX structure. No application support yet.
500 *) Support for companion algorithm specific ASN1 signing routines.
501 New function ASN1_item_sign_ctx() signs a pre-initialised
502 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
503 the appropriate parameters.
506 *) Add new algorithm specific ASN1 verification initialisation function
507 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
508 handling will be the same no matter what EVP_PKEY_METHOD is used.
509 Add a PSS handler to support verification of PSS signatures: checked
510 against a number of sample certificates.
513 *) Add signature printing for PSS. Add PSS OIDs.
514 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
516 *) Add algorithm specific signature printing. An individual ASN1 method
517 can now print out signatures instead of the standard hex dump.
519 More complex signatures (e.g. PSS) can print out more meaningful
520 information. Include DSA version that prints out the signature
524 *) Password based recipient info support for CMS library: implementing
528 *) Split password based encryption into PBES2 and PBKDF2 functions. This
529 neatly separates the code into cipher and PBE sections and is required
530 for some algorithms that split PBES2 into separate pieces (such as
534 *) Session-handling fixes:
535 - Fix handling of connections that are resuming with a session ID,
536 but also support Session Tickets.
537 - Fix a bug that suppressed issuing of a new ticket if the client
538 presented a ticket with an expired session.
539 - Try to set the ticket lifetime hint to something reasonable.
540 - Make tickets shorter by excluding irrelevant information.
541 - On the client side, don't ignore renewed tickets.
542 [Adam Langley, Bodo Moeller (Google)]
544 *) Fix PSK session representation.
547 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
549 This work was sponsored by Intel.
552 *) Add GCM support to TLS library. Some custom code is needed to split
553 the IV between the fixed (from PRF) and explicit (from TLS record)
554 portions. This adds all GCM ciphersuites supported by RFC5288 and
555 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
556 add a special AESGCM string for GCM only.
559 *) Expand range of ctrls for AES GCM. Permit setting invocation
560 field on decrypt and retrieval of invocation field only on encrypt.
563 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
564 As required by RFC5289 these ciphersuites cannot be used if for
565 versions of TLS earlier than 1.2.
568 *) For FIPS capable OpenSSL interpret a NULL default public key method
569 as unset and return the appopriate default but do *not* set the default.
570 This means we can return the appopriate method in applications that
571 swicth between FIPS and non-FIPS modes.
574 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
575 ENGINE is used then we cannot handle that in the FIPS module so we
576 keep original code iff non-FIPS operations are allowed.
579 *) Add -attime option to openssl utilities.
580 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
582 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
585 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
586 FIPS EC methods unconditionally for now.
589 *) New build option no-ec2m to disable characteristic 2 code.
592 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
593 all cases can be covered as some introduce binary incompatibilities.
596 *) Redirect RSA operations to FIPS module including keygen,
597 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
600 *) Add similar low level API blocking to ciphers.
603 *) Low level digest APIs are not approved in FIPS mode: any attempt
604 to use these will cause a fatal error. Applications that *really* want
605 to use them can use the private_* version instead.
608 *) Redirect cipher operations to FIPS module for FIPS builds.
611 *) Redirect digest operations to FIPS module for FIPS builds.
614 *) Update build system to add "fips" flag which will link in fipscanister.o
615 for static and shared library builds embedding a signature if needed.
618 *) Output TLS supported curves in preference order instead of numerical
619 order. This is currently hardcoded for the highest order curves first.
620 This should be configurable so applications can judge speed vs strength.
623 *) Add TLS v1.2 server support for client authentication.
626 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
630 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
631 FIPS modules versions.
634 *) Add TLS v1.2 client side support for client authentication. Keep cache
635 of handshake records longer as we don't know the hash algorithm to use
636 until after the certificate request message is received.
639 *) Initial TLS v1.2 client support. Add a default signature algorithms
640 extension including all the algorithms we support. Parse new signature
641 format in client key exchange. Relax some ECC signing restrictions for
642 TLS v1.2 as indicated in RFC5246.
645 *) Add server support for TLS v1.2 signature algorithms extension. Switch
646 to new signature format when needed using client digest preference.
647 All server ciphersuites should now work correctly in TLS v1.2. No client
648 support yet and no support for client certificates.
651 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
652 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
653 ciphersuites. At present only RSA key exchange ciphersuites work with
654 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
655 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
656 and version checking.
659 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
660 with this defined it will not be affected by any changes to ssl internal
661 structures. Add several utility functions to allow openssl application
662 to work with OPENSSL_NO_SSL_INTERN defined.
666 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
668 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
671 *) Permit abbreviated handshakes when renegotiating using the function
672 SSL_renegotiate_abbreviated().
673 [Robin Seggelmann <seggelmann@fh-muenster.de>]
675 *) Add call to ENGINE_register_all_complete() to
676 ENGINE_load_builtin_engines(), so some implementations get used
677 automatically instead of needing explicit application support.
680 *) Add support for TLS key exporter as described in RFC5705.
681 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
683 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
684 a few changes are required:
686 Add SSL_OP_NO_TLSv1_1 flag.
688 Update version checking logic to handle version 1.1.
689 Add explicit IV handling (ported from DTLS code).
690 Add command line options to s_client/s_server.
693 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
695 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
696 in CMS and PKCS7 code. When RSA decryption fails use a random key for
697 content decryption and always return the same error. Note: this attack
698 needs on average 2^20 messages so it only affects automated senders. The
699 old behaviour can be reenabled in the CMS code by setting the
700 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
701 an MMA defence is not necessary.
702 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
703 this issue. (CVE-2012-0884)
706 *) Fix CVE-2011-4619: make sure we really are receiving a
707 client hello before rejecting multiple SGC restarts. Thanks to
708 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
711 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
713 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
714 Thanks to Antonio Martin, Enterprise Secure Access Research and
715 Development, Cisco Systems, Inc. for discovering this bug and
716 preparing a fix. (CVE-2012-0050)
719 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
721 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
722 of the Vaudenay padding oracle attack on CBC mode encryption
723 which enables an efficient plaintext recovery attack against
724 the OpenSSL implementation of DTLS. Their attack exploits timing
725 differences arising during decryption processing. A research
726 paper describing this attack can be found at:
727 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
728 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
729 Security Group at Royal Holloway, University of London
730 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
731 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
732 for preparing the fix. (CVE-2011-4108)
733 [Robin Seggelmann, Michael Tuexen]
735 *) Clear bytes used for block padding of SSL 3.0 records.
737 [Adam Langley (Google)]
739 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
740 Kadianakis <desnacked@gmail.com> for discovering this issue and
741 Adam Langley for preparing the fix. (CVE-2011-4619)
742 [Adam Langley (Google)]
744 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
745 [Andrey Kulikov <amdeich@gmail.com>]
747 *) Prevent malformed RFC3779 data triggering an assertion failure.
748 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
749 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
750 [Rob Austein <sra@hactrn.net>]
752 *) Improved PRNG seeding for VOS.
753 [Paul Green <Paul.Green@stratus.com>]
755 *) Fix ssl_ciph.c set-up race.
756 [Adam Langley (Google)]
758 *) Fix spurious failures in ecdsatest.c.
759 [Emilia Käsper (Google)]
761 *) Fix the BIO_f_buffer() implementation (which was mixing different
762 interpretations of the '..._len' fields).
763 [Adam Langley (Google)]
765 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
766 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
767 threads won't reuse the same blinding coefficients.
769 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
770 lock to call BN_BLINDING_invert_ex, and avoids one use of
771 BN_BLINDING_update for each BN_BLINDING structure (previously,
772 the last update always remained unused).
773 [Emilia Käsper (Google)]
775 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
776 [Bob Buckholz (Google)]
778 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
780 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
781 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
782 [Kaspar Brand <ossl@velox.ch>]
784 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
785 for multi-threaded use of ECDH. (CVE-2011-3210)
786 [Adam Langley (Google)]
788 *) Fix x509_name_ex_d2i memory leak on bad inputs.
791 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
792 signature public key algorithm by using OID xref utilities instead.
793 Before this you could only use some ECC ciphersuites with SHA1 only.
796 *) Add protection against ECDSA timing attacks as mentioned in the paper
797 by Billy Bob Brumley and Nicola Tuveri, see:
799 http://eprint.iacr.org/2011/232.pdf
801 [Billy Bob Brumley and Nicola Tuveri]
803 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
805 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
806 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
808 *) Fix bug in string printing code: if *any* escaping is enabled we must
809 escape the escape character (backslash) or the resulting string is
813 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
815 *) Disable code workaround for ancient and obsolete Netscape browsers
816 and servers: an attacker can use it in a ciphersuite downgrade attack.
817 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
820 *) Fixed J-PAKE implementation error, originally discovered by
821 Sebastien Martini, further info and confirmation from Stefan
822 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
825 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
827 *) Fix extension code to avoid race conditions which can result in a buffer
828 overrun vulnerability: resumed sessions must not be modified as they can
829 be shared by multiple threads. CVE-2010-3864
832 *) Fix WIN32 build system to correctly link an ENGINE directory into
836 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
838 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
840 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
842 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
844 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
845 context. The operation can be customised via the ctrl mechanism in
846 case ENGINEs want to include additional functionality.
849 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
852 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
853 output hashes compatible with older versions of OpenSSL.
854 [Willy Weisz <weisz@vcpc.univie.ac.at>]
856 *) Fix compression algorithm handling: if resuming a session use the
857 compression algorithm of the resumed session instead of determining
858 it from client hello again. Don't allow server to change algorithm.
861 *) Add load_crls() function to apps tidying load_certs() too. Add option
862 to verify utility to allow additional CRLs to be included.
865 *) Update OCSP request code to permit adding custom headers to the request:
866 some responders need this.
869 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
871 [Julia Lawall <julia@diku.dk>]
873 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
874 needlessly dereferenced structures, used obsolete functions and
875 didn't handle all updated verify codes correctly.
878 *) Disable MD2 in the default configuration.
881 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
882 indicate the initial BIO being pushed or popped. This makes it possible
883 to determine whether the BIO is the one explicitly called or as a result
884 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
885 it handles reference counts correctly and doesn't zero out the I/O bio
886 when it is not being explicitly popped. WARNING: applications which
887 included workarounds for the old buggy behaviour will need to be modified
888 or they could free up already freed BIOs.
891 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
892 renaming to all platforms (within the 0.9.8 branch, this was
893 done conditionally on Netware platforms to avoid a name clash).
894 [Guenter <lists@gknw.net>]
896 *) Add ECDHE and PSK support to DTLS.
897 [Michael Tuexen <tuexen@fh-muenster.de>]
899 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
903 *) Add "missing" function EVP_MD_flags() (without this the only way to
904 retrieve a digest flags is by accessing the structure directly. Update
905 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
906 or cipher is registered as in the "from" argument. Print out all
907 registered digests in the dgst usage message instead of manually
908 attempting to work them out.
911 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
912 this allows the use of compression and extensions. Change default cipher
913 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
914 by default unless an application cipher string requests it.
917 *) Alter match criteria in PKCS12_parse(). It used to try to use local
918 key ids to find matching certificates and keys but some PKCS#12 files
919 don't follow the (somewhat unwritten) rules and this strategy fails.
920 Now just gather all certificates together and the first private key
921 then look for the first certificate that matches the key.
924 *) Support use of registered digest and cipher names for dgst and cipher
925 commands instead of having to add each one as a special case. So now
932 openssl dgst -sha256 foo
934 and this works for ENGINE based algorithms too.
938 *) Update Gost ENGINE to support parameter files.
939 [Victor B. Wagner <vitus@cryptocom.ru>]
941 *) Support GeneralizedTime in ca utility.
942 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
944 *) Enhance the hash format used for certificate directory links. The new
945 form uses the canonical encoding (meaning equivalent names will work
946 even if they aren't identical) and uses SHA1 instead of MD5. This form
947 is incompatible with the older format and as a result c_rehash should
948 be used to rebuild symbolic links.
951 *) Make PKCS#8 the default write format for private keys, replacing the
952 traditional format. This form is standardised, more secure and doesn't
953 include an implicit MD5 dependency.
956 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
957 committed to OpenSSL should pass this lot as a minimum.
960 *) Add session ticket override functionality for use by EAP-FAST.
961 [Jouni Malinen <j@w1.fi>]
963 *) Modify HMAC functions to return a value. Since these can be implemented
964 in an ENGINE errors can occur.
967 *) Type-checked OBJ_bsearch_ex.
970 *) Type-checked OBJ_bsearch. Also some constification necessitated
971 by type-checking. Still to come: TXT_DB, bsearch(?),
972 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
976 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
977 seconds to a tm structure directly, instead of going through OS
978 specific date routines. This avoids any issues with OS routines such
979 as the year 2038 bug. New *_adj() functions for ASN1 time structures
980 and X509_time_adj_ex() to cover the extended range. The existing
981 X509_time_adj() is still usable and will no longer have any date issues.
984 *) Delta CRL support. New use deltas option which will attempt to locate
985 and search any appropriate delta CRLs available.
987 This work was sponsored by Google.
990 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
991 code and add additional score elements. Validate alternate CRL paths
992 as part of the CRL checking and indicate a new error "CRL path validation
993 error" in this case. Applications wanting additional details can use
994 the verify callback and check the new "parent" field. If this is not
995 NULL CRL path validation is taking place. Existing applications wont
996 see this because it requires extended CRL support which is off by
999 This work was sponsored by Google.
1002 *) Support for freshest CRL extension.
1004 This work was sponsored by Google.
1007 *) Initial indirect CRL support. Currently only supported in the CRLs
1008 passed directly and not via lookup. Process certificate issuer
1009 CRL entry extension and lookup CRL entries by bother issuer name
1010 and serial number. Check and process CRL issuer entry in IDP extension.
1012 This work was sponsored by Google.
1015 *) Add support for distinct certificate and CRL paths. The CRL issuer
1016 certificate is validated separately in this case. Only enabled if
1017 an extended CRL support flag is set: this flag will enable additional
1018 CRL functionality in future.
1020 This work was sponsored by Google.
1023 *) Add support for policy mappings extension.
1025 This work was sponsored by Google.
1028 *) Fixes to pathlength constraint, self issued certificate handling,
1029 policy processing to align with RFC3280 and PKITS tests.
1031 This work was sponsored by Google.
1034 *) Support for name constraints certificate extension. DN, email, DNS
1035 and URI types are currently supported.
1037 This work was sponsored by Google.
1040 *) To cater for systems that provide a pointer-based thread ID rather
1041 than numeric, deprecate the current numeric thread ID mechanism and
1042 replace it with a structure and associated callback type. This
1043 mechanism allows a numeric "hash" to be extracted from a thread ID in
1044 either case, and on platforms where pointers are larger than 'long',
1045 mixing is done to help ensure the numeric 'hash' is usable even if it
1046 can't be guaranteed unique. The default mechanism is to use "&errno"
1047 as a pointer-based thread ID to distinguish between threads.
1049 Applications that want to provide their own thread IDs should now use
1050 CRYPTO_THREADID_set_callback() to register a callback that will call
1051 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
1053 Note that ERR_remove_state() is now deprecated, because it is tied
1054 to the assumption that thread IDs are numeric. ERR_remove_state(0)
1055 to free the current thread's error state should be replaced by
1056 ERR_remove_thread_state(NULL).
1058 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
1059 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
1060 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
1061 application was previously providing a numeric thread callback that
1062 was inappropriate for distinguishing threads, then uniqueness might
1063 have been obtained with &errno that happened immediately in the
1064 intermediate development versions of OpenSSL; this is no longer the
1065 case, the numeric thread callback will now override the automatic use
1067 [Geoff Thorpe, with help from Bodo Moeller]
1069 *) Initial support for different CRL issuing certificates. This covers a
1070 simple case where the self issued certificates in the chain exist and
1071 the real CRL issuer is higher in the existing chain.
1073 This work was sponsored by Google.
1076 *) Removed effectively defunct crypto/store from the build.
1079 *) Revamp of STACK to provide stronger type-checking. Still to come:
1080 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
1081 ASN1_STRING, CONF_VALUE.
1084 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
1085 RAM on SSL connections. This option can save about 34k per idle SSL.
1088 *) Revamp of LHASH to provide stronger type-checking. Still to come:
1089 STACK, TXT_DB, bsearch, qsort.
1092 *) Initial support for Cryptographic Message Syntax (aka CMS) based
1093 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
1094 support for data, signedData, compressedData, digestedData and
1095 encryptedData, envelopedData types included. Scripts to check against
1096 RFC4134 examples draft and interop and consistency checks of many
1097 content types and variants.
1100 *) Add options to enc utility to support use of zlib compression BIO.
1103 *) Extend mk1mf to support importing of options and assembly language
1104 files from Configure script, currently only included in VC-WIN32.
1105 The assembly language rules can now optionally generate the source
1106 files from the associated perl scripts.
1109 *) Implement remaining functionality needed to support GOST ciphersuites.
1110 Interop testing has been performed using CryptoPro implementations.
1111 [Victor B. Wagner <vitus@cryptocom.ru>]
1113 *) s390x assembler pack.
1116 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
1120 *) Implement Opaque PRF Input TLS extension as specified in
1121 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
1122 official specification yet and no extension type assignment by
1123 IANA exists, this extension (for now) will have to be explicitly
1124 enabled when building OpenSSL by providing the extension number
1125 to use. For example, specify an option
1127 -DTLSEXT_TYPE_opaque_prf_input=0x9527
1129 to the "config" or "Configure" script to enable the extension,
1130 assuming extension number 0x9527 (which is a completely arbitrary
1131 and unofficial assignment based on the MD5 hash of the Internet
1132 Draft). Note that by doing so, you potentially lose
1133 interoperability with other TLS implementations since these might
1134 be using the same extension number for other purposes.
1136 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
1137 opaque PRF input value to use in the handshake. This will create
1138 an interal copy of the length-'len' string at 'src', and will
1139 return non-zero for success.
1141 To get more control and flexibility, provide a callback function
1144 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
1145 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
1149 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
1152 Callback function 'cb' will be called in handshakes, and is
1153 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
1154 Argument 'arg' is for application purposes (the value as given to
1155 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
1156 be provided to the callback function). The callback function
1157 has to return non-zero to report success: usually 1 to use opaque
1158 PRF input just if possible, or 2 to enforce use of the opaque PRF
1159 input. In the latter case, the library will abort the handshake
1160 if opaque PRF input is not successfully negotiated.
1162 Arguments 'peerinput' and 'len' given to the callback function
1163 will always be NULL and 0 in the case of a client. A server will
1164 see the client's opaque PRF input through these variables if
1165 available (NULL and 0 otherwise). Note that if the server
1166 provides an opaque PRF input, the length must be the same as the
1167 length of the client's opaque PRF input.
1169 Note that the callback function will only be called when creating
1170 a new session (session resumption can resume whatever was
1171 previously negotiated), and will not be called in SSL 2.0
1172 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
1173 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
1174 for applications that need to enforce opaque PRF input.
1178 *) Update ssl code to support digests other than SHA1+MD5 for handshake
1181 [Victor B. Wagner <vitus@cryptocom.ru>]
1183 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1184 RFC4507bis. The encrypted ticket format is an encrypted encoded
1185 SSL_SESSION structure, that way new session features are automatically
1188 If a client application caches session in an SSL_SESSION structure
1189 support is transparent because tickets are now stored in the encoded
1192 The SSL_CTX structure automatically generates keys for ticket
1193 protection in servers so again support should be possible
1194 with no application modification.
1196 If a client or server wishes to disable RFC4507 support then the option
1197 SSL_OP_NO_TICKET can be set.
1199 Add a TLS extension debugging callback to allow the contents of any client
1200 or server extensions to be examined.
1202 This work was sponsored by Google.
1205 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1206 OpenSSL should now compile cleanly on gcc 4.2
1207 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1209 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1210 support including streaming MAC support: this is required for GOST
1211 ciphersuite support.
1212 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1214 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1215 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1216 to output in BER and PEM format.
1219 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1220 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1221 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1222 ENGINE support for HMAC keys which are unextractable. New -mac and
1223 -macopt options to dgst utility.
1226 *) New option -sigopt to dgst utility. Update dgst to use
1227 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1228 alternative signing paramaters such as X9.31 or PSS in the dgst
1232 *) Change ssl_cipher_apply_rule(), the internal function that does
1233 the work each time a ciphersuite string requests enabling
1234 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1235 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1236 the order of disabled ciphersuites such that those ciphersuites
1237 that most recently went from enabled to disabled not only stay
1238 in order with respect to each other, but also have higher priority
1239 than other disabled ciphersuites the next time ciphersuites are
1242 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1243 the same ciphersuites as with "HIGH" alone, but in a specific
1244 order where the PSK ciphersuites come first (since they are the
1245 most recently disabled ciphersuites when "HIGH" is parsed).
1247 Also, change ssl_create_cipher_list() (using this new
1248 funcionality) such that between otherwise identical
1249 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
1253 *) Change ssl_create_cipher_list() so that it automatically
1254 arranges the ciphersuites in reasonable order before starting
1255 to process the rule string. Thus, the definition for "DEFAULT"
1256 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
1257 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
1258 This makes it much easier to arrive at a reasonable default order
1259 in applications for which anonymous ciphers are OK (meaning
1260 that you can't actually use DEFAULT).
1261 [Bodo Moeller; suggested by Victor Duchovni]
1263 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
1264 processing) into multiple integers instead of setting
1265 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
1266 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
1267 (These masks as well as the individual bit definitions are hidden
1268 away into the non-exported interface ssl/ssl_locl.h, so this
1269 change to the definition of the SSL_CIPHER structure shouldn't
1270 affect applications.) This give us more bits for each of these
1271 categories, so there is no longer a need to coagulate AES128 and
1272 AES256 into a single algorithm bit, and to coagulate Camellia128
1273 and Camellia256 into a single algorithm bit, which has led to all
1276 Thus, among other things, the kludge introduced in 0.9.7m and
1277 0.9.8e for masking out AES256 independently of AES128 or masking
1278 out Camellia256 independently of AES256 is not needed here in 0.9.9.
1280 With the change, we also introduce new ciphersuite aliases that
1281 so far were missing: "AES128", "AES256", "CAMELLIA128", and
1285 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
1286 Use the leftmost N bytes of the signature input if the input is
1287 larger than the prime q (with N being the size in bytes of q).
1290 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1291 it yet and it is largely untested.
1294 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1297 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1298 some compilers (gcc 4.2 and later) reject their use. Safestack is
1299 reimplemented. Update ASN1 to avoid use of legacy functions.
1302 *) Win32/64 targets are linked with Winsock2.
1305 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1306 to external functions. This can be used to increase CRL handling
1307 efficiency especially when CRLs are very large by (for example) storing
1308 the CRL revoked certificates in a database.
1311 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1312 new CRLs added to a directory can be used. New command line option
1313 -verify_return_error to s_client and s_server. This causes real errors
1314 to be returned by the verify callback instead of carrying on no matter
1315 what. This reflects the way a "real world" verify callback would behave.
1318 *) GOST engine, supporting several GOST algorithms and public key formats.
1319 Kindly donated by Cryptocom.
1322 *) Partial support for Issuing Distribution Point CRL extension. CRLs
1323 partitioned by DP are handled but no indirect CRL or reason partitioning
1324 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
1325 selected via a scoring technique which handles IDP and AKID in CRLs.
1328 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1329 will ultimately be used for all verify operations: this will remove the
1330 X509_STORE dependency on certificate verification and allow alternative
1331 lookup methods. X509_STORE based implementations of these two callbacks.
1334 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1335 Modify get_crl() to find a valid (unexpired) CRL if possible.
1338 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1339 this would be called X509_CRL_cmp() but that name is already used by
1340 a function that just compares CRL issuer names. Cache several CRL
1341 extensions in X509_CRL structure and cache CRLDP in X509.
1344 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1345 this maps equivalent X509_NAME structures into a consistent structure.
1346 Name comparison can then be performed rapidly using memcmp().
1349 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1353 *) Allow digests to supply their own micalg string for S/MIME type using
1354 the ctrl EVP_MD_CTRL_MICALG.
1357 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1358 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1359 ctrl. It can then customise the structure before and/or after signing
1363 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1364 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1365 to free up any added signature OIDs.
1368 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1369 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1370 digest and cipher tables. New options added to openssl utility:
1371 list-message-digest-algorithms and list-cipher-algorithms.
1374 *) Change the array representation of binary polynomials: the list
1375 of degrees of non-zero coefficients is now terminated with -1.
1376 Previously it was terminated with 0, which was also part of the
1377 value; thus, the array representation was not applicable to
1378 polynomials where t^0 has coefficient zero. This change makes
1379 the array representation useful in a more general context.
1382 *) Various modifications and fixes to SSL/TLS cipher string
1383 handling. For ECC, the code now distinguishes between fixed ECDH
1384 with RSA certificates on the one hand and with ECDSA certificates
1385 on the other hand, since these are separate ciphersuites. The
1386 unused code for Fortezza ciphersuites has been removed.
1388 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1389 (not "ECDHE"). For consistency with the code for DH
1390 certificates, use of ECDH certificates is now considered ECDH
1391 authentication, not RSA or ECDSA authentication (the latter is
1392 merely the CA's signing algorithm and not actively used in the
1395 The temporary ciphersuite alias "ECCdraft" is no longer
1396 available, and ECC ciphersuites are no longer excluded from "ALL"
1397 and "DEFAULT". The following aliases now exist for RFC 4492
1398 ciphersuites, most of these by analogy with the DH case:
1400 kECDHr - ECDH cert, signed with RSA
1401 kECDHe - ECDH cert, signed with ECDSA
1402 kECDH - ECDH cert (signed with either RSA or ECDSA)
1403 kEECDH - ephemeral ECDH
1404 ECDH - ECDH cert or ephemeral ECDH
1410 AECDH - anonymous ECDH
1411 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1415 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1416 Use correct micalg parameters depending on digest(s) in signed message.
1419 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1420 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1423 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1424 an engine to register a method. Add ENGINE lookups for methods and
1425 functional reference processing.
1428 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1429 EVP_{Sign,Verify}* which allow an application to customise the signature
1433 *) New -resign option to smime utility. This adds one or more signers
1434 to an existing PKCS#7 signedData structure. Also -md option to use an
1435 alternative message digest algorithm for signing.
1438 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1439 create PKCS7 structures containing multiple signers. Update smime
1440 application to support multiple signers.
1443 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1447 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1448 Reorganize PBE internals to lookup from a static table using NIDs,
1449 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1450 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1451 PRF which will be automatically used with PBES2.
1454 *) Replace the algorithm specific calls to generate keys in "req" with the
1458 *) Update PKCS#7 enveloped data routines to use new API. This is now
1459 supported by any public key method supporting the encrypt operation. A
1460 ctrl is added to allow the public key algorithm to examine or modify
1461 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1465 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1466 a default digest type to use. In most cases this will be SHA1 but some
1467 algorithms (such as GOST) need to specify an alternative digest. The
1468 return value indicates how strong the prefernce is 1 means optional and
1469 2 is mandatory (that is it is the only supported type). Modify
1470 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1471 use the default md. Update openssl utilities to use the default digest
1472 type for signing if it is not explicitly indicated.
1475 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1476 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1477 signing method from the key type. This effectively removes the link
1478 between digests and public key types.
1481 *) Add an OID cross reference table and utility functions. Its purpose is to
1482 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1483 rsaEncryption. This will allow some of the algorithm specific hackery
1484 needed to use the correct OID to be removed.
1487 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1488 structures for PKCS7_sign(). They are now set up by the relevant public
1492 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1495 *) Add support for key derivation (agreement) in the API, DH method and
1499 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1500 public and private key formats. As a side effect these add additional
1501 command line functionality not previously available: DSA signatures can be
1502 generated and verified using pkeyutl and DH key support and generation in
1507 [Oliver Tappe <zooey@hirschkaefer.de>]
1509 *) New make target "install_html_docs" installs HTML renditions of the
1511 [Oliver Tappe <zooey@hirschkaefer.de>]
1513 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1514 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1515 support key and parameter generation and add initial key generation
1516 functionality for RSA.
1519 *) Add functions for main EVP_PKEY_method operations. The undocumented
1520 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1521 EVP_PKEY_{encrypt,decrypt}_old.
1524 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1525 key API, doesn't do much yet.
1528 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1529 public key algorithms. New option to openssl utility:
1530 "list-public-key-algorithms" to print out info.
1533 *) Implement the Supported Elliptic Curves Extension for
1534 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1537 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1538 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1541 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1542 utilities such as rsa, dsa, dsaparam etc except they process any key
1546 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1547 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1548 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1552 *) Initial support for pluggable public key ASN1.
1553 De-spaghettify the public key ASN1 handling. Move public and private
1554 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1555 algorithm specific handling to a single module within the relevant
1556 algorithm directory. Add functions to allow (near) opaque processing
1557 of public and private key structures.
1560 *) Implement the Supported Point Formats Extension for
1561 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1564 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1565 for the psk identity [hint] and the psk callback functions to the
1566 SSL_SESSION, SSL and SSL_CTX structure.
1569 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1573 SSL_CTX_use_psk_identity_hint
1574 SSL_get_psk_identity_hint
1575 SSL_get_psk_identity
1576 SSL_use_psk_identity_hint
1578 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1580 *) Add RFC 3161 compliant time stamp request creation, response generation
1581 and response verification functionality.
1582 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1584 *) Add initial support for TLS extensions, specifically for the server_name
1585 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1586 have new members for a host name. The SSL data structure has an
1587 additional member SSL_CTX *initial_ctx so that new sessions can be
1588 stored in that context to allow for session resumption, even after the
1589 SSL has been switched to a new SSL_CTX in reaction to a client's
1590 server_name extension.
1592 New functions (subject to change):
1594 SSL_get_servername()
1595 SSL_get_servername_type()
1598 New CTRL codes and macros (subject to change):
1600 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1601 - SSL_CTX_set_tlsext_servername_callback()
1602 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1603 - SSL_CTX_set_tlsext_servername_arg()
1604 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1606 openssl s_client has a new '-servername ...' option.
1608 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1609 '-key2 ...', '-servername_fatal' (subject to change). This allows
1610 testing the HostName extension for a specific single host name ('-cert'
1611 and '-key' remain fallbacks for handshakes without HostName
1612 negotiation). If the unrecogninzed_name alert has to be sent, this by
1613 default is a warning; it becomes fatal with the '-servername_fatal'
1616 [Peter Sylvester, Remy Allais, Christophe Renou]
1618 *) Whirlpool hash implementation is added.
1621 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1622 bn(64,32). Because of instruction set limitations it doesn't have
1623 any negative impact on performance. This was done mostly in order
1624 to make it possible to share assembler modules, such as bn_mul_mont
1625 implementations, between 32- and 64-bit builds without hassle.
1628 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1629 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1633 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1634 dedicated Montgomery multiplication procedure, is introduced.
1635 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1636 "64-bit" performance on certain 32-bit targets.
1639 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1640 in SSL structures. New SSL ctrl to set maximum send fragment size.
1641 Save memory by seeting the I/O buffer sizes dynamically instead of
1642 using the maximum available value.
1645 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1646 in addition to the text details.
1649 *) Very, very preliminary EXPERIMENTAL support for printing of general
1650 ASN1 structures. This currently produces rather ugly output and doesn't
1651 handle several customised structures at all.
1654 *) Integrated support for PVK file format and some related formats such
1655 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1656 these in the 'rsa' and 'dsa' utilities.
1659 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1662 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1663 place for the (very old) "NETSCAPE" format certificates which are now
1664 handled using new ASN1 code equivalents.
1667 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1668 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1669 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1672 *) Modify CRL distribution points extension code to print out previously
1673 unsupported fields. Enhance extension setting code to allow setting of
1677 *) Add print and set support for Issuing Distribution Point CRL extension.
1680 *) Change 'Configure' script to enable Camellia by default.
1683 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
1685 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1686 Thanks to Antonio Martin, Enterprise Secure Access Research and
1687 Development, Cisco Systems, Inc. for discovering this bug and
1688 preparing a fix. (CVE-2012-0050)
1691 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
1693 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1694 of the Vaudenay padding oracle attack on CBC mode encryption
1695 which enables an efficient plaintext recovery attack against
1696 the OpenSSL implementation of DTLS. Their attack exploits timing
1697 differences arising during decryption processing. A research
1698 paper describing this attack can be found at:
1699 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1700 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1701 Security Group at Royal Holloway, University of London
1702 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1703 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1704 for preparing the fix. (CVE-2011-4108)
1705 [Robin Seggelmann, Michael Tuexen]
1707 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
1708 [Ben Laurie, Kasper <ekasper@google.com>]
1710 *) Clear bytes used for block padding of SSL 3.0 records.
1712 [Adam Langley (Google)]
1714 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1715 Kadianakis <desnacked@gmail.com> for discovering this issue and
1716 Adam Langley for preparing the fix. (CVE-2011-4619)
1717 [Adam Langley (Google)]
1719 *) Prevent malformed RFC3779 data triggering an assertion failure.
1720 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1721 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1722 [Rob Austein <sra@hactrn.net>]
1724 *) Fix ssl_ciph.c set-up race.
1725 [Adam Langley (Google)]
1727 *) Fix spurious failures in ecdsatest.c.
1728 [Emilia Käsper (Google)]
1730 *) Fix the BIO_f_buffer() implementation (which was mixing different
1731 interpretations of the '..._len' fields).
1732 [Adam Langley (Google)]
1734 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1735 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1736 threads won't reuse the same blinding coefficients.
1738 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1739 lock to call BN_BLINDING_invert_ex, and avoids one use of
1740 BN_BLINDING_update for each BN_BLINDING structure (previously,
1741 the last update always remained unused).
1742 [Emilia Käsper (Google)]
1744 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1745 for multi-threaded use of ECDH.
1746 [Adam Langley (Google)]
1748 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1751 *) Add protection against ECDSA timing attacks as mentioned in the paper
1752 by Billy Bob Brumley and Nicola Tuveri, see:
1754 http://eprint.iacr.org/2011/232.pdf
1756 [Billy Bob Brumley and Nicola Tuveri]
1758 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
1760 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1761 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1763 *) Fix bug in string printing code: if *any* escaping is enabled we must
1764 escape the escape character (backslash) or the resulting string is
1768 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
1770 *) Disable code workaround for ancient and obsolete Netscape browsers
1771 and servers: an attacker can use it in a ciphersuite downgrade attack.
1772 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1775 *) Fixed J-PAKE implementation error, originally discovered by
1776 Sebastien Martini, further info and confirmation from Stefan
1777 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1780 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
1782 *) Fix extension code to avoid race conditions which can result in a buffer
1783 overrun vulnerability: resumed sessions must not be modified as they can
1784 be shared by multiple threads. CVE-2010-3864
1787 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
1790 *) Don't reencode certificate when calculating signature: cache and use
1791 the original encoding instead. This makes signature verification of
1792 some broken encodings work correctly.
1795 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
1796 is also one of the inputs.
1797 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1799 *) Don't repeatedly append PBE algorithms to table if they already exist.
1800 Sort table on each new add. This effectively makes the table read only
1801 after all algorithms are added and subsequent calls to PKCS12_pbe_add
1805 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
1807 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
1810 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
1811 access or freeing data twice (CVE-2010-0742)
1812 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
1814 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
1815 common in certificates and some applications which only call
1816 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
1820 Reduce copying into .apps and .test in makevms.com
1821 Don't try to use blank CA certificate in CA.com
1822 Allow use of C files from original directories in maketests.com
1823 [Steven M. Schweda" <sms@antinode.info>]
1825 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1827 *) When rejecting SSL/TLS records due to an incorrect version number, never
1828 update s->server with a new major version number. As of
1829 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1830 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1831 the previous behavior could result in a read attempt at NULL when
1832 receiving specific incorrect SSL/TLS records once record payload
1833 protection is active. (CVE-2010-0740)
1834 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1836 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1837 could be crashed if the relevant tables were not present (e.g. chrooted).
1838 [Tomas Hoger <thoger@redhat.com>]
1840 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1842 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1843 [Martin Olsson, Neel Mehta]
1845 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1846 accommodate for stack sorting, always a write lock!).
1849 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1850 excessive delays in the RAND_poll(): over a minute. As a workaround
1851 include a time check in the inner Heap32Next loop too.
1854 *) The code that handled flushing of data in SSL/TLS originally used the
1855 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1856 the problem outlined in PR#1949. The fix suggested there however can
1857 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1858 of Apache). So instead simplify the code to flush unconditionally.
1859 This should be fine since flushing with no data to flush is a no op.
1862 *) Handle TLS versions 2.0 and later properly and correctly use the
1863 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1864 off ancient servers have a habit of sticking around for a while...
1867 *) Modify compression code so it frees up structures without using the
1868 ex_data callbacks. This works around a problem where some applications
1869 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1870 restarting) then use compression (e.g. SSL with compression) later.
1871 This results in significant per-connection memory leaks and
1872 has caused some security issues including CVE-2008-1678 and
1876 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1877 change when encrypting or decrypting.
1880 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1881 connect and renegotiate with servers which do not support RI.
1882 Until RI is more widely deployed this option is enabled by default.
1885 *) Add "missing" ssl ctrls to clear options and mode.
1888 *) If client attempts to renegotiate and doesn't support RI respond with
1889 a no_renegotiation alert as required by RFC5746. Some renegotiating
1890 TLS clients will continue a connection gracefully when they receive
1891 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1892 waiting for a server hello which it will never receive. Now we treat a
1893 received no_renegotiation alert as a fatal error. This is because
1894 applications requesting a renegotiation might well expect it to succeed
1895 and would have no code in place to handle the server denying it so the
1896 only safe thing to do is to terminate the connection.
1899 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1900 peer supports secure renegotiation and 0 otherwise. Print out peer
1901 renegotiation support in s_client/s_server.
1904 *) Replace the highly broken and deprecated SPKAC certification method with
1905 the updated NID creation version. This should correctly handle UTF8.
1908 *) Implement RFC5746. Re-enable renegotiation but require the extension
1909 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
1910 turns out to be a bad idea. It has been replaced by
1911 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
1912 SSL_CTX_set_options(). This is really not recommended unless you
1913 know what you are doing.
1914 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
1916 *) Fixes to stateless session resumption handling. Use initial_ctx when
1917 issuing and attempting to decrypt tickets in case it has changed during
1918 servername handling. Use a non-zero length session ID when attempting
1919 stateless session resumption: this makes it possible to determine if
1920 a resumption has occurred immediately after receiving server hello
1921 (several places in OpenSSL subtly assume this) instead of later in
1925 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1926 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1927 fixes for a few places where the return code is not checked
1929 [Julia Lawall <julia@diku.dk>]
1931 *) Add --strict-warnings option to Configure script to include devteam
1932 warnings in other configurations.
1935 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1936 makes it possible to install openssl libraries in locations which
1937 have names other than "lib", for example "/usr/lib64" which some
1939 [Steve Henson, based on patch from Jeremy Utley]
1941 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1942 X690 8.9.12 and can produce some misleading textual output of OIDs.
1943 [Steve Henson, reported by Dan Kaminsky]
1945 *) Delete MD2 from algorithm tables. This follows the recommendation in
1946 several standards that it is not used in new applications due to
1947 several cryptographic weaknesses. For binary compatibility reasons
1948 the MD2 API is still compiled in by default.
1951 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1955 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1956 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1958 [Guenter <lists@gknw.net>]
1960 *) Fix the server certificate chain building code to use X509_verify_cert(),
1961 it used to have an ad-hoc builder which was unable to cope with anything
1962 other than a simple chain.
1963 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1965 *) Don't check self signed certificate signatures in X509_verify_cert()
1966 by default (a flag can override this): it just wastes time without
1967 adding any security. As a useful side effect self signed root CAs
1968 with non-FIPS digests are now usable in FIPS mode.
1971 *) In dtls1_process_out_of_seq_message() the check if the current message
1972 is already buffered was missing. For every new message was memory
1973 allocated, allowing an attacker to perform an denial of service attack
1974 with sending out of seq handshake messages until there is no memory
1975 left. Additionally every future messege was buffered, even if the
1976 sequence number made no sense and would be part of another handshake.
1977 So only messages with sequence numbers less than 10 in advance will be
1978 buffered. (CVE-2009-1378)
1979 [Robin Seggelmann, discovered by Daniel Mentz]
1981 *) Records are buffered if they arrive with a future epoch to be
1982 processed after finishing the corresponding handshake. There is
1983 currently no limitation to this buffer allowing an attacker to perform
1984 a DOS attack with sending records with future epochs until there is no
1985 memory left. This patch adds the pqueue_size() function to detemine
1986 the size of a buffer and limits the record buffer to 100 entries.
1988 [Robin Seggelmann, discovered by Daniel Mentz]
1990 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
1991 parent structure is freed. (CVE-2009-1379)
1994 *) Handle non-blocking I/O properly in SSL_shutdown() call.
1995 [Darryl Miles <darryl-mailinglists@netbauds.net>]
1998 [Ilya O. <vrghost@gmail.com>]
2000 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2002 *) Disable renegotiation completely - this fixes a severe security
2003 problem (CVE-2009-3555) at the cost of breaking all
2004 renegotiation. Renegotiation can be re-enabled by setting
2005 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2006 run-time. This is really not recommended unless you know what
2010 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2012 *) Don't set val to NULL when freeing up structures, it is freed up by
2013 underlying code. If sizeof(void *) > sizeof(long) this can result in
2014 zeroing past the valid field. (CVE-2009-0789)
2015 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2017 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2018 checked correctly. This would allow some invalid signed attributes to
2019 appear to verify correctly. (CVE-2009-0591)
2020 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2022 *) Reject UniversalString and BMPString types with invalid lengths. This
2023 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2024 a legal length. (CVE-2009-0590)
2027 *) Set S/MIME signing as the default purpose rather than setting it
2028 unconditionally. This allows applications to override it at the store
2032 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2033 to handle some structures.
2036 *) Improve efficiency of mem_gets: don't search whole buffer each time
2038 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2040 *) New -hex option for openssl rand.
2043 *) Print out UTF8String and NumericString when parsing ASN1.
2046 *) Support NumericString type for name components.
2049 *) Allow CC in the environment to override the automatically chosen
2050 compiler. Note that nothing is done to ensure flags work with the
2054 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2056 *) Properly check EVP_VerifyFinal() and similar return values
2058 [Ben Laurie, Bodo Moeller, Google Security Team]
2060 *) Enable TLS extensions by default.
2063 *) Allow the CHIL engine to be loaded, whether the application is
2064 multithreaded or not. (This does not release the developer from the
2065 obligation to set up the dynamic locking callbacks.)
2066 [Sander Temme <sander@temme.net>]
2068 *) Use correct exit code if there is an error in dgst command.
2069 [Steve Henson; problem pointed out by Roland Dirlewanger]
2071 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2072 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2075 *) Add experimental JPAKE support, including demo authentication in
2076 s_client and s_server.
2079 *) Set the comparison function in v3_addr_canonize().
2080 [Rob Austein <sra@hactrn.net>]
2082 *) Add support for XMPP STARTTLS in s_client.
2083 [Philip Paeps <philip@freebsd.org>]
2085 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2086 to ensure that even with this option, only ciphersuites in the
2087 server's preference list will be accepted. (Note that the option
2088 applies only when resuming a session, so the earlier behavior was
2089 just about the algorithm choice for symmetric cryptography.)
2092 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2094 *) Fix NULL pointer dereference if a DTLS server received
2095 ChangeCipherSpec as first record (CVE-2009-1386).
2098 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
2099 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2102 *) The fix in 0.9.8c that supposedly got rid of unsafe
2103 double-checked locking was incomplete for RSA blinding,
2104 addressing just one layer of what turns out to have been
2105 doubly unsafe triple-checked locking.
2107 So now fix this for real by retiring the MONT_HELPER macro
2108 in crypto/rsa/rsa_eay.c.
2110 [Bodo Moeller; problem pointed out by Marius Schilder]
2112 *) Various precautionary measures:
2114 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2116 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2117 (NB: This would require knowledge of the secret session ticket key
2118 to exploit, in which case you'd be SOL either way.)
2120 - Change bn_nist.c so that it will properly handle input BIGNUMs
2121 outside the expected range.
2123 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2126 [Neel Mehta, Bodo Moeller]
2128 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2129 the load fails. Useful for distros.
2130 [Ben Laurie and the FreeBSD team]
2132 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2135 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2138 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2140 This work was sponsored by Logica.
2143 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2144 keystores. Support for SSL/TLS client authentication too.
2145 Not compiled unless enable-capieng specified to Configure.
2147 This work was sponsored by Logica.
2150 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2151 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2152 attribute creation routines such as certifcate requests and PKCS#12
2156 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2158 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2159 handshake which could lead to a cilent crash as found using the
2160 Codenomicon TLS test suite (CVE-2008-1672)
2161 [Steve Henson, Mark Cox]
2163 *) Fix double free in TLS server name extensions which could lead to
2164 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
2167 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
2169 Clear the error queue to ensure that error entries left from
2170 older function calls do not interfere with the correct operation.
2171 [Lutz Jaenicke, Erik de Castro Lopo]
2173 *) Remove root CA certificates of commercial CAs:
2175 The OpenSSL project does not recommend any specific CA and does not
2176 have any policy with respect to including or excluding any CA.
2177 Therefore it does not make any sense to ship an arbitrary selection
2178 of root CA certificates with the OpenSSL software.
2181 *) RSA OAEP patches to fix two separate invalid memory reads.
2182 The first one involves inputs when 'lzero' is greater than
2183 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
2184 before the beginning of from). The second one involves inputs where
2185 the 'db' section contains nothing but zeroes (there is a one-byte
2186 invalid read after the end of 'db').
2187 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2189 *) Partial backport from 0.9.9-dev:
2191 Introduce bn_mul_mont (dedicated Montgomery multiplication
2192 procedure) as a candidate for BIGNUM assembler implementation.
2193 While 0.9.9-dev uses assembler for various architectures, only
2194 x86_64 is available by default here in the 0.9.8 branch, and
2195 32-bit x86 is available through a compile-time setting.
2197 To try the 32-bit x86 assembler implementation, use Configure
2198 option "enable-montasm" (which exists only for this backport).
2200 As "enable-montasm" for 32-bit x86 disclaims code stability
2201 anyway, in this constellation we activate additional code
2202 backported from 0.9.9-dev for further performance improvements,
2203 namely BN_from_montgomery_word. (To enable this otherwise,
2204 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
2206 [Andy Polyakov (backport partially by Bodo Moeller)]
2208 *) Add TLS session ticket callback. This allows an application to set
2209 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
2210 values. This is useful for key rollover for example where several key
2211 sets may exist with different names.
2214 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
2215 This was broken until now in 0.9.8 releases, such that the only way
2216 a registered ENGINE could be used (assuming it initialises
2217 successfully on the host) was to explicitly set it as the default
2218 for the relevant algorithms. This is in contradiction with 0.9.7
2219 behaviour and the documentation. With this fix, when an ENGINE is
2220 registered into a given algorithm's table of implementations, the
2221 'uptodate' flag is reset so that auto-discovery will be used next
2222 time a new context for that algorithm attempts to select an
2224 [Ian Lister (tweaked by Geoff Thorpe)]
2226 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
2227 implemention in the following ways:
2229 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
2232 Lack of BER streaming support means one pass streaming processing is
2233 only supported if data is detached: setting the streaming flag is
2234 ignored for embedded content.
2236 CMS support is disabled by default and must be explicitly enabled
2237 with the enable-cms configuration option.
2240 *) Update the GMP engine glue to do direct copies between BIGNUM and
2241 mpz_t when openssl and GMP use the same limb size. Otherwise the
2242 existing "conversion via a text string export" trick is still used.
2243 [Paul Sheer <paulsheer@gmail.com>]
2245 *) Zlib compression BIO. This is a filter BIO which compressed and
2246 uncompresses any data passed through it.
2249 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
2250 RFC3394 compatible AES key wrapping.
2253 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
2254 sets string data without copying. X509_ALGOR_set0() and
2255 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
2256 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
2257 from an X509_ATTRIBUTE structure optionally checking it occurs only
2258 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
2262 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
2263 to get the expected BN_FLG_CONSTTIME behavior.
2264 [Bodo Moeller (Google)]
2268 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
2269 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
2270 - added some more tests to do_tests.pl
2271 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
2272 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
2273 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
2274 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
2275 - various changes to netware.pl to enable gcc-cross builds on Win32
2277 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
2278 - various changes to fix missing prototype warnings
2279 - fixed x86nasm.pl to create correct asm files for NASM COFF output
2280 - added AES, WHIRLPOOL and CPUID assembler code to build files
2281 - added missing AES assembler make rules to mk1mf.pl
2282 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
2283 [Guenter Knauf <eflash@gmx.net>]
2285 *) Implement certificate status request TLS extension defined in RFC3546.
2286 A client can set the appropriate parameters and receive the encoded
2287 OCSP response via a callback. A server can query the supplied parameters
2288 and set the encoded OCSP response in the callback. Add simplified examples
2289 to s_client and s_server.
2292 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
2294 *) Fix various bugs:
2295 + Binary incompatibility of ssl_ctx_st structure
2296 + DTLS interoperation with non-compliant servers
2297 + Don't call get_session_cb() without proposed session
2298 + Fix ia64 assembler code
2299 [Andy Polyakov, Steve Henson]
2301 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
2303 *) DTLS Handshake overhaul. There were longstanding issues with
2304 OpenSSL DTLS implementation, which were making it impossible for
2305 RFC 4347 compliant client to communicate with OpenSSL server.
2306 Unfortunately just fixing these incompatibilities would "cut off"
2307 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
2308 server keeps tolerating non RFC compliant syntax. The opposite is
2309 not true, 0.9.8f client can not communicate with earlier server.
2310 This update even addresses CVE-2007-4995.
2313 *) Changes to avoid need for function casts in OpenSSL: some compilers
2314 (gcc 4.2 and later) reject their use.
2315 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
2318 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2319 RFC4507bis. The encrypted ticket format is an encrypted encoded
2320 SSL_SESSION structure, that way new session features are automatically
2323 If a client application caches session in an SSL_SESSION structure
2324 support is transparent because tickets are now stored in the encoded
2327 The SSL_CTX structure automatically generates keys for ticket
2328 protection in servers so again support should be possible
2329 with no application modification.
2331 If a client or server wishes to disable RFC4507 support then the option
2332 SSL_OP_NO_TICKET can be set.
2334 Add a TLS extension debugging callback to allow the contents of any client
2335 or server extensions to be examined.
2337 This work was sponsored by Google.
2340 *) Add initial support for TLS extensions, specifically for the server_name
2341 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2342 have new members for a host name. The SSL data structure has an
2343 additional member SSL_CTX *initial_ctx so that new sessions can be
2344 stored in that context to allow for session resumption, even after the
2345 SSL has been switched to a new SSL_CTX in reaction to a client's
2346 server_name extension.
2348 New functions (subject to change):
2350 SSL_get_servername()
2351 SSL_get_servername_type()
2354 New CTRL codes and macros (subject to change):
2356 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2357 - SSL_CTX_set_tlsext_servername_callback()
2358 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2359 - SSL_CTX_set_tlsext_servername_arg()
2360 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2362 openssl s_client has a new '-servername ...' option.
2364 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2365 '-key2 ...', '-servername_fatal' (subject to change). This allows
2366 testing the HostName extension for a specific single host name ('-cert'
2367 and '-key' remain fallbacks for handshakes without HostName
2368 negotiation). If the unrecogninzed_name alert has to be sent, this by
2369 default is a warning; it becomes fatal with the '-servername_fatal'
2372 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2374 *) Add AES and SSE2 assembly language support to VC++ build.
2377 *) Mitigate attack on final subtraction in Montgomery reduction.
2380 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2381 (which previously caused an internal error).
2384 *) Squeeze another 10% out of IGE mode when in != out.
2387 *) AES IGE mode speedup.
2388 [Dean Gaudet (Google)]
2390 *) Add the Korean symmetric 128-bit cipher SEED (see
2391 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2392 add SEED ciphersuites from RFC 4162:
2394 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2395 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2396 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2397 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2399 To minimize changes between patchlevels in the OpenSSL 0.9.8
2400 series, SEED remains excluded from compilation unless OpenSSL
2401 is configured with 'enable-seed'.
2402 [KISA, Bodo Moeller]
2404 *) Mitigate branch prediction attacks, which can be practical if a
2405 single processor is shared, allowing a spy process to extract
2406 information. For detailed background information, see
2407 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2408 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2409 and Necessary Software Countermeasures"). The core of the change
2410 are new versions BN_div_no_branch() and
2411 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2412 respectively, which are slower, but avoid the security-relevant
2413 conditional branches. These are automatically called by BN_div()
2414 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2415 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2416 remove a conditional branch.
2418 BN_FLG_CONSTTIME is the new name for the previous
2419 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2420 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2421 in the exponent causes BN_mod_exp_mont() to use the alternative
2422 implementation in BN_mod_exp_mont_consttime().) The old name
2423 remains as a deprecated alias.
2425 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2426 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2427 constant-time implementations for more than just exponentiation.
2428 Here too the old name is kept as a deprecated alias.
2430 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2431 the BN_BLINDING structure gets an independent copy of the
2432 modulus. This means that the previous "BIGNUM *m" argument to
2433 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2434 essentially becomes "const BIGNUM *m", although we can't actually
2435 change this in the header file before 0.9.9. It allows
2436 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2437 enable BN_FLG_CONSTTIME.
2439 [Matthew D Wood (Intel Corp)]
2441 *) In the SSL/TLS server implementation, be strict about session ID
2442 context matching (which matters if an application uses a single
2443 external cache for different purposes). Previously,
2444 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2445 set. This did ensure strict client verification, but meant that,
2446 with applications using a single external cache for quite
2447 different requirements, clients could circumvent ciphersuite
2448 restrictions for a given session ID context by starting a session
2449 in a different context.
2452 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2453 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2454 authentication-only ciphersuites.
2457 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2458 not complete and could lead to a possible single byte overflow
2459 (CVE-2007-5135) [Ben Laurie]
2461 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2463 *) Since AES128 and AES256 (and similarly Camellia128 and
2464 Camellia256) share a single mask bit in the logic of
2465 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2466 kludge to work properly if AES128 is available and AES256 isn't
2467 (or if Camellia128 is available and Camellia256 isn't).
2470 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2471 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2472 When a point or a seed is encoded in a BIT STRING, we need to
2473 prevent the removal of trailing zero bits to get the proper DER
2474 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2475 of a NamedBitList, for which trailing 0 bits need to be removed.)
2478 *) Have SSL/TLS server implementation tolerate "mismatched" record
2479 protocol version while receiving ClientHello even if the
2480 ClientHello is fragmented. (The server can't insist on the
2481 particular protocol version it has chosen before the ServerHello
2482 message has informed the client about his choice.)
2485 *) Add RFC 3779 support.
2486 [Rob Austein for ARIN, Ben Laurie]
2488 *) Load error codes if they are not already present instead of using a
2489 static variable. This allows them to be cleanly unloaded and reloaded.
2490 Improve header file function name parsing.
2493 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2494 or CAPABILITY handshake as required by RFCs.
2497 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2499 *) Introduce limits to prevent malicious keys being able to
2500 cause a denial of service. (CVE-2006-2940)
2501 [Steve Henson, Bodo Moeller]
2503 *) Fix ASN.1 parsing of certain invalid structures that can result
2504 in a denial of service. (CVE-2006-2937) [Steve Henson]
2506 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2507 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2509 *) Fix SSL client code which could crash if connecting to a
2510 malicious SSLv2 server. (CVE-2006-4343)
2511 [Tavis Ormandy and Will Drewry, Google Security Team]
2513 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2514 match only those. Before that, "AES256-SHA" would be interpreted
2515 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2516 the same strength classification in 0.9.7h) as we currently only
2517 have a single AES bit in the ciphersuite description bitmap.
2518 That change, however, also applied to ciphersuite strings such as
2519 "RC4-MD5" that intentionally matched multiple ciphersuites --
2520 namely, SSL 2.0 ciphersuites in addition to the more common ones
2521 from SSL 3.0/TLS 1.0.
2523 So we change the selection algorithm again: Naming an explicit
2524 ciphersuite selects this one ciphersuite, and any other similar
2525 ciphersuite (same bitmap) from *other* protocol versions.
2526 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2527 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2529 Since SSL 2.0 does not have any ciphersuites for which the
2530 128/256 bit distinction would be relevant, this works for now.
2531 The proper fix will be to use different bits for AES128 and
2532 AES256, which would have avoided the problems from the beginning;
2533 however, bits are scarce, so we can only do this in a new release
2534 (not just a patchlevel) when we can change the SSL_CIPHER
2535 definition to split the single 'unsigned long mask' bitmap into
2536 multiple values to extend the available space.
2540 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2542 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2543 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2545 *) Add AES IGE and biIGE modes.
2548 *) Change the Unix randomness entropy gathering to use poll() when
2549 possible instead of select(), since the latter has some
2550 undesirable limitations.
2551 [Darryl Miles via Richard Levitte and Bodo Moeller]
2553 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2554 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2555 cannot be implicitly activated as part of, e.g., the "AES" alias.
2556 However, please upgrade to OpenSSL 0.9.9[-dev] for
2557 non-experimental use of the ECC ciphersuites to get TLS extension
2558 support, which is required for curve and point format negotiation
2559 to avoid potential handshake problems.
2562 *) Disable rogue ciphersuites:
2564 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2565 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2566 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2568 The latter two were purportedly from
2569 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2572 Also deactivate the remaining ciphersuites from
2573 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2574 unofficial, and the ID has long expired.
2577 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2578 dual-core machines) and other potential thread-safety issues.
2581 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2582 versions), which is now available for royalty-free use
2583 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2584 Also, add Camellia TLS ciphersuites from RFC 4132.
2586 To minimize changes between patchlevels in the OpenSSL 0.9.8
2587 series, Camellia remains excluded from compilation unless OpenSSL
2588 is configured with 'enable-camellia'.
2591 *) Disable the padding bug check when compression is in use. The padding
2592 bug check assumes the first packet is of even length, this is not
2593 necessarily true if compresssion is enabled and can result in false
2594 positives causing handshake failure. The actual bug test is ancient
2595 code so it is hoped that implementations will either have fixed it by
2596 now or any which still have the bug do not support compression.
2599 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2601 *) When applying a cipher rule check to see if string match is an explicit
2602 cipher suite and only match that one cipher suite if it is.
2605 *) Link in manifests for VC++ if needed.
2606 [Austin Ziegler <halostatue@gmail.com>]
2608 *) Update support for ECC-based TLS ciphersuites according to
2609 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2610 TLS extensions, which are supported starting with the 0.9.9
2611 branch, not in the OpenSSL 0.9.8 branch).
2614 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2615 opaque EVP_CIPHER_CTX handling.
2618 *) Fixes and enhancements to zlib compression code. We now only use
2619 "zlib1.dll" and use the default __cdecl calling convention on Win32
2620 to conform with the standards mentioned here:
2621 http://www.zlib.net/DLL_FAQ.txt
2622 Static zlib linking now works on Windows and the new --with-zlib-include
2623 --with-zlib-lib options to Configure can be used to supply the location
2624 of the headers and library. Gracefully handle case where zlib library
2628 *) Several fixes and enhancements to the OID generation code. The old code
2629 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2630 handle numbers larger than ULONG_MAX, truncated printing and had a
2631 non standard OBJ_obj2txt() behaviour.
2634 *) Add support for building of engines under engine/ as shared libraries
2635 under VC++ build system.
2638 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2639 Hopefully, we will not see any false combination of paths any more.
2642 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2644 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2645 (part of SSL_OP_ALL). This option used to disable the
2646 countermeasure against man-in-the-middle protocol-version
2647 rollback in the SSL 2.0 server implementation, which is a bad
2648 idea. (CVE-2005-2969)
2650 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2651 for Information Security, National Institute of Advanced Industrial
2652 Science and Technology [AIST], Japan)]
2654 *) Add two function to clear and return the verify parameter flags.
2657 *) Keep cipherlists sorted in the source instead of sorting them at
2658 runtime, thus removing the need for a lock.
2661 *) Avoid some small subgroup attacks in Diffie-Hellman.
2662 [Nick Mathewson and Ben Laurie]
2664 *) Add functions for well-known primes.
2667 *) Extended Windows CE support.
2668 [Satoshi Nakamura and Andy Polyakov]
2670 *) Initialize SSL_METHOD structures at compile time instead of during
2671 runtime, thus removing the need for a lock.
2674 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2675 attempting to decrypt each encrypted key in turn. Add support to
2679 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2681 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2684 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2687 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2688 key into the same file any more.
2691 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2694 *) Add -utf8 command line and config file option to 'ca'.
2695 [Stefan <stf@udoma.org]
2697 *) Removed the macro des_crypt(), as it seems to conflict with some
2698 libraries. Use DES_crypt().
2701 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2702 involves renaming the source and generated shared-libs for
2703 both. The engines will accept the corrected or legacy ids
2704 ('ncipher' and '4758_cca' respectively) when binding. NB,
2705 this only applies when building 'shared'.
2706 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2708 *) Add attribute functions to EVP_PKEY structure. Modify
2709 PKCS12_create() to recognize a CSP name attribute and
2710 use it. Make -CSP option work again in pkcs12 utility.
2713 *) Add new functionality to the bn blinding code:
2714 - automatic re-creation of the BN_BLINDING parameters after
2715 a fixed number of uses (currently 32)
2716 - add new function for parameter creation
2717 - introduce flags to control the update behaviour of the
2718 BN_BLINDING parameters
2719 - hide BN_BLINDING structure
2720 Add a second BN_BLINDING slot to the RSA structure to improve
2721 performance when a single RSA object is shared among several
2725 *) Add support for DTLS.
2726 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2728 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2729 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2732 *) Remove buggy and incompletet DH cert support from
2733 ssl/ssl_rsa.c and ssl/s3_both.c
2736 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2737 the apps/openssl applications.
2740 *) Compile clean with "-Wall -Wmissing-prototypes
2741 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2742 DEBUG_SAFESTACK must also be set.
2745 *) Change ./Configure so that certain algorithms can be disabled by default.
2746 The new counterpiece to "no-xxx" is "enable-xxx".
2748 The patented RC5 and MDC2 algorithms will now be disabled unless
2749 "enable-rc5" and "enable-mdc2", respectively, are specified.
2751 (IDEA remains enabled despite being patented. This is because IDEA
2752 is frequently required for interoperability, and there is no license
2753 fee for non-commercial use. As before, "no-idea" can be used to
2754 avoid this algorithm.)
2758 *) Add processing of proxy certificates (see RFC 3820). This work was
2759 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2760 EGEE (Enabling Grids for E-science in Europe).
2763 *) RC4 performance overhaul on modern architectures/implementations, such
2764 as Intel P4, IA-64 and AMD64.
2767 *) New utility extract-section.pl. This can be used specify an alternative
2768 section number in a pod file instead of having to treat each file as
2769 a separate case in Makefile. This can be done by adding two lines to the
2772 =for comment openssl_section:XXX
2774 The blank line is mandatory.
2778 *) New arguments -certform, -keyform and -pass for s_client and s_server
2779 to allow alternative format key and certificate files and passphrase
2783 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2784 update associated structures and add various utility functions.
2786 Add new policy related verify parameters, include policy checking in
2787 standard verify code. Enhance 'smime' application with extra parameters
2788 to support policy checking and print out.
2791 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2792 Nehemiah processors. These extensions support AES encryption in hardware
2793 as well as RNG (though RNG support is currently disabled).
2794 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2796 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2799 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2800 [Andy Polyakov and a number of other people]
2802 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2803 implementation contributed by IBM.
2804 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2806 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2807 exponent rather than 'unsigned long'. There is a corresponding change to
2808 the new 'rsa_keygen' element of the RSA_METHOD structure.
2809 [Jelte Jansen, Geoff Thorpe]
2811 *) Functionality for creating the initial serial number file is now
2812 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2814 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2815 number file to 1, which is bound to cause problems. To avoid
2816 the problems while respecting compatibility between different 0.9.7
2817 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2818 CA.pl for serial number initialization. With the new release 0.9.8,
2819 we can fix the problem directly in the 'ca' utility.)
2822 *) Reduced header interdepencies by declaring more opaque objects in
2823 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2824 give fewer recursive includes, which could break lazy source code - so
2825 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2826 developers should define this symbol when building and using openssl to
2827 ensure they track the recommended behaviour, interfaces, [etc], but
2828 backwards-compatible behaviour prevails when this isn't defined.
2831 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2834 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2835 This will generate a random key of the appropriate length based on the
2836 cipher context. The EVP_CIPHER can provide its own random key generation
2837 routine to support keys of a specific form. This is used in the des and
2838 3des routines to generate a key of the correct parity. Update S/MIME
2839 code to use new functions and hence generate correct parity DES keys.
2840 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2841 valid (weak or incorrect parity).
2844 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2845 as looking them up. This is useful when the verified structure may contain
2846 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2847 present unless the new PKCS7_NO_CRL flag is asserted.
2850 *) Extend ASN1 oid configuration module. It now additionally accepts the
2853 shortName = some long name, 1.2.3.4
2856 *) Reimplemented the BN_CTX implementation. There is now no more static
2857 limitation on the number of variables it can handle nor the depth of the
2858 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2859 information can now expand as required, and rather than having a single
2860 static array of bignums, BN_CTX now uses a linked-list of such arrays
2861 allowing it to expand on demand whilst maintaining the usefulness of
2862 BN_CTX's "bundling".
2865 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2866 to allow all RSA operations to function using a single BN_CTX.
2869 *) Preliminary support for certificate policy evaluation and checking. This
2870 is initially intended to pass the tests outlined in "Conformance Testing
2871 of Relying Party Client Certificate Path Processing Logic" v1.07.
2874 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2875 remained unused and not that useful. A variety of other little bignum
2876 tweaks and fixes have also been made continuing on from the audit (see
2880 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2881 associated ASN1, EVP and SSL functions and old ASN1 macros.
2884 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2885 and this should never fail. So the return value from the use of
2886 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2887 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2890 *) BN_CTX_get() should return zero-valued bignums, providing the same
2891 initialised value as BN_new().
2892 [Geoff Thorpe, suggested by Ulf Möller]
2894 *) Support for inhibitAnyPolicy certificate extension.
2897 *) An audit of the BIGNUM code is underway, for which debugging code is
2898 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2899 is considered valid when processing BIGNUMs, and causes execution to
2900 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2901 further steps are taken to deliberately pollute unused data in BIGNUM
2902 structures to try and expose faulty code further on. For now, openssl will
2903 (in its default mode of operation) continue to tolerate the inconsistent
2904 forms that it has tolerated in the past, but authors and packagers should
2905 consider trying openssl and their own applications when compiled with
2906 these debugging symbols defined. It will help highlight potential bugs in
2907 their own code, and will improve the test coverage for OpenSSL itself. At
2908 some point, these tighter rules will become openssl's default to improve
2909 maintainability, though the assert()s and other overheads will remain only
2910 in debugging configurations. See bn.h for more details.
2911 [Geoff Thorpe, Nils Larsch, Ulf Möller]
2913 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
2914 that can only be obtained through BN_CTX_new() (which implicitly
2915 initialises it). The presence of this function only made it possible
2916 to overwrite an existing structure (and cause memory leaks).
2919 *) Because of the callback-based approach for implementing LHASH as a
2920 template type, lh_insert() adds opaque objects to hash-tables and
2921 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2922 to clean up those corresponding objects before destroying the hash table
2923 (and losing the object pointers). So some over-zealous constifications in
2924 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2925 objects as "const" and the lh_doall[_arg] callback wrappers are not
2926 prototyped to have "const" restrictions on the object pointers they are
2927 given (and so aren't required to cast them away any more).
2930 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2931 (speed) prefers to use its own implementation. The two implementations
2932 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2933 its object type properly exposed (MS_TM) instead of casting to/from "char
2934 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2935 aren't necessarily the greatest nomenclatures - but this is what was used
2936 internally to the implementation so I've used that for now.
2939 *) Ensure that deprecated functions do not get compiled when
2940 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2941 the self-tests were still using deprecated key-generation functions so
2942 these have been updated also.
2945 *) Reorganise PKCS#7 code to separate the digest location functionality
2946 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2947 New function PKCS7_set_digest() to set the digest type for PKCS#7
2948 digestedData type. Add additional code to correctly generate the
2949 digestedData type and add support for this type in PKCS7 initialization
2953 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2954 structure of type "other".
2957 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2958 sure the loop does correctly stop and breaking ("division by zero")
2959 modulus operations are not performed. The (pre-generated) prime
2960 table crypto/bn/bn_prime.h was already correct, but it could not be
2961 re-generated on some platforms because of the "division by zero"
2962 situation in the script.
2963 [Ralf S. Engelschall]
2965 *) Update support for ECC-based TLS ciphersuites according to
2966 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2967 SHA-1 now is only used for "small" curves (where the
2968 representation of a field element takes up to 24 bytes); for
2969 larger curves, the field element resulting from ECDH is directly
2970 used as premaster secret.
2971 [Douglas Stebila (Sun Microsystems Laboratories)]
2973 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
2974 curve secp160r1 to the tests.
2975 [Douglas Stebila (Sun Microsystems Laboratories)]
2977 *) Add the possibility to load symbols globally with DSO.
2978 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
2980 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
2981 control of the error stack.
2984 *) Add support for STORE in ENGINE.
2987 *) Add the STORE type. The intention is to provide a common interface
2988 to certificate and key stores, be they simple file-based stores, or
2989 HSM-type store, or LDAP stores, or...
2990 NOTE: The code is currently UNTESTED and isn't really used anywhere.
2993 *) Add a generic structure called OPENSSL_ITEM. This can be used to
2994 pass a list of arguments to any function as well as provide a way
2995 for a function to pass data back to the caller.
2998 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
2999 works like BUF_strdup() but can be used to duplicate a portion of
3000 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
3004 *) Add the function sk_find_ex() which works like sk_find(), but will
3005 return an index to an element even if an exact match couldn't be
3006 found. The index is guaranteed to point at the element where the
3007 searched-for key would be inserted to preserve sorting order.
3010 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
3011 takes an extra flags argument for optional functionality. Currently,
3012 the following flags are defined:
3014 OBJ_BSEARCH_VALUE_ON_NOMATCH
3015 This one gets OBJ_bsearch_ex() to return a pointer to the first
3016 element where the comparing function returns a negative or zero
3019 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
3020 This one gets OBJ_bsearch_ex() to return a pointer to the first
3021 element where the comparing function returns zero. This is useful
3022 if there are more than one element where the comparing function
3026 *) Make it possible to create self-signed certificates with 'openssl ca'
3027 in such a way that the self-signed certificate becomes part of the
3028 CA database and uses the same mechanisms for serial number generation
3029 as all other certificate signing. The new flag '-selfsign' enables
3030 this functionality. Adapt CA.sh and CA.pl.in.
3033 *) Add functionality to check the public key of a certificate request
3034 against a given private. This is useful to check that a certificate
3035 request can be signed by that key (self-signing).
3038 *) Make it possible to have multiple active certificates with the same
3039 subject in the CA index file. This is done only if the keyword
3040 'unique_subject' is set to 'no' in the main CA section (default
3041 if 'CA_default') of the configuration file. The value is saved
3042 with the database itself in a separate index attribute file,
3043 named like the index file with '.attr' appended to the name.
3046 *) Generate muti valued AVAs using '+' notation in config files for
3050 *) Support for nameConstraints certificate extension.
3053 *) Support for policyConstraints certificate extension.
3056 *) Support for policyMappings certificate extension.
3059 *) Make sure the default DSA_METHOD implementation only uses its
3060 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3061 and change its own handlers to be NULL so as to remove unnecessary
3062 indirection. This lets alternative implementations fallback to the
3063 default implementation more easily.
3066 *) Support for directoryName in GeneralName related extensions
3070 *) Make it possible to link applications using Makefile.shared.
3071 Make that possible even when linking against static libraries!
3074 *) Support for single pass processing for S/MIME signing. This now
3075 means that S/MIME signing can be done from a pipe, in addition
3076 cleartext signing (multipart/signed type) is effectively streaming
3077 and the signed data does not need to be all held in memory.
3079 This is done with a new flag PKCS7_STREAM. When this flag is set
3080 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3081 is done after the data is output (and digests calculated) in
3082 SMIME_write_PKCS7().
3085 *) Add full support for -rpath/-R, both in shared libraries and
3086 applications, at least on the platforms where it's known how
3090 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3091 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3092 will now compute a table of multiples of the generator that
3093 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3094 faster (notably in the case of a single point multiplication,
3095 scalar * generator).
3096 [Nils Larsch, Bodo Moeller]
3098 *) IPv6 support for certificate extensions. The various extensions
3099 which use the IP:a.b.c.d can now take IPv6 addresses using the
3100 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3104 *) Added an ENGINE that implements RSA by performing private key
3105 exponentiations with the GMP library. The conversions to and from
3106 GMP's mpz_t format aren't optimised nor are any montgomery forms
3107 cached, and on x86 it appears OpenSSL's own performance has caught up.
3108 However there are likely to be other architectures where GMP could
3109 provide a boost. This ENGINE is not built in by default, but it can be
3110 specified at Configure time and should be accompanied by the necessary
3111 linker additions, eg;
3112 ./config -DOPENSSL_USE_GMP -lgmp
3115 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3116 testing availability of engines with "-t" - the old behaviour is
3117 produced by increasing the feature's verbosity with "-tt".
3120 *) ECDSA routines: under certain error conditions uninitialized BN objects
3121 could be freed. Solution: make sure initialization is performed early
3122 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3126 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3127 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3128 software implementations. For DSA and DH, parameter generation can
3129 also be overriden by providing the appropriate method callbacks.
3132 *) Change the "progress" mechanism used in key-generation and
3133 primality testing to functions that take a new BN_GENCB pointer in
3134 place of callback/argument pairs. The new API functions have "_ex"
3135 postfixes and the older functions are reimplemented as wrappers for
3136 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3137 declarations of the old functions to help (graceful) attempts to
3138 migrate to the new functions. Also, the new key-generation API
3139 functions operate on a caller-supplied key-structure and return
3140 success/failure rather than returning a key or NULL - this is to
3141 help make "keygen" another member function of RSA_METHOD etc.
3143 Example for using the new callback interface:
3145 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3149 BN_GENCB_set(&my_cb, my_callback, my_arg);
3151 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3152 /* For the meaning of a, b in calls to my_callback(), see the
3153 * documentation of the function that calls the callback.
3154 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3155 * my_callback should return 1 if it wants BN_is_prime_ex()
3156 * to continue, or 0 to stop.
3161 *) Change the ZLIB compression method to be stateful, and make it
3162 available to TLS with the number defined in
3163 draft-ietf-tls-compression-04.txt.
3166 *) Add the ASN.1 structures and functions for CertificatePair, which
3167 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
3169 CertificatePair ::= SEQUENCE {
3170 forward [0] Certificate OPTIONAL,
3171 reverse [1] Certificate OPTIONAL,
3172 -- at least one of the pair shall be present -- }
3174 Also implement the PEM functions to read and write certificate
3175 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
3177 This needed to be defined, mostly for the sake of the LDAP
3178 attribute crossCertificatePair, but may prove useful elsewhere as
3182 *) Make it possible to inhibit symlinking of shared libraries in
3183 Makefile.shared, for Cygwin's sake.
3186 *) Extend the BIGNUM API by creating a function
3187 void BN_set_negative(BIGNUM *a, int neg);
3188 and a macro that behave like
3189 int BN_is_negative(const BIGNUM *a);
3191 to avoid the need to access 'a->neg' directly in applications.
3194 *) Implement fast modular reduction for pseudo-Mersenne primes
3195 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
3196 EC_GROUP_new_curve_GFp() will now automatically use this
3198 [Nils Larsch <nla@trustcenter.de>]
3200 *) Add new lock type (CRYPTO_LOCK_BN).
3203 *) Change the ENGINE framework to automatically load engines
3204 dynamically from specific directories unless they could be
3205 found to already be built in or loaded. Move all the
3206 current engines except for the cryptodev one to a new
3208 The engines in engines/ are built as shared libraries if
3209 the "shared" options was given to ./Configure or ./config.
3210 Otherwise, they are inserted in libcrypto.a.
3211 /usr/local/ssl/engines is the default directory for dynamic
3212 engines, but that can be overriden at configure time through
3213 the usual use of --prefix and/or --openssldir, and at run
3214 time with the environment variable OPENSSL_ENGINES.
3215 [Geoff Thorpe and Richard Levitte]
3217 *) Add Makefile.shared, a helper makefile to build shared
3218 libraries. Addapt Makefile.org.
3221 *) Add version info to Win32 DLLs.
3222 [Peter 'Luna' Runestig" <peter@runestig.com>]
3224 *) Add new 'medium level' PKCS#12 API. Certificates and keys
3225 can be added using this API to created arbitrary PKCS#12
3226 files while avoiding the low level API.
3228 New options to PKCS12_create(), key or cert can be NULL and
3229 will then be omitted from the output file. The encryption
3230 algorithm NIDs can be set to -1 for no encryption, the mac
3231 iteration count can be set to 0 to omit the mac.
3233 Enhance pkcs12 utility by making the -nokeys and -nocerts
3234 options work when creating a PKCS#12 file. New option -nomac
3235 to omit the mac, NONE can be set for an encryption algorithm.
3236 New code is modified to use the enhanced PKCS12_create()
3237 instead of the low level API.
3240 *) Extend ASN1 encoder to support indefinite length constructed
3241 encoding. This can output sequences tags and octet strings in
3242 this form. Modify pk7_asn1.c to support indefinite length
3243 encoding. This is experimental and needs additional code to
3244 be useful, such as an ASN1 bio and some enhanced streaming
3247 Extend template encode functionality so that tagging is passed
3248 down to the template encoder.
3251 *) Let 'openssl req' fail if an argument to '-newkey' is not
3252 recognized instead of using RSA as a default.
3255 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
3256 As these are not official, they are not included in "ALL";
3257 the "ECCdraft" ciphersuite group alias can be used to select them.
3258 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
3260 *) Add ECDH engine support.
3261 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
3263 *) Add ECDH in new directory crypto/ecdh/.
3264 [Douglas Stebila (Sun Microsystems Laboratories)]
3266 *) Let BN_rand_range() abort with an error after 100 iterations
3267 without success (which indicates a broken PRNG).
3270 *) Change BN_mod_sqrt() so that it verifies that the input value
3271 is really the square of the return value. (Previously,
3272 BN_mod_sqrt would show GIGO behaviour.)
3275 *) Add named elliptic curves over binary fields from X9.62, SECG,
3276 and WAP/WTLS; add OIDs that were still missing.
3278 [Sheueling Chang Shantz and Douglas Stebila
3279 (Sun Microsystems Laboratories)]
3281 *) Extend the EC library for elliptic curves over binary fields
3282 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
3285 EC_GF2m_simple_method
3289 EC_GROUP_new_curve_GF2m
3290 EC_GROUP_set_curve_GF2m
3291 EC_GROUP_get_curve_GF2m
3292 EC_POINT_set_affine_coordinates_GF2m
3293 EC_POINT_get_affine_coordinates_GF2m
3294 EC_POINT_set_compressed_coordinates_GF2m
3296 Point compression for binary fields is disabled by default for
3297 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
3300 As binary polynomials are represented as BIGNUMs, various members
3301 of the EC_GROUP and EC_POINT data structures can be shared
3302 between the implementations for prime fields and binary fields;
3303 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
3304 are essentially identical to their ..._GFp counterparts.
3305 (For simplicity, the '..._GFp' prefix has been dropped from
3306 various internal method names.)
3308 An internal 'field_div' method (similar to 'field_mul' and
3309 'field_sqr') has been added; this is used only for binary fields.
3311 [Sheueling Chang Shantz and Douglas Stebila
3312 (Sun Microsystems Laboratories)]
3314 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
3315 through methods ('mul', 'precompute_mult').
3317 The generic implementations (now internally called 'ec_wNAF_mul'
3318 and 'ec_wNAF_precomputed_mult') remain the default if these
3319 methods are undefined.
3321 [Sheueling Chang Shantz and Douglas Stebila
3322 (Sun Microsystems Laboratories)]
3324 *) New function EC_GROUP_get_degree, which is defined through
3325 EC_METHOD. For curves over prime fields, this returns the bit
3326 length of the modulus.
3328 [Sheueling Chang Shantz and Douglas Stebila
3329 (Sun Microsystems Laboratories)]
3331 *) New functions EC_GROUP_dup, EC_POINT_dup.
3332 (These simply call ..._new and ..._copy).
3334 [Sheueling Chang Shantz and Douglas Stebila
3335 (Sun Microsystems Laboratories)]
3337 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3338 Polynomials are represented as BIGNUMs (where the sign bit is not
3339 used) in the following functions [macros]:
3342 BN_GF2m_sub [= BN_GF2m_add]
3343 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3344 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3345 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3347 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3348 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3349 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3350 BN_GF2m_cmp [= BN_ucmp]
3352 (Note that only the 'mod' functions are actually for fields GF(2^m).
3353 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3355 For some functions, an the irreducible polynomial defining a
3356 field can be given as an 'unsigned int[]' with strictly
3357 decreasing elements giving the indices of those bits that are set;
3358 i.e., p[] represents the polynomial
3359 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3361 p[0] > p[1] > ... > p[k] = 0.
3362 This applies to the following functions:
3367 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3368 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3370 BN_GF2m_mod_sqrt_arr
3371 BN_GF2m_mod_solve_quad_arr
3375 Conversion can be performed by the following functions:
3380 bntest.c has additional tests for binary polynomial arithmetic.
3382 Two implementations for BN_GF2m_mod_div() are available.
3383 The default algorithm simply uses BN_GF2m_mod_inv() and
3384 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3385 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3386 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3388 [Sheueling Chang Shantz and Douglas Stebila
3389 (Sun Microsystems Laboratories)]
3391 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3392 functionality is disabled at compile-time.
3393 [Douglas Stebila <douglas.stebila@sun.com>]
3395 *) Change default behaviour of 'openssl asn1parse' so that more
3396 information is visible when viewing, e.g., a certificate:
3398 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3399 mode the content of non-printable OCTET STRINGs is output in a
3400 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3401 avoid the appearance of a printable string.
3402 [Nils Larsch <nla@trustcenter.de>]
3404 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3406 EC_GROUP_set_asn1_flag()
3407 EC_GROUP_get_asn1_flag()
3408 EC_GROUP_set_point_conversion_form()
3409 EC_GROUP_get_point_conversion_form()
3410 These control ASN1 encoding details:
3411 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3412 has been set to OPENSSL_EC_NAMED_CURVE.
3413 - Points are encoded in uncompressed form by default; options for
3414 asn1_for are as for point2oct, namely
3415 POINT_CONVERSION_COMPRESSED
3416 POINT_CONVERSION_UNCOMPRESSED
3417 POINT_CONVERSION_HYBRID
3419 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3422 EC_GROUP_get0_seed()
3423 EC_GROUP_get_seed_len()
3424 This is used only for ASN1 purposes (so far).
3425 [Nils Larsch <nla@trustcenter.de>]
3427 *) Add 'field_type' member to EC_METHOD, which holds the NID
3428 of the appropriate field type OID. The new function
3429 EC_METHOD_get_field_type() returns this value.
3430 [Nils Larsch <nla@trustcenter.de>]
3435 EC_POINT_point2hex()
3436 EC_POINT_hex2point()
3437 providing useful interfaces to EC_POINT_point2oct() and
3438 EC_POINT_oct2point().
3439 [Nils Larsch <nla@trustcenter.de>]
3441 *) Change internals of the EC library so that the functions
3442 EC_GROUP_set_generator()
3443 EC_GROUP_get_generator()
3444 EC_GROUP_get_order()
3445 EC_GROUP_get_cofactor()
3446 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3447 to methods, which would lead to unnecessary code duplication when
3448 adding different types of curves.
3449 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3451 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3452 arithmetic, and such that modified wNAFs are generated
3453 (which avoid length expansion in many cases).
3456 *) Add a function EC_GROUP_check_discriminant() (defined via
3457 EC_METHOD) that verifies that the curve discriminant is non-zero.
3459 Add a function EC_GROUP_check() that makes some sanity tests
3460 on a EC_GROUP, its generator and order. This includes
3461 EC_GROUP_check_discriminant().
3462 [Nils Larsch <nla@trustcenter.de>]
3464 *) Add ECDSA in new directory crypto/ecdsa/.
3466 Add applications 'openssl ecparam' and 'openssl ecdsa'
3467 (these are based on 'openssl dsaparam' and 'openssl dsa').
3469 ECDSA support is also included in various other files across the
3470 library. Most notably,
3471 - 'openssl req' now has a '-newkey ecdsa:file' option;
3472 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3473 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3474 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3475 them suitable for ECDSA where domain parameters must be
3476 extracted before the specific public key;
3477 - ECDSA engine support has been added.
3478 [Nils Larsch <nla@trustcenter.de>]
3480 *) Include some named elliptic curves, and add OIDs from X9.62,
3481 SECG, and WAP/WTLS. Each curve can be obtained from the new
3483 EC_GROUP_new_by_curve_name(),
3484 and the list of available named curves can be obtained with
3485 EC_get_builtin_curves().
3486 Also add a 'curve_name' member to EC_GROUP objects, which can be
3488 EC_GROUP_set_curve_name()
3489 EC_GROUP_get_curve_name()
3490 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3492 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3493 was actually never needed) and in BN_mul(). The removal in BN_mul()
3494 required a small change in bn_mul_part_recursive() and the addition
3495 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3496 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3497 bn_sub_words() and bn_add_words() except they take arrays with
3501 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3503 *) Cleanse PEM buffers before freeing them since they may contain
3505 [Benjamin Bennett <ben@psc.edu>]
3507 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3508 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3509 authentication-only ciphersuites.
3512 *) Since AES128 and AES256 share a single mask bit in the logic of
3513 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3514 kludge to work properly if AES128 is available and AES256 isn't.
3517 *) Expand security boundary to match 1.1.1 module.
3520 *) Remove redundant features: hash file source, editing of test vectors
3521 modify fipsld to use external fips_premain.c signature.
3524 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3525 run algorithm test programs.
3528 *) Make algorithm test programs more tolerant of whitespace.
3531 *) Have SSL/TLS server implementation tolerate "mismatched" record
3532 protocol version while receiving ClientHello even if the
3533 ClientHello is fragmented. (The server can't insist on the
3534 particular protocol version it has chosen before the ServerHello
3535 message has informed the client about his choice.)
3538 *) Load error codes if they are not already present instead of using a
3539 static variable. This allows them to be cleanly unloaded and reloaded.
3542 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3544 *) Introduce limits to prevent malicious keys being able to
3545 cause a denial of service. (CVE-2006-2940)
3546 [Steve Henson, Bodo Moeller]
3548 *) Fix ASN.1 parsing of certain invalid structures that can result
3549 in a denial of service. (CVE-2006-2937) [Steve Henson]
3551 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3552 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3554 *) Fix SSL client code which could crash if connecting to a
3555 malicious SSLv2 server. (CVE-2006-4343)
3556 [Tavis Ormandy and Will Drewry, Google Security Team]
3558 *) Change ciphersuite string processing so that an explicit
3559 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3560 will no longer include "AES128-SHA"), and any other similar
3561 ciphersuite (same bitmap) from *other* protocol versions (so that
3562 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3563 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3564 changes from 0.9.8b and 0.9.8d.
3567 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3569 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3570 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3572 *) Change the Unix randomness entropy gathering to use poll() when
3573 possible instead of select(), since the latter has some
3574 undesirable limitations.
3575 [Darryl Miles via Richard Levitte and Bodo Moeller]
3577 *) Disable rogue ciphersuites:
3579 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3580 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3581 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3583 The latter two were purportedly from
3584 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3587 Also deactive the remaining ciphersuites from
3588 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3589 unofficial, and the ID has long expired.
3592 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3593 dual-core machines) and other potential thread-safety issues.
3596 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3598 *) Adapt fipsld and the build system to link against the validated FIPS
3599 module in FIPS mode.
3602 *) Fixes for VC++ 2005 build under Windows.
3605 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3606 from a Windows bash shell such as MSYS. It is autodetected from the
3607 "config" script when run from a VC++ environment. Modify standard VC++
3608 build to use fipscanister.o from the GNU make build.
3611 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3613 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3614 The value now differs depending on if you build for FIPS or not.
3615 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3616 safely run with a non-FIPSed libcrypto, as it may crash because of
3617 the difference induced by this change.
3620 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3622 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3623 (part of SSL_OP_ALL). This option used to disable the
3624 countermeasure against man-in-the-middle protocol-version
3625 rollback in the SSL 2.0 server implementation, which is a bad
3626 idea. (CVE-2005-2969)
3628 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3629 for Information Security, National Institute of Advanced Industrial
3630 Science and Technology [AIST], Japan)]
3632 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3633 mainly for FIPS compliance and not fully integrated at this stage.
3636 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3637 the exponentiation using a fixed-length exponent. (Otherwise,
3638 the information leaked through timing could expose the secret key
3639 after many signatures; cf. Bleichenbacher's attack on DSA with
3643 *) Make a new fixed-window mod_exp implementation the default for
3644 RSA, DSA, and DH private-key operations so that the sequence of
3645 squares and multiplies and the memory access pattern are
3646 independent of the particular secret key. This will mitigate
3647 cache-timing and potential related attacks.
3649 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3650 and this is automatically used by BN_mod_exp_mont() if the new flag
3651 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3652 will use this BN flag for private exponents unless the flag
3653 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3654 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3656 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3658 *) Change the client implementation for SSLv23_method() and
3659 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3660 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3661 (Previously, the SSL 2.0 backwards compatible Client Hello
3662 message format would be used even with SSL_OP_NO_SSLv2.)
3665 *) Add support for smime-type MIME parameter in S/MIME messages which some
3669 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3670 a threadsafe manner. Modify rsa code to use new function and add calls
3671 to dsa and dh code (which had race conditions before).
3674 *) Include the fixed error library code in the C error file definitions
3675 instead of fixing them up at runtime. This keeps the error code
3676 structures constant.
3679 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3681 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3684 *) Fixes for newer kerberos headers. NB: the casts are needed because
3685 the 'length' field is signed on one version and unsigned on another
3686 with no (?) obvious way to tell the difference, without these VC++
3687 complains. Also the "definition" of FAR (blank) is no longer included
3688 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3689 some needed definitions.
3692 *) Undo Cygwin change.
3695 *) Added support for proxy certificates according to RFC 3820.
3696 Because they may be a security thread to unaware applications,
3697 they must be explicitely allowed in run-time. See
3698 docs/HOWTO/proxy_certificates.txt for further information.
3701 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3703 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3704 server and client random values. Previously
3705 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3706 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3708 This change has negligible security impact because:
3710 1. Server and client random values still have 24 bytes of pseudo random
3713 2. Server and client random values are sent in the clear in the initial
3716 3. The master secret is derived using the premaster secret (48 bytes in
3717 size for static RSA ciphersuites) as well as client server and random
3720 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3723 [Stephen Henson, reported by UK NISCC]
3725 *) Use Windows randomness collection on Cygwin.
3728 *) Fix hang in EGD/PRNGD query when communication socket is closed
3729 prematurely by EGD/PRNGD.
3730 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3732 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3735 *) Back-port of selected performance improvements from development
3736 branch, as well as improved support for PowerPC platforms.
3739 *) Add lots of checks for memory allocation failure, error codes to indicate
3740 failure and freeing up memory if a failure occurs.
3741 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3743 *) Add new -passin argument to dgst.
3746 *) Perform some character comparisons of different types in X509_NAME_cmp:
3747 this is needed for some certificates that reencode DNs into UTF8Strings
3748 (in violation of RFC3280) and can't or wont issue name rollover
3752 *) Make an explicit check during certificate validation to see that
3753 the CA setting in each certificate on the chain is correct. As a
3754 side effect always do the following basic checks on extensions,
3755 not just when there's an associated purpose to the check:
3757 - if there is an unhandled critical extension (unless the user
3758 has chosen to ignore this fault)
3759 - if the path length has been exceeded (if one is set at all)
3760 - that certain extensions fit the associated purpose (if one has
3764 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3766 *) Avoid a race condition when CRLs are checked in a multi threaded
3767 environment. This would happen due to the reordering of the revoked
3768 entries during signature checking and serial number lookup. Now the
3769 encoding is cached and the serial number sort performed under a lock.
3770 Add new STACK function sk_is_sorted().
3773 *) Add Delta CRL to the extension code.
3776 *) Various fixes to s3_pkt.c so alerts are sent properly.
3777 [David Holmes <d.holmes@f5.com>]
3779 *) Reduce the chances of duplicate issuer name and serial numbers (in
3780 violation of RFC3280) using the OpenSSL certificate creation utilities.
3781 This is done by creating a random 64 bit value for the initial serial
3782 number when a serial number file is created or when a self signed
3783 certificate is created using 'openssl req -x509'. The initial serial
3784 number file is created using 'openssl x509 -next_serial' in CA.pl
3785 rather than being initialized to 1.
3788 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3790 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3791 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3792 [Joe Orton, Steve Henson]
3794 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3796 [Joe Orton, Steve Henson]
3798 *) Make it possible to have multiple active certificates with the same
3799 subject in the CA index file. This is done only if the keyword
3800 'unique_subject' is set to 'no' in the main CA section (default
3801 if 'CA_default') of the configuration file. The value is saved
3802 with the database itself in a separate index attribute file,
3803 named like the index file with '.attr' appended to the name.
3806 *) X509 verify fixes. Disable broken certificate workarounds when
3807 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3808 keyUsage extension present. Don't accept CRLs with unhandled critical
3809 extensions: since verify currently doesn't process CRL extensions this
3810 rejects a CRL with *any* critical extensions. Add new verify error codes
3814 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3815 A clarification of RFC2560 will require the use of OCTET STRINGs and
3816 some implementations cannot handle the current raw format. Since OpenSSL
3817 copies and compares OCSP nonces as opaque blobs without any attempt at
3818 parsing them this should not create any compatibility issues.
3821 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3822 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3823 this HMAC (and other) operations are several times slower than OpenSSL
3827 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3828 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3830 *) Use the correct content when signing type "other".
3833 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3835 *) Fix various bugs revealed by running the NISCC test suite:
3837 Stop out of bounds reads in the ASN1 code when presented with
3838 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3840 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3842 If verify callback ignores invalid public key errors don't try to check
3843 certificate signature with the NULL public key.
3847 *) New -ignore_err option in ocsp application to stop the server
3848 exiting on the first error in a request.
3851 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3852 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3856 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3857 extra data after the compression methods not only for TLS 1.0
3858 but also for SSL 3.0 (as required by the specification).
3859 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3861 *) Change X509_certificate_type() to mark the key as exported/exportable
3862 when it's 512 *bits* long, not 512 bytes.
3865 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3866 blocks during encryption.
3869 *) Various fixes to base64 BIO and non blocking I/O. On write
3870 flushes were not handled properly if the BIO retried. On read
3871 data was not being buffered properly and had various logic bugs.
3872 This also affects blocking I/O when the data being decoded is a
3876 *) Various S/MIME bugfixes and compatibility changes:
3877 output correct application/pkcs7 MIME type if
3878 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3879 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3880 of files as .eml work). Correctly handle very long lines in MIME
3884 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3886 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3887 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3888 a protocol version number mismatch like a decryption error
3889 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3892 *) Turn on RSA blinding by default in the default implementation
3893 to avoid a timing attack. Applications that don't want it can call
3894 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3895 They would be ill-advised to do so in most cases.
3896 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3898 *) Change RSA blinding code so that it works when the PRNG is not
3899 seeded (in this case, the secret RSA exponent is abused as
3900 an unpredictable seed -- if it is not unpredictable, there
3901 is no point in blinding anyway). Make RSA blinding thread-safe
3902 by remembering the creator's thread ID in rsa->blinding and
3903 having all other threads use local one-time blinding factors
3904 (this requires more computation than sharing rsa->blinding, but
3905 avoids excessive locking; and if an RSA object is not shared
3906 between threads, blinding will still be very fast).
3909 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
3910 ENGINE as defaults for all supported algorithms irrespective of
3911 the 'flags' parameter. 'flags' is now honoured, so applications
3912 should make sure they are passing it correctly.
3915 *) Target "mingw" now allows native Windows code to be generated in
3916 the Cygwin environment as well as with the MinGW compiler.
3919 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
3921 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3922 via timing by performing a MAC computation even if incorrrect
3923 block cipher padding has been found. This is a countermeasure
3924 against active attacks where the attacker has to distinguish
3925 between bad padding and a MAC verification error. (CVE-2003-0078)
3927 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3928 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3929 Martin Vuagnoux (EPFL, Ilion)]
3931 *) Make the no-err option work as intended. The intention with no-err
3932 is not to have the whole error stack handling routines removed from
3933 libcrypto, it's only intended to remove all the function name and
3934 reason texts, thereby removing some of the footprint that may not
3935 be interesting if those errors aren't displayed anyway.
3937 NOTE: it's still possible for any application or module to have it's
3938 own set of error texts inserted. The routines are there, just not
3939 used by default when no-err is given.
3942 *) Add support for FreeBSD on IA64.
3943 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
3945 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
3946 Kerberos function mit_des_cbc_cksum(). Before this change,
3947 the value returned by DES_cbc_cksum() was like the one from
3948 mit_des_cbc_cksum(), except the bytes were swapped.
3949 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
3951 *) Allow an application to disable the automatic SSL chain building.
3952 Before this a rather primitive chain build was always performed in
3953 ssl3_output_cert_chain(): an application had no way to send the
3954 correct chain if the automatic operation produced an incorrect result.
3956 Now the chain builder is disabled if either:
3958 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
3960 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
3962 The reasoning behind this is that an application would not want the
3963 auto chain building to take place if extra chain certificates are
3964 present and it might also want a means of sending no additional
3965 certificates (for example the chain has two certificates and the
3969 *) Add the possibility to build without the ENGINE framework.
3970 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3972 *) Under Win32 gmtime() can return NULL: check return value in
3973 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
3976 *) DSA routines: under certain error conditions uninitialized BN objects
3977 could be freed. Solution: make sure initialization is performed early
3978 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
3979 Nils Larsch <nla@trustcenter.de> via PR#459)
3982 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
3983 checked on reconnect on the client side, therefore session resumption
3984 could still fail with a "ssl session id is different" error. This
3985 behaviour is masked when SSL_OP_ALL is used due to
3986 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
3987 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3988 followup to PR #377.
3991 *) IA-32 assembler support enhancements: unified ELF targets, support
3992 for SCO/Caldera platforms, fix for Cygwin shared build.
3995 *) Add support for FreeBSD on sparc64. As a consequence, support for
3996 FreeBSD on non-x86 processors is separate from x86 processors on
3997 the config script, much like the NetBSD support.
3998 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
4000 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
4002 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
4005 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
4006 code (06) was taken as the first octet of the session ID and the last
4007 octet was ignored consequently. As a result SSLv2 client side session
4008 caching could not have worked due to the session ID mismatch between
4010 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4014 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
4015 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
4019 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
4020 seems that in spite of existing for more than a year, many application
4021 author have done nothing to provide the necessary callbacks, which
4022 means that this particular engine will not work properly anywhere.
4023 This is a very unfortunate situation which forces us, in the name
4024 of usability, to give the hw_ncipher.c a static lock, which is part
4026 NOTE: This is for the 0.9.7 series ONLY. This hack will never
4027 appear in 0.9.8 or later. We EXPECT application authors to have
4028 dealt properly with this when 0.9.8 is released (unless we actually
4029 make such changes in the libcrypto locking code that changes will
4030 have to be made anyway).
4033 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
4034 octets have been read, EOF or an error occurs. Without this change
4035 some truncated ASN1 structures will not produce an error.
4038 *) Disable Heimdal support, since it hasn't been fully implemented.
4039 Still give the possibility to force the use of Heimdal, but with
4040 warnings and a request that patches get sent to openssl-dev.
4043 *) Add the VC-CE target, introduce the WINCE sysname, and add
4044 INSTALL.WCE and appropriate conditionals to make it build.
4045 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4047 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4048 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4049 edit numbers of the version.
4050 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4052 *) Introduce safe string copy and catenation functions
4053 (BUF_strlcpy() and BUF_strlcat()).
4054 [Ben Laurie (CHATS) and Richard Levitte]
4056 *) Avoid using fixed-size buffers for one-line DNs.
4057 [Ben Laurie (CHATS)]
4059 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4060 resizing buffers containing secrets, and use where appropriate.
4061 [Ben Laurie (CHATS)]
4063 *) Avoid using fixed size buffers for configuration file location.
4064 [Ben Laurie (CHATS)]
4066 *) Avoid filename truncation for various CA files.
4067 [Ben Laurie (CHATS)]
4069 *) Use sizeof in preference to magic numbers.
4070 [Ben Laurie (CHATS)]
4072 *) Avoid filename truncation in cert requests.
4073 [Ben Laurie (CHATS)]
4075 *) Add assertions to check for (supposedly impossible) buffer
4077 [Ben Laurie (CHATS)]
4079 *) Don't cache truncated DNS entries in the local cache (this could
4080 potentially lead to a spoofing attack).
4081 [Ben Laurie (CHATS)]
4083 *) Fix various buffers to be large enough for hex/decimal
4084 representations in a platform independent manner.
4085 [Ben Laurie (CHATS)]
4087 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4088 resizing buffers containing secrets, and use where appropriate.
4089 [Ben Laurie (CHATS)]
4091 *) Add BIO_indent() to avoid much slightly worrying code to do
4093 [Ben Laurie (CHATS)]
4095 *) Convert sprintf()/BIO_puts() to BIO_printf().
4096 [Ben Laurie (CHATS)]
4098 *) buffer_gets() could terminate with the buffer only half
4100 [Ben Laurie (CHATS)]
4102 *) Add assertions to prevent user-supplied crypto functions from
4103 overflowing internal buffers by having large block sizes, etc.
4104 [Ben Laurie (CHATS)]
4106 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4108 [Ben Laurie (CHATS)]
4110 *) Eliminate unused copy of key in RC4.
4111 [Ben Laurie (CHATS)]
4113 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4114 [Ben Laurie (CHATS)]
4116 *) Fix off-by-one error in EGD path.
4117 [Ben Laurie (CHATS)]
4119 *) If RANDFILE path is too long, ignore instead of truncating.
4120 [Ben Laurie (CHATS)]
4122 *) Eliminate unused and incorrectly sized X.509 structure
4124 [Ben Laurie (CHATS)]
4126 *) Eliminate unused and dangerous function knumber().
4127 [Ben Laurie (CHATS)]
4129 *) Eliminate unused and dangerous structure, KSSL_ERR.
4130 [Ben Laurie (CHATS)]
4132 *) Protect against overlong session ID context length in an encoded
4133 session object. Since these are local, this does not appear to be
4135 [Ben Laurie (CHATS)]
4137 *) Change from security patch (see 0.9.6e below) that did not affect
4138 the 0.9.6 release series:
4140 Remote buffer overflow in SSL3 protocol - an attacker could
4141 supply an oversized master key in Kerberos-enabled versions.
4143 [Ben Laurie (CHATS)]
4145 *) Change the SSL kerb5 codes to match RFC 2712.
4148 *) Make -nameopt work fully for req and add -reqopt switch.
4149 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4151 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4152 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4154 *) Make sure tests can be performed even if the corresponding algorithms
4155 have been removed entirely. This was also the last step to make
4156 OpenSSL compilable with DJGPP under all reasonable conditions.
4157 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
4159 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
4160 to allow version independent disabling of normally unselected ciphers,
4161 which may be activated as a side-effect of selecting a single cipher.
4163 (E.g., cipher list string "RSA" enables ciphersuites that are left
4164 out of "ALL" because they do not provide symmetric encryption.
4165 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
4166 [Lutz Jaenicke, Bodo Moeller]
4168 *) Add appropriate support for separate platform-dependent build
4169 directories. The recommended way to make a platform-dependent
4170 build directory is the following (tested on Linux), maybe with
4173 # Place yourself outside of the OpenSSL source tree. In
4174 # this example, the environment variable OPENSSL_SOURCE
4175 # is assumed to contain the absolute OpenSSL source directory.
4176 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
4177 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
4178 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
4179 mkdir -p `dirname $F`
4180 ln -s $OPENSSL_SOURCE/$F $F
4183 To be absolutely sure not to disturb the source tree, a "make clean"
4184 is a good thing. If it isn't successfull, don't worry about it,
4185 it probably means the source directory is very clean.
4188 *) Make sure any ENGINE control commands make local copies of string
4189 pointers passed to them whenever necessary. Otherwise it is possible
4190 the caller may have overwritten (or deallocated) the original string
4191 data when a later ENGINE operation tries to use the stored values.
4192 [Götz Babin-Ebell <babinebell@trustcenter.de>]
4194 *) Improve diagnostics in file reading and command-line digests.
4195 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
4197 *) Add AES modes CFB and OFB to the object database. Correct an
4198 error in AES-CFB decryption.
4201 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
4202 allows existing EVP_CIPHER_CTX structures to be reused after
4203 calling EVP_*Final(). This behaviour is used by encryption
4204 BIOs and some applications. This has the side effect that
4205 applications must explicitly clean up cipher contexts with
4206 EVP_CIPHER_CTX_cleanup() or they will leak memory.
4209 *) Check the values of dna and dnb in bn_mul_recursive before calling
4210 bn_mul_comba (a non zero value means the a or b arrays do not contain
4211 n2 elements) and fallback to bn_mul_normal if either is not zero.
4214 *) Fix escaping of non-ASCII characters when using the -subj option
4215 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
4218 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
4219 form for "surname", serialNumber has no short form.
4220 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
4221 therefore remove "mail" short name for "internet 7".
4222 The OID for unique identifiers in X509 certificates is
4223 x500UniqueIdentifier, not uniqueIdentifier.
4224 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
4227 *) Add an "init" command to the ENGINE config module and auto initialize
4228 ENGINEs. Without any "init" command the ENGINE will be initialized
4229 after all ctrl commands have been executed on it. If init=1 the
4230 ENGINE is initailized at that point (ctrls before that point are run
4231 on the uninitialized ENGINE and after on the initialized one). If
4232 init=0 then the ENGINE will not be iniatialized at all.
4235 *) Fix the 'app_verify_callback' interface so that the user-defined
4236 argument is actually passed to the callback: In the
4237 SSL_CTX_set_cert_verify_callback() prototype, the callback
4238 declaration has been changed from
4241 int (*cb)(X509_STORE_CTX *,void *);
4242 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
4243 i=s->ctx->app_verify_callback(&ctx)
4244 has been changed into
4245 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
4247 To update applications using SSL_CTX_set_cert_verify_callback(),
4248 a dummy argument can be added to their callback functions.
4249 [D. K. Smetters <smetters@parc.xerox.com>]
4251 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
4252 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
4254 *) Add and OPENSSL_LOAD_CONF define which will cause
4255 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
4256 This allows older applications to transparently support certain
4257 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
4258 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
4259 load the config file and OPENSSL_add_all_algorithms_conf() which will
4260 always load it have also been added.
4263 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
4264 Adjust NIDs and EVP layer.
4265 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4267 *) Config modules support in openssl utility.
4269 Most commands now load modules from the config file,
4270 though in a few (such as version) this isn't done
4271 because it couldn't be used for anything.
4273 In the case of ca and req the config file used is
4274 the same as the utility itself: that is the -config
4275 command line option can be used to specify an
4279 *) Move default behaviour from OPENSSL_config(). If appname is NULL
4280 use "openssl_conf" if filename is NULL use default openssl config file.
4283 *) Add an argument to OPENSSL_config() to allow the use of an alternative
4284 config section name. Add a new flag to tolerate a missing config file
4285 and move code to CONF_modules_load_file().
4288 *) Support for crypto accelerator cards from Accelerated Encryption
4289 Processing, www.aep.ie. (Use engine 'aep')
4290 The support was copied from 0.9.6c [engine] and adapted/corrected
4291 to work with the new engine framework.
4292 [AEP Inc. and Richard Levitte]
4294 *) Support for SureWare crypto accelerator cards from Baltimore
4295 Technologies. (Use engine 'sureware')
4296 The support was copied from 0.9.6c [engine] and adapted
4297 to work with the new engine framework.
4300 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
4301 make the newer ENGINE framework commands for the CHIL engine work.
4302 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
4304 *) Make it possible to produce shared libraries on ReliantUNIX.
4305 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
4307 *) Add the configuration target debug-linux-ppro.
4308 Make 'openssl rsa' use the general key loading routines
4309 implemented in apps.c, and make those routines able to
4310 handle the key format FORMAT_NETSCAPE and the variant
4312 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4314 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4315 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4317 *) Add -keyform to rsautl, and document -engine.
4318 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
4320 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
4321 BIO_R_NO_SUCH_FILE error code rather than the generic
4322 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
4325 *) Add new functions
4327 ERR_peek_last_error_line
4328 ERR_peek_last_error_line_data.
4329 These are similar to
4332 ERR_peek_error_line_data,
4333 but report on the latest error recorded rather than the first one
4334 still in the error queue.
4335 [Ben Laurie, Bodo Moeller]
4337 *) default_algorithms option in ENGINE config module. This allows things
4339 default_algorithms = ALL
4340 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4343 *) Prelminary ENGINE config module.
4346 *) New experimental application configuration code.
4349 *) Change the AES code to follow the same name structure as all other
4350 symmetric ciphers, and behave the same way. Move everything to
4351 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4352 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4354 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4355 [Ben Laurie and Theo de Raadt]
4357 *) Add option to output public keys in req command.
4358 [Massimiliano Pala madwolf@openca.org]
4360 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4361 (up to about 10% better than before for P-192 and P-224).
4364 *) New functions/macros
4366 SSL_CTX_set_msg_callback(ctx, cb)
4367 SSL_CTX_set_msg_callback_arg(ctx, arg)
4368 SSL_set_msg_callback(ssl, cb)
4369 SSL_set_msg_callback_arg(ssl, arg)
4371 to request calling a callback function
4373 void cb(int write_p, int version, int content_type,
4374 const void *buf, size_t len, SSL *ssl, void *arg)
4376 whenever a protocol message has been completely received
4377 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4378 protocol version according to which the SSL library interprets
4379 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4380 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4381 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4382 specification (change_cipher_spec(20), alert(21), handshake(22)).
4383 'buf' and 'len' point to the actual message, 'ssl' to the
4384 SSL object, and 'arg' is the application-defined value set by
4385 SSL[_CTX]_set_msg_callback_arg().
4387 'openssl s_client' and 'openssl s_server' have new '-msg' options
4388 to enable a callback that displays all protocol messages.
4391 *) Change the shared library support so shared libraries are built as
4392 soon as the corresponding static library is finished, and thereby get
4393 openssl and the test programs linked against the shared library.
4394 This still only happens when the keyword "shard" has been given to
4395 the configuration scripts.
4397 NOTE: shared library support is still an experimental thing, and
4398 backward binary compatibility is still not guaranteed.
4399 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4401 *) Add support for Subject Information Access extension.
4402 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4404 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4405 additional bytes when new memory had to be allocated, not just
4406 when reusing an existing buffer.
4409 *) New command line and configuration option 'utf8' for the req command.
4410 This allows field values to be specified as UTF8 strings.
4413 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4414 runs for the former and machine-readable output for the latter.
4417 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4418 of the e-mail address in the DN (i.e., it will go into a certificate
4419 extension only). The new configuration file option 'email_in_dn = no'
4420 has the same effect.
4421 [Massimiliano Pala madwolf@openca.org]
4423 *) Change all functions with names starting with des_ to be starting
4424 with DES_ instead. Add wrappers that are compatible with libdes,
4425 but are named _ossl_old_des_*. Finally, add macros that map the
4426 des_* symbols to the corresponding _ossl_old_des_* if libdes
4427 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4428 desired, the des_* symbols will be mapped to DES_*, with one
4431 Since we provide two compatibility mappings, the user needs to
4432 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4433 compatibility is desired. The default (i.e., when that macro
4434 isn't defined) is OpenSSL 0.9.6c compatibility.
4436 There are also macros that enable and disable the support of old
4437 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4438 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4439 are defined, the default will apply: to support the old des routines.
4441 In either case, one must include openssl/des.h to get the correct
4442 definitions. Do not try to just include openssl/des_old.h, that
4445 NOTE: This is a major break of an old API into a new one. Software
4446 authors are encouraged to switch to the DES_ style functions. Some
4447 time in the future, des_old.h and the libdes compatibility functions
4448 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4449 default), and then completely removed.
4452 *) Test for certificates which contain unsupported critical extensions.
4453 If such a certificate is found during a verify operation it is
4454 rejected by default: this behaviour can be overridden by either
4455 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4456 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4457 X509_supported_extension() has also been added which returns 1 if a
4458 particular extension is supported.
4461 *) Modify the behaviour of EVP cipher functions in similar way to digests
4462 to retain compatibility with existing code.
4465 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4466 compatibility with existing code. In particular the 'ctx' parameter does
4467 not have to be to be initialized before the call to EVP_DigestInit() and
4468 it is tidied up after a call to EVP_DigestFinal(). New function
4469 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4470 EVP_MD_CTX_copy() changed to not require the destination to be
4471 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4472 requires the destination to be valid.
4474 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4475 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4478 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4479 so that complete 'Handshake' protocol structures are kept in memory
4480 instead of overwriting 'msg_type' and 'length' with 'body' data.
4483 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4484 [Massimo Santin via Richard Levitte]
4486 *) Major restructuring to the underlying ENGINE code. This includes
4487 reduction of linker bloat, separation of pure "ENGINE" manipulation
4488 (initialisation, etc) from functionality dealing with implementations
4489 of specific crypto iterfaces. This change also introduces integrated
4490 support for symmetric ciphers and digest implementations - so ENGINEs
4491 can now accelerate these by providing EVP_CIPHER and EVP_MD
4492 implementations of their own. This is detailed in crypto/engine/README
4493 as it couldn't be adequately described here. However, there are a few
4494 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4495 were changed in the original introduction of ENGINE code have now
4496 reverted back - the hooking from this code to ENGINE is now a good
4497 deal more passive and at run-time, operations deal directly with
4498 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4499 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4500 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4501 they were not being used by the framework as there is no concept of a
4502 BIGNUM_METHOD and they could not be generalised to the new
4503 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4504 ENGINE_cpy() has been removed as it cannot be consistently defined in
4508 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4511 *) Change mkdef.pl to sort symbols that get the same entry number,
4512 and make sure the automatically generated functions ERR_load_*
4513 become part of libeay.num as well.
4516 *) New function SSL_renegotiate_pending(). This returns true once
4517 renegotiation has been requested (either SSL_renegotiate() call
4518 or HelloRequest/ClientHello receveived from the peer) and becomes
4519 false once a handshake has been completed.
4520 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4521 sends a HelloRequest, but does not ensure that a handshake takes
4522 place. SSL_renegotiate_pending() is useful for checking if the
4523 client has followed the request.)
4526 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4527 By default, clients may request session resumption even during
4528 renegotiation (if session ID contexts permit); with this option,
4529 session resumption is possible only in the first handshake.
4531 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4532 more bits available for options that should not be part of
4533 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4536 *) Add some demos for certificate and certificate request creation.
4539 *) Make maximum certificate chain size accepted from the peer application
4540 settable (SSL*_get/set_max_cert_list()), as proposed by
4541 "Douglas E. Engert" <deengert@anl.gov>.
4544 *) Add support for shared libraries for Unixware-7
4545 (Boyd Lynn Gerber <gerberb@zenez.com>).
4548 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4549 be done prior to destruction. Use this to unload error strings from
4550 ENGINEs that load their own error strings. NB: This adds two new API
4551 functions to "get" and "set" this destroy handler in an ENGINE.
4554 *) Alter all existing ENGINE implementations (except "openssl" and
4555 "openbsd") to dynamically instantiate their own error strings. This
4556 makes them more flexible to be built both as statically-linked ENGINEs
4557 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4558 Also, add stub code to each that makes building them as self-contained
4559 shared-libraries easier (see README.ENGINE).
4562 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4563 implementations into applications that are completely implemented in
4564 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4565 commands that can be used to configure what shared-library to load and
4566 to control aspects of the way it is handled. Also, made an update to
4567 the README.ENGINE file that brings its information up-to-date and
4568 provides some information and instructions on the "dynamic" ENGINE
4569 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4572 *) Make it possible to unload ranges of ERR strings with a new
4573 "ERR_unload_strings" function.
4576 *) Add a copy() function to EVP_MD.
4579 *) Make EVP_MD routines take a context pointer instead of just the
4580 md_data void pointer.
4583 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4584 that the digest can only process a single chunk of data
4585 (typically because it is provided by a piece of
4586 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4587 is only going to provide a single chunk of data, and hence the
4588 framework needn't accumulate the data for oneshot drivers.
4591 *) As with "ERR", make it possible to replace the underlying "ex_data"
4592 functions. This change also alters the storage and management of global
4593 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4594 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4595 index counters. The API functions that use this state have been changed
4596 to take a "class_index" rather than pointers to the class's local STACK
4597 and counter, and there is now an API function to dynamically create new
4598 classes. This centralisation allows us to (a) plug a lot of the
4599 thread-safety problems that existed, and (b) makes it possible to clean
4600 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4601 such data would previously have always leaked in application code and
4602 workarounds were in place to make the memory debugging turn a blind eye
4603 to it. Application code that doesn't use this new function will still
4604 leak as before, but their memory debugging output will announce it now
4605 rather than letting it slide.
4607 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4608 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4609 has a return value to indicate success or failure.
4612 *) Make it possible to replace the underlying "ERR" functions such that the
4613 global state (2 LHASH tables and 2 locks) is only used by the "default"
4614 implementation. This change also adds two functions to "get" and "set"
4615 the implementation prior to it being automatically set the first time
4616 any other ERR function takes place. Ie. an application can call "get",
4617 pass the return value to a module it has just loaded, and that module
4618 can call its own "set" function using that value. This means the
4619 module's "ERR" operations will use (and modify) the error state in the
4620 application and not in its own statically linked copy of OpenSSL code.
4623 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4624 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4625 the operation, and provides a more encapsulated way for external code
4626 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4627 to use these functions rather than manually incrementing the counts.
4629 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4632 *) Add EVP test program.
4635 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4638 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4639 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4640 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4641 These allow a CRL to be built without having to access X509_CRL fields
4642 directly. Modify 'ca' application to use new functions.
4645 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4646 bug workarounds. Rollback attack detection is a security feature.
4647 The problem will only arise on OpenSSL servers when TLSv1 is not
4648 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4649 Software authors not wanting to support TLSv1 will have special reasons
4650 for their choice and can explicitly enable this option.
4651 [Bodo Moeller, Lutz Jaenicke]
4653 *) Rationalise EVP so it can be extended: don't include a union of
4654 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4655 (similar to those existing for EVP_CIPHER_CTX).
4660 EVP_MD_CTX_init(&md); /* new function call */
4661 EVP_DigestInit(&md, EVP_sha1());
4662 EVP_DigestUpdate(&md, in, len);
4663 EVP_DigestFinal(&md, out, NULL);
4664 EVP_MD_CTX_cleanup(&md); /* new function call */
4668 *) Make DES key schedule conform to the usual scheme, as well as
4669 correcting its structure. This means that calls to DES functions
4670 now have to pass a pointer to a des_key_schedule instead of a
4671 plain des_key_schedule (which was actually always a pointer
4674 des_key_schedule ks;
4676 des_set_key_checked(..., &ks);
4677 des_ncbc_encrypt(..., &ks, ...);
4679 (Note that a later change renames 'des_...' into 'DES_...'.)
4682 *) Initial reduction of linker bloat: the use of some functions, such as
4683 PEM causes large amounts of unused functions to be linked in due to
4684 poor organisation. For example pem_all.c contains every PEM function
4685 which has a knock on effect of linking in large amounts of (unused)
4686 ASN1 code. Grouping together similar functions and splitting unrelated
4687 functions prevents this.
4690 *) Cleanup of EVP macros.
4693 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4694 correct _ecb suffix.
4697 *) Add initial OCSP responder support to ocsp application. The
4698 revocation information is handled using the text based index
4699 use by the ca application. The responder can either handle
4700 requests generated internally, supplied in files (for example
4701 via a CGI script) or using an internal minimal server.
4704 *) Add configuration choices to get zlib compression for TLS.
4707 *) Changes to Kerberos SSL for RFC 2712 compliance:
4708 1. Implemented real KerberosWrapper, instead of just using
4709 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4710 2. Implemented optional authenticator field of KerberosWrapper.
4712 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4713 and authenticator structs; see crypto/krb5/.
4715 Generalized Kerberos calls to support multiple Kerberos libraries.
4716 [Vern Staats <staatsvr@asc.hpc.mil>,
4717 Jeffrey Altman <jaltman@columbia.edu>
4718 via Richard Levitte]
4720 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4721 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4722 values for each of the key sizes rather than having just
4723 parameters (and 'speed' generating keys each time).
4726 *) Speed up EVP routines.
4729 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4730 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4731 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4732 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4734 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4735 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4736 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4739 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4741 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4744 *) Added the OS2-EMX target.
4745 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4747 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4748 to support NCONF routines in extension code. New function CONF_set_nconf()
4749 to allow functions which take an NCONF to also handle the old LHASH
4750 structure: this means that the old CONF compatible routines can be
4751 retained (in particular wrt extensions) without having to duplicate the
4752 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4755 *) Enhance the general user interface with mechanisms for inner control
4756 and with possibilities to have yes/no kind of prompts.
4759 *) Change all calls to low level digest routines in the library and
4760 applications to use EVP. Add missing calls to HMAC_cleanup() and
4761 don't assume HMAC_CTX can be copied using memcpy().
4762 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4764 *) Add the possibility to control engines through control names but with
4765 arbitrary arguments instead of just a string.
4766 Change the key loaders to take a UI_METHOD instead of a callback
4767 function pointer. NOTE: this breaks binary compatibility with earlier
4768 versions of OpenSSL [engine].
4769 Adapt the nCipher code for these new conditions and add a card insertion
4773 *) Enhance the general user interface with mechanisms to better support
4774 dialog box interfaces, application-defined prompts, the possibility
4775 to use defaults (for example default passwords from somewhere else)
4776 and interrupts/cancellations.
4779 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4780 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4783 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4784 tidy up some unnecessarily weird code in 'sk_new()').
4785 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4787 *) Change the key loading routines for ENGINEs to use the same kind
4788 callback (pem_password_cb) as all other routines that need this
4792 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4793 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4794 than this minimum value is recommended.
4797 *) New random seeder for OpenVMS, using the system process statistics
4798 that are easily reachable.
4801 *) Windows apparently can't transparently handle global
4802 variables defined in DLLs. Initialisations such as:
4804 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4806 wont compile. This is used by the any applications that need to
4807 declare their own ASN1 modules. This was fixed by adding the option
4808 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4809 needed for static libraries under Win32.
4812 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4813 setting of purpose and trust fields. New X509_STORE trust and
4814 purpose functions and tidy up setting in other SSL functions.
4817 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4818 structure. These are inherited by X509_STORE_CTX when it is
4819 initialised. This allows various defaults to be set in the
4820 X509_STORE structure (such as flags for CRL checking and custom
4821 purpose or trust settings) for functions which only use X509_STORE_CTX
4822 internally such as S/MIME.
4824 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4825 trust settings if they are not set in X509_STORE. This allows X509_STORE
4826 purposes and trust (in S/MIME for example) to override any set by default.
4828 Add command line options for CRL checking to smime, s_client and s_server
4832 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4833 are set then the CRL is looked up in the X509_STORE structure and
4834 its validity and signature checked, then if the certificate is found
4835 in the CRL the verify fails with a revoked error.
4837 Various new CRL related callbacks added to X509_STORE_CTX structure.
4839 Command line options added to 'verify' application to support this.
4841 This needs some additional work, such as being able to handle multiple
4842 CRLs with different times, extension based lookup (rather than just
4843 by subject name) and ultimately more complete V2 CRL extension
4847 *) Add a general user interface API (crypto/ui/). This is designed
4848 to replace things like des_read_password and friends (backward
4849 compatibility functions using this new API are provided).
4850 The purpose is to remove prompting functions from the DES code
4851 section as well as provide for prompting through dialog boxes in
4852 a window system and the like.
4855 *) Add "ex_data" support to ENGINE so implementations can add state at a
4856 per-structure level rather than having to store it globally.
4859 *) Make it possible for ENGINE structures to be copied when retrieved by
4860 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4861 This causes the "original" ENGINE structure to act like a template,
4862 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4863 operational state can be localised to each ENGINE structure, despite the
4864 fact they all share the same "methods". New ENGINE structures returned in
4865 this case have no functional references and the return value is the single
4866 structural reference. This matches the single structural reference returned
4867 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4871 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4872 needs to match any other type at all we need to manually clear the
4876 *) Changes to the "openssl engine" utility to include;
4877 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4878 about an ENGINE's available control commands.
4879 - executing control commands from command line arguments using the
4880 '-pre' and '-post' switches. '-post' is only used if '-t' is
4881 specified and the ENGINE is successfully initialised. The syntax for
4882 the individual commands are colon-separated, for example;
4883 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4886 *) New dynamic control command support for ENGINEs. ENGINEs can now
4887 declare their own commands (numbers), names (strings), descriptions,
4888 and input types for run-time discovery by calling applications. A
4889 subset of these commands are implicitly classed as "executable"
4890 depending on their input type, and only these can be invoked through
4891 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4892 can be based on user input, config files, etc). The distinction is
4893 that "executable" commands cannot return anything other than a boolean
4894 result and can only support numeric or string input, whereas some
4895 discoverable commands may only be for direct use through
4896 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4897 pointers, or other custom uses. The "executable" commands are to
4898 support parameterisations of ENGINE behaviour that can be
4899 unambiguously defined by ENGINEs and used consistently across any
4900 OpenSSL-based application. Commands have been added to all the
4901 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4902 control over shared-library paths without source code alterations.
4905 *) Changed all ENGINE implementations to dynamically allocate their
4906 ENGINEs rather than declaring them statically. Apart from this being
4907 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
4908 this also allows the implementations to compile without using the
4909 internal engine_int.h header.
4912 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
4913 'const' value. Any code that should be able to modify a RAND_METHOD
4914 should already have non-const pointers to it (ie. they should only
4915 modify their own ones).
4918 *) Made a variety of little tweaks to the ENGINE code.
4919 - "atalla" and "ubsec" string definitions were moved from header files
4920 to C code. "nuron" string definitions were placed in variables
4921 rather than hard-coded - allowing parameterisation of these values
4922 later on via ctrl() commands.
4923 - Removed unused "#if 0"'d code.
4924 - Fixed engine list iteration code so it uses ENGINE_free() to release
4925 structural references.
4926 - Constified the RAND_METHOD element of ENGINE structures.
4927 - Constified various get/set functions as appropriate and added
4928 missing functions (including a catch-all ENGINE_cpy that duplicates
4929 all ENGINE values onto a new ENGINE except reference counts/state).
4930 - Removed NULL parameter checks in get/set functions. Setting a method
4931 or function to NULL is a way of cancelling out a previously set
4932 value. Passing a NULL ENGINE parameter is just plain stupid anyway
4933 and doesn't justify the extra error symbols and code.
4934 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
4935 flags from engine_int.h to engine.h.
4936 - Changed prototypes for ENGINE handler functions (init(), finish(),
4937 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
4940 *) Implement binary inversion algorithm for BN_mod_inverse in addition
4941 to the algorithm using long division. The binary algorithm can be
4942 used only if the modulus is odd. On 32-bit systems, it is faster
4943 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
4944 roughly 5-15% for 256-bit moduli), so we use it only for moduli
4945 up to 450 bits. In 64-bit environments, the binary algorithm
4946 appears to be advantageous for much longer moduli; here we use it
4947 for moduli up to 2048 bits.
4950 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
4951 could not support the combine flag in choice fields.
4954 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
4955 extensions from a certificate request to the certificate.
4958 *) Allow multiple 'certopt' and 'nameopt' options to be separated
4959 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
4960 file: this allows the display of the certificate about to be
4961 signed to be customised, to allow certain fields to be included
4962 or excluded and extension details. The old system didn't display
4963 multicharacter strings properly, omitted fields not in the policy
4964 and couldn't display additional details such as extensions.
4967 *) Function EC_POINTs_mul for multiple scalar multiplication
4968 of an arbitrary number of elliptic curve points
4969 \sum scalars[i]*points[i],
4970 optionally including the generator defined for the EC_GROUP:
4971 scalar*generator + \sum scalars[i]*points[i].
4973 EC_POINT_mul is a simple wrapper function for the typical case
4974 that the point list has just one item (besides the optional
4978 *) First EC_METHODs for curves over GF(p):
4980 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
4981 operations and provides various method functions that can also
4982 operate with faster implementations of modular arithmetic.
4984 EC_GFp_mont_method() reuses most functions that are part of
4985 EC_GFp_simple_method, but uses Montgomery arithmetic.
4987 [Bodo Moeller; point addition and point doubling
4988 implementation directly derived from source code provided by
4989 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
4991 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
4992 crypto/ec/ec_lib.c):
4994 Curves are EC_GROUP objects (with an optional group generator)
4995 based on EC_METHODs that are built into the library.
4997 Points are EC_POINT objects based on EC_GROUP objects.
4999 Most of the framework would be able to handle curves over arbitrary
5000 finite fields, but as there are no obvious types for fields other
5001 than GF(p), some functions are limited to that for now.
5004 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
5005 that the file contains a complete HTTP response.
5008 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
5009 change the def and num file printf format specifier from "%-40sXXX"
5010 to "%-39s XXX". The latter will always guarantee a space after the
5011 field while the former will cause them to run together if the field
5012 is 40 of more characters long.
5015 *) Constify the cipher and digest 'method' functions and structures
5016 and modify related functions to take constant EVP_MD and EVP_CIPHER
5020 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
5021 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
5024 *) Modify EVP_Digest*() routines so they now return values. Although the
5025 internal software routines can never fail additional hardware versions
5029 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
5031 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
5032 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
5035 ERR_R_NESTED_ASN1_ERROR
5037 ERR_R_MISSING_ASN1_EOS
5038 were 4 .. 9, conflicting with
5039 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5041 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5042 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5044 Add new error code 'ERR_R_INTERNAL_ERROR'.
5047 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5051 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5052 sets the subject name for a new request or supersedes the
5053 subject name in a given request. Formats that can be parsed are
5054 'CN=Some Name, OU=myOU, C=IT'
5056 'CN=Some Name/OU=myOU/C=IT'.
5058 Add options '-batch' and '-verbose' to 'openssl req'.
5059 [Massimiliano Pala <madwolf@hackmasters.net>]
5061 *) Introduce the possibility to access global variables through
5062 functions on platform were that's the best way to handle exporting
5063 global variables in shared libraries. To enable this functionality,
5064 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5065 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5066 is normally done by Configure or something similar).
5068 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5069 in the source file (foo.c) like this:
5071 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5072 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5074 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5075 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5077 OPENSSL_DECLARE_GLOBAL(int,foo);
5078 #define foo OPENSSL_GLOBAL_REF(foo)
5079 OPENSSL_DECLARE_GLOBAL(double,bar);
5080 #define bar OPENSSL_GLOBAL_REF(bar)
5082 The #defines are very important, and therefore so is including the
5083 header file everywhere where the defined globals are used.
5085 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5086 of ASN.1 items, but that structure is a bit different.
5088 The largest change is in util/mkdef.pl which has been enhanced with
5089 better and easier to understand logic to choose which symbols should
5090 go into the Windows .def files as well as a number of fixes and code
5091 cleanup (among others, algorithm keywords are now sorted
5092 lexicographically to avoid constant rewrites).
5095 *) In BN_div() keep a copy of the sign of 'num' before writing the
5096 result to 'rm' because if rm==num the value will be overwritten
5097 and produce the wrong result if 'num' is negative: this caused
5098 problems with BN_mod() and BN_nnmod().
5101 *) Function OCSP_request_verify(). This checks the signature on an
5102 OCSP request and verifies the signer certificate. The signer
5103 certificate is just checked for a generic purpose and OCSP request
5107 *) Add OCSP_check_validity() function to check the validity of OCSP
5108 responses. OCSP responses are prepared in real time and may only
5109 be a few seconds old. Simply checking that the current time lies
5110 between thisUpdate and nextUpdate max reject otherwise valid responses
5111 caused by either OCSP responder or client clock inaccuracy. Instead
5112 we allow thisUpdate and nextUpdate to fall within a certain period of
5113 the current time. The age of the response can also optionally be
5114 checked. Two new options -validity_period and -status_age added to
5118 *) If signature or public key algorithm is unrecognized print out its
5119 OID rather that just UNKNOWN.
5122 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5123 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5124 ID to be generated from the issuer certificate alone which can then be
5125 passed to OCSP_id_issuer_cmp().
5128 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5129 ASN1 modules to export functions returning ASN1_ITEM pointers
5130 instead of the ASN1_ITEM structures themselves. This adds several
5131 new macros which allow the underlying ASN1 function/structure to
5132 be accessed transparently. As a result code should not use ASN1_ITEM
5133 references directly (such as &X509_it) but instead use the relevant
5134 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5135 use of the new ASN1 code on platforms where exporting structures
5136 is problematical (for example in shared libraries) but exporting
5137 functions returning pointers to structures is not.
5140 *) Add support for overriding the generation of SSL/TLS session IDs.
5141 These callbacks can be registered either in an SSL_CTX or per SSL.
5142 The purpose of this is to allow applications to control, if they wish,
5143 the arbitrary values chosen for use as session IDs, particularly as it
5144 can be useful for session caching in multiple-server environments. A
5145 command-line switch for testing this (and any client code that wishes
5146 to use such a feature) has been added to "s_server".
5147 [Geoff Thorpe, Lutz Jaenicke]
5149 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5150 of the form '#if defined(...) || defined(...) || ...' and
5151 '#if !defined(...) && !defined(...) && ...'. This also avoids
5152 the growing number of special cases it was previously handling.
5155 *) Make all configuration macros available for application by making
5156 sure they are available in opensslconf.h, by giving them names starting
5157 with "OPENSSL_" to avoid conflicts with other packages and by making
5158 sure e_os2.h will cover all platform-specific cases together with
5160 Additionally, it is now possible to define configuration/platform-
5161 specific names (called "system identities"). In the C code, these
5162 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
5163 macro with the name beginning with "OPENSSL_SYS_", which is determined
5164 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
5168 *) New option -set_serial to 'req' and 'x509' this allows the serial
5169 number to use to be specified on the command line. Previously self
5170 signed certificates were hard coded with serial number 0 and the
5171 CA options of 'x509' had to use a serial number in a file which was
5175 *) New options to 'ca' utility to support V2 CRL entry extensions.
5176 Currently CRL reason, invalidity date and hold instruction are
5177 supported. Add new CRL extensions to V3 code and some new objects.
5180 *) New function EVP_CIPHER_CTX_set_padding() this is used to
5181 disable standard block padding (aka PKCS#5 padding) in the EVP
5182 API, which was previously mandatory. This means that the data is
5183 not padded in any way and so the total length much be a multiple
5184 of the block size, otherwise an error occurs.
5187 *) Initial (incomplete) OCSP SSL support.
5190 *) New function OCSP_parse_url(). This splits up a URL into its host,
5191 port and path components: primarily to parse OCSP URLs. New -url
5192 option to ocsp utility.
5195 *) New nonce behavior. The return value of OCSP_check_nonce() now
5196 reflects the various checks performed. Applications can decide
5197 whether to tolerate certain situations such as an absent nonce
5198 in a response when one was present in a request: the ocsp application
5199 just prints out a warning. New function OCSP_add1_basic_nonce()
5200 this is to allow responders to include a nonce in a response even if
5201 the request is nonce-less.
5204 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
5205 skipped when using openssl x509 multiple times on a single input file,
5206 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
5209 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
5210 set string type: to handle setting ASN1_TIME structures. Fix ca
5211 utility to correctly initialize revocation date of CRLs.
5214 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
5215 the clients preferred ciphersuites and rather use its own preferences.
5216 Should help to work around M$ SGC (Server Gated Cryptography) bug in
5217 Internet Explorer by ensuring unchanged hash method during stepup.
5218 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
5221 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
5222 to aes and add a new 'exist' option to print out symbols that don't
5226 *) Additional options to ocsp utility to allow flags to be set and
5227 additional certificates supplied.
5230 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
5231 OCSP client a number of certificate to only verify the response
5235 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
5236 handle the new API. Currently only ECB, CBC modes supported. Add new
5239 Add TLS AES ciphersuites as described in RFC3268, "Advanced
5240 Encryption Standard (AES) Ciphersuites for Transport Layer
5241 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
5242 not enabled by default and were not part of the "ALL" ciphersuite
5243 alias because they were not yet official; they could be
5244 explicitly requested by specifying the "AESdraft" ciphersuite
5245 group alias. In the final release of OpenSSL 0.9.7, the group
5246 alias is called "AES" and is part of "ALL".)
5247 [Ben Laurie, Steve Henson, Bodo Moeller]
5249 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
5250 request to response.
5253 *) Functions for OCSP responders. OCSP_request_onereq_count(),
5254 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
5255 extract information from a certificate request. OCSP_response_create()
5256 creates a response and optionally adds a basic response structure.
5257 OCSP_basic_add1_status() adds a complete single response to a basic
5258 response and returns the OCSP_SINGLERESP structure just added (to allow
5259 extensions to be included for example). OCSP_basic_add1_cert() adds a
5260 certificate to a basic response and OCSP_basic_sign() signs a basic
5261 response with various flags. New helper functions ASN1_TIME_check()
5262 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
5263 (converts ASN1_TIME to GeneralizedTime).
5266 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
5267 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
5268 structure from a certificate. X509_pubkey_digest() digests the public_key
5269 contents: this is used in various key identifiers.
5272 *) Make sk_sort() tolerate a NULL argument.
5273 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
5275 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
5276 passed by the function are trusted implicitly. If any of them signed the
5277 response then it is assumed to be valid and is not verified.
5280 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
5281 to data. This was previously part of the PKCS7 ASN1 code. This
5282 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
5283 [Steve Henson, reported by Kenneth R. Robinette
5284 <support@securenetterm.com>]
5286 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
5287 routines: without these tracing memory leaks is very painful.
5288 Fix leaks in PKCS12 and PKCS7 routines.
5291 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
5292 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
5293 effectively meant GeneralizedTime would never be used. Now it
5294 is initialised to -1 but X509_time_adj() now has to check the value
5295 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
5296 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
5297 [Steve Henson, reported by Kenneth R. Robinette
5298 <support@securenetterm.com>]
5300 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
5301 result in a zero length in the ASN1_INTEGER structure which was
5302 not consistent with the structure when d2i_ASN1_INTEGER() was used
5303 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
5304 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
5305 where it did not print out a minus for negative ASN1_INTEGER.
5308 *) Add summary printout to ocsp utility. The various functions which
5309 convert status values to strings have been renamed to:
5310 OCSP_response_status_str(), OCSP_cert_status_str() and
5311 OCSP_crl_reason_str() and are no longer static. New options
5312 to verify nonce values and to disable verification. OCSP response
5313 printout format cleaned up.
5316 *) Add additional OCSP certificate checks. These are those specified
5317 in RFC2560. This consists of two separate checks: the CA of the
5318 certificate being checked must either be the OCSP signer certificate
5319 or the issuer of the OCSP signer certificate. In the latter case the
5320 OCSP signer certificate must contain the OCSP signing extended key
5321 usage. This check is performed by attempting to match the OCSP
5322 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
5323 in the OCSP_CERTID structures of the response.
5326 *) Initial OCSP certificate verification added to OCSP_basic_verify()
5327 and related routines. This uses the standard OpenSSL certificate
5328 verify routines to perform initial checks (just CA validity) and
5329 to obtain the certificate chain. Then additional checks will be
5330 performed on the chain. Currently the root CA is checked to see
5331 if it is explicitly trusted for OCSP signing. This is used to set
5332 a root CA as a global signing root: that is any certificate that
5333 chains to that CA is an acceptable OCSP signing certificate.
5336 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5337 extensions from a separate configuration file.
5338 As when reading extensions from the main configuration file,
5339 the '-extensions ...' option may be used for specifying the
5341 [Massimiliano Pala <madwolf@comune.modena.it>]
5343 *) New OCSP utility. Allows OCSP requests to be generated or
5344 read. The request can be sent to a responder and the output
5345 parsed, outputed or printed in text form. Not complete yet:
5346 still needs to check the OCSP response validity.
5349 *) New subcommands for 'openssl ca':
5350 'openssl ca -status <serial>' prints the status of the cert with
5351 the given serial number (according to the index file).
5352 'openssl ca -updatedb' updates the expiry status of certificates
5354 [Massimiliano Pala <madwolf@comune.modena.it>]
5356 *) New '-newreq-nodes' command option to CA.pl. This is like
5357 '-newreq', but calls 'openssl req' with the '-nodes' option
5358 so that the resulting key is not encrypted.
5359 [Damien Miller <djm@mindrot.org>]
5361 *) New configuration for the GNU Hurd.
5362 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5364 *) Initial code to implement OCSP basic response verify. This
5365 is currently incomplete. Currently just finds the signer's
5366 certificate and verifies the signature on the response.
5369 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5370 value of OPENSSLDIR. This is available via the new '-d' option
5371 to 'openssl version', and is also included in 'openssl version -a'.
5374 *) Allowing defining memory allocation callbacks that will be given
5375 file name and line number information in additional arguments
5376 (a const char* and an int). The basic functionality remains, as
5377 well as the original possibility to just replace malloc(),
5378 realloc() and free() by functions that do not know about these
5379 additional arguments. To register and find out the current
5380 settings for extended allocation functions, the following
5381 functions are provided:
5383 CRYPTO_set_mem_ex_functions
5384 CRYPTO_set_locked_mem_ex_functions
5385 CRYPTO_get_mem_ex_functions
5386 CRYPTO_get_locked_mem_ex_functions
5388 These work the same way as CRYPTO_set_mem_functions and friends.
5389 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5390 extended allocation function is enabled.
5391 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5392 a conventional allocation function is enabled.
5393 [Richard Levitte, Bodo Moeller]
5395 *) Finish off removing the remaining LHASH function pointer casts.
5396 There should no longer be any prototype-casting required when using
5397 the LHASH abstraction, and any casts that remain are "bugs". See
5398 the callback types and macros at the head of lhash.h for details
5399 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5402 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5403 If /dev/[u]random devices are not available or do not return enough
5404 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5406 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5407 /etc/entropy will be queried once each in this sequence, quering stops
5408 when enough entropy was collected without querying more sockets.
5411 *) Change the Unix RAND_poll() variant to be able to poll several
5412 random devices, as specified by DEVRANDOM, until a sufficient amount
5413 of data has been collected. We spend at most 10 ms on each file
5414 (select timeout) and read in non-blocking mode. DEVRANDOM now
5415 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5416 (previously it was just the string "/dev/urandom"), so on typical
5417 platforms the 10 ms delay will never occur.
5418 Also separate out the Unix variant to its own file, rand_unix.c.
5419 For VMS, there's a currently-empty rand_vms.c.
5422 *) Move OCSP client related routines to ocsp_cl.c. These
5423 provide utility functions which an application needing
5424 to issue a request to an OCSP responder and analyse the
5425 response will typically need: as opposed to those which an
5426 OCSP responder itself would need which will be added later.
5428 OCSP_request_sign() signs an OCSP request with an API similar
5429 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5430 response. OCSP_response_get1_basic() extracts basic response
5431 from response. OCSP_resp_find_status(): finds and extracts status
5432 information from an OCSP_CERTID structure (which will be created
5433 when the request structure is built). These are built from lower
5434 level functions which work on OCSP_SINGLERESP structures but
5435 wont normally be used unless the application wishes to examine
5436 extensions in the OCSP response for example.
5438 Replace nonce routines with a pair of functions.
5439 OCSP_request_add1_nonce() adds a nonce value and optionally
5440 generates a random value. OCSP_check_nonce() checks the
5441 validity of the nonce in an OCSP response.
5444 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5445 This doesn't copy the supplied OCSP_CERTID and avoids the
5446 need to free up the newly created id. Change return type
5447 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5448 This can then be used to add extensions to the request.
5449 Deleted OCSP_request_new(), since most of its functionality
5450 is now in OCSP_REQUEST_new() (and the case insensitive name
5451 clash) apart from the ability to set the request name which
5452 will be added elsewhere.
5455 *) Update OCSP API. Remove obsolete extensions argument from
5456 various functions. Extensions are now handled using the new
5457 OCSP extension code. New simple OCSP HTTP function which
5458 can be used to send requests and parse the response.
5461 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5462 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5463 uses the special reorder version of SET OF to sort the attributes
5464 and reorder them to match the encoded order. This resolves a long
5465 standing problem: a verify on a PKCS7 structure just after signing
5466 it used to fail because the attribute order did not match the
5467 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5468 it uses the received order. This is necessary to tolerate some broken
5469 software that does not order SET OF. This is handled by encoding
5470 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5471 to produce the required SET OF.
5474 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5475 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5476 files to get correct declarations of the ASN.1 item variables.
5479 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5480 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5481 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5482 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5483 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5484 ASN1_ITEM and no wrapper functions.
5487 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5488 replace the old function pointer based I/O routines. Change most of
5489 the *_d2i_bio() and *_d2i_fp() functions to use these.
5492 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5493 lines, recognice more "algorithms" that can be deselected, and make
5494 it complain about algorithm deselection that isn't recognised.
5497 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5498 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5499 to use new functions. Add NO_ASN1_OLD which can be set to remove
5500 some old style ASN1 functions: this can be used to determine if old
5501 code will still work when these eventually go away.
5504 *) New extension functions for OCSP structures, these follow the
5505 same conventions as certificates and CRLs.
5508 *) New function X509V3_add1_i2d(). This automatically encodes and
5509 adds an extension. Its behaviour can be customised with various
5510 flags to append, replace or delete. Various wrappers added for
5511 certifcates and CRLs.
5514 *) Fix to avoid calling the underlying ASN1 print routine when
5515 an extension cannot be parsed. Correct a typo in the
5516 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5519 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5520 entries for variables.
5523 *) Add functionality to apps/openssl.c for detecting locking
5524 problems: As the program is single-threaded, all we have
5525 to do is register a locking callback using an array for
5526 storing which locks are currently held by the program.
5529 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5530 SSL_get_ex_data_X509_STORE_idx(), which is used in
5531 ssl_verify_cert_chain() and thus can be called at any time
5532 during TLS/SSL handshakes so that thread-safety is essential.
5533 Unfortunately, the ex_data design is not at all suited
5534 for multi-threaded use, so it probably should be abolished.
5537 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5538 [Broadcom, tweaked and integrated by Geoff Thorpe]
5540 *) Move common extension printing code to new function
5541 X509V3_print_extensions(). Reorganise OCSP print routines and
5542 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5545 *) New function X509_signature_print() to remove duplication in some
5549 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5550 set (this was treated exactly the same as SET OF previously). This
5551 is used to reorder the STACK representing the structure to match the
5552 encoding. This will be used to get round a problem where a PKCS7
5553 structure which was signed could not be verified because the STACK
5554 order did not reflect the encoded order.
5557 *) Reimplement the OCSP ASN1 module using the new code.
5560 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5561 for its ASN1 operations. The old style function pointers still exist
5562 for now but they will eventually go away.
5565 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5566 completely replaces the old ASN1 functionality with a table driven
5567 encoder and decoder which interprets an ASN1_ITEM structure describing
5568 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5569 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5570 has also been converted to the new form.
5573 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5574 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5575 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5576 for negative moduli.
5579 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5580 of not touching the result's sign bit.
5583 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5587 *) Changed the LHASH code to use prototypes for callbacks, and created
5588 macros to declare and implement thin (optionally static) functions
5589 that provide type-safety and avoid function pointer casting for the
5590 type-specific callbacks.
5593 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5595 [Veers Staats <staatsvr@asc.hpc.mil>,
5596 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5598 *) Reformat the FAQ so the different questions and answers can be divided
5599 in sections depending on the subject.
5602 *) Have the zlib compression code load ZLIB.DLL dynamically under
5606 *) New function BN_mod_sqrt for computing square roots modulo a prime
5607 (using the probabilistic Tonelli-Shanks algorithm unless
5608 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5609 be handled deterministically).
5610 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5612 *) Make BN_mod_inverse faster by explicitly handling small quotients
5613 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5614 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5617 *) New function BN_kronecker.
5620 *) Fix BN_gcd so that it works on negative inputs; the result is
5621 positive unless both parameters are zero.
5622 Previously something reasonably close to an infinite loop was
5623 possible because numbers could be growing instead of shrinking
5624 in the implementation of Euclid's algorithm.
5627 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5628 sign of the number in question.
5630 Fix BN_is_word(a,w) to work correctly for w == 0.
5632 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5633 because its test if the absolute value of 'a' equals 'w'.
5634 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5635 it exists mostly for use in the implementations of BN_is_zero(),
5636 BN_is_one(), and BN_is_word().
5639 *) New function BN_swap.
5642 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5643 the exponentiation functions are more likely to produce reasonable
5644 results on negative inputs.
5647 *) Change BN_mod_mul so that the result is always non-negative.
5648 Previously, it could be negative if one of the factors was negative;
5649 I don't think anyone really wanted that behaviour.
5652 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5653 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5654 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5655 and add new functions:
5664 BN_mod_lshift1_quick
5668 These functions always generate non-negative results.
5670 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5671 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5673 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5674 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5675 be reduced modulo m.
5676 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5679 The following entry accidentily appeared in the CHANGES file
5680 distributed with OpenSSL 0.9.7. The modifications described in
5681 it do *not* apply to OpenSSL 0.9.7.
5683 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5684 was actually never needed) and in BN_mul(). The removal in BN_mul()
5685 required a small change in bn_mul_part_recursive() and the addition
5686 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5687 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5688 bn_sub_words() and bn_add_words() except they take arrays with
5693 *) In 'openssl passwd', verify passwords read from the terminal
5694 unless the '-salt' option is used (which usually means that
5695 verification would just waste user's time since the resulting
5696 hash is going to be compared with some given password hash)
5697 or the new '-noverify' option is used.
5699 This is an incompatible change, but it does not affect
5700 non-interactive use of 'openssl passwd' (passwords on the command
5701 line, '-stdin' option, '-in ...' option) and thus should not
5705 *) Remove all references to RSAref, since there's no more need for it.
5708 *) Make DSO load along a path given through an environment variable
5709 (SHLIB_PATH) with shl_load().
5712 *) Constify the ENGINE code as a result of BIGNUM constification.
5713 Also constify the RSA code and most things related to it. In a
5714 few places, most notable in the depth of the ASN.1 code, ugly
5715 casts back to non-const were required (to be solved at a later
5719 *) Make it so the openssl application has all engines loaded by default.
5722 *) Constify the BIGNUM routines a little more.
5725 *) Add the following functions:
5727 ENGINE_load_cswift()
5729 ENGINE_load_atalla()
5731 ENGINE_load_builtin_engines()
5733 That way, an application can itself choose if external engines that
5734 are built-in in OpenSSL shall ever be used or not. The benefit is
5735 that applications won't have to be linked with libdl or other dso
5736 libraries unless it's really needed.
5738 Changed 'openssl engine' to load all engines on demand.
5739 Changed the engine header files to avoid the duplication of some
5740 declarations (they differed!).
5743 *) 'openssl engine' can now list capabilities.
5746 *) Better error reporting in 'openssl engine'.
5749 *) Never call load_dh_param(NULL) in s_server.
5752 *) Add engine application. It can currently list engines by name and
5753 identity, and test if they are actually available.
5756 *) Improve RPM specification file by forcing symbolic linking and making
5757 sure the installed documentation is also owned by root.root.
5758 [Damien Miller <djm@mindrot.org>]
5760 *) Give the OpenSSL applications more possibilities to make use of
5761 keys (public as well as private) handled by engines.
5764 *) Add OCSP code that comes from CertCo.
5767 *) Add VMS support for the Rijndael code.
5770 *) Added untested support for Nuron crypto accelerator.
5773 *) Add support for external cryptographic devices. This code was
5774 previously distributed separately as the "engine" branch.
5775 [Geoff Thorpe, Richard Levitte]
5777 *) Rework the filename-translation in the DSO code. It is now possible to
5778 have far greater control over how a "name" is turned into a filename
5779 depending on the operating environment and any oddities about the
5780 different shared library filenames on each system.
5783 *) Support threads on FreeBSD-elf in Configure.
5786 *) Fix for SHA1 assembly problem with MASM: it produces
5787 warnings about corrupt line number information when assembling
5788 with debugging information. This is caused by the overlapping
5790 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5793 NCONF_get_number() has no error checking at all. As a replacement,
5794 NCONF_get_number_e() is defined (_e for "error checking") and is
5795 promoted strongly. The old NCONF_get_number is kept around for
5796 binary backward compatibility.
5797 Make it possible for methods to load from something other than a BIO,
5798 by providing a function pointer that is given a name instead of a BIO.
5799 For example, this could be used to load configuration data from an
5803 *) Fix for non blocking accept BIOs. Added new I/O special reason
5804 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5805 with non blocking I/O was not possible because no retry code was
5806 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5810 *) Added the beginnings of Rijndael support.
5813 *) Fix for bug in DirectoryString mask setting. Add support for
5814 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5815 to allow certificate printing to more controllable, additional
5816 'certopt' option to 'x509' to allow new printing options to be
5820 *) Clean old EAY MD5 hack from e_os.h.
5823 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5825 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5826 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5827 [Joe Orton, Steve Henson]
5829 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5831 *) Fix additional bug revealed by the NISCC test suite:
5833 Stop bug triggering large recursion when presented with
5834 certain ASN.1 tags (CVE-2003-0851)
5837 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5839 *) Fix various bugs revealed by running the NISCC test suite:
5841 Stop out of bounds reads in the ASN1 code when presented with
5842 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5844 If verify callback ignores invalid public key errors don't try to check
5845 certificate signature with the NULL public key.
5849 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5850 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5854 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5855 extra data after the compression methods not only for TLS 1.0
5856 but also for SSL 3.0 (as required by the specification).
5857 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5859 *) Change X509_certificate_type() to mark the key as exported/exportable
5860 when it's 512 *bits* long, not 512 bytes.
5863 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5865 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5866 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5867 a protocol version number mismatch like a decryption error
5868 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5871 *) Turn on RSA blinding by default in the default implementation
5872 to avoid a timing attack. Applications that don't want it can call
5873 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5874 They would be ill-advised to do so in most cases.
5875 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5877 *) Change RSA blinding code so that it works when the PRNG is not
5878 seeded (in this case, the secret RSA exponent is abused as
5879 an unpredictable seed -- if it is not unpredictable, there
5880 is no point in blinding anyway). Make RSA blinding thread-safe
5881 by remembering the creator's thread ID in rsa->blinding and
5882 having all other threads use local one-time blinding factors
5883 (this requires more computation than sharing rsa->blinding, but
5884 avoids excessive locking; and if an RSA object is not shared
5885 between threads, blinding will still be very fast).
5888 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5890 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5891 via timing by performing a MAC computation even if incorrrect
5892 block cipher padding has been found. This is a countermeasure
5893 against active attacks where the attacker has to distinguish
5894 between bad padding and a MAC verification error. (CVE-2003-0078)
5896 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5897 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5898 Martin Vuagnoux (EPFL, Ilion)]
5900 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5902 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5903 memory from it's contents. This is done with a counter that will
5904 place alternating values in each byte. This can be used to solve
5905 two issues: 1) the removal of calls to memset() by highly optimizing
5906 compilers, and 2) cleansing with other values than 0, since those can
5907 be read through on certain media, for example a swap space on disk.
5910 *) Bugfix: client side session caching did not work with external caching,
5911 because the session->cipher setting was not restored when reloading
5912 from the external cache. This problem was masked, when
5913 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
5914 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
5917 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
5918 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
5919 [Zeev Lieber <zeev-l@yahoo.com>]
5921 *) Undo an undocumented change introduced in 0.9.6e which caused
5922 repeated calls to OpenSSL_add_all_ciphers() and
5923 OpenSSL_add_all_digests() to be ignored, even after calling
5927 *) Change the default configuration reader to deal with last line not
5928 being properly terminated.
5931 *) Change X509_NAME_cmp() so it applies the special rules on handling
5932 DN values that are of type PrintableString, as well as RDNs of type
5933 emailAddress where the value has the type ia5String.
5934 [stefank@valicert.com via Richard Levitte]
5936 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
5937 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
5938 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
5939 the bitwise-OR of the two for use by the majority of applications
5940 wanting this behaviour, and update the docs. The documented
5941 behaviour and actual behaviour were inconsistent and had been
5942 changing anyway, so this is more a bug-fix than a behavioural
5944 [Geoff Thorpe, diagnosed by Nadav Har'El]
5946 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
5947 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
5950 *) Fix initialization code race conditions in
5951 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
5952 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
5953 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
5954 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
5955 ssl2_get_cipher_by_char(),
5956 ssl3_get_cipher_by_char().
5957 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
5959 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
5960 the cached sessions are flushed, as the remove_cb() might use ex_data
5961 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
5962 (see [openssl.org #212]).
5963 [Geoff Thorpe, Lutz Jaenicke]
5965 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
5966 length, instead of the encoding length to d2i_ASN1_OBJECT.
5969 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
5971 *) [In 0.9.6g-engine release:]
5972 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
5973 [Lynn Gazis <lgazis@rainbow.com>]
5975 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
5977 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
5978 and get fix the header length calculation.
5979 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
5980 Alon Kantor <alonk@checkpoint.com> (and others),
5983 *) Use proper error handling instead of 'assertions' in buffer
5984 overflow checks added in 0.9.6e. This prevents DoS (the
5985 assertions could call abort()).
5986 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
5988 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
5990 *) Add various sanity checks to asn1_get_length() to reject
5991 the ASN1 length bytes if they exceed sizeof(long), will appear
5992 negative or the content length exceeds the length of the
5994 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5996 *) Fix cipher selection routines: ciphers without encryption had no flags
5997 for the cipher strength set and where therefore not handled correctly
5998 by the selection routines (PR #130).
6001 *) Fix EVP_dsa_sha macro.
6005 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
6006 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
6007 that was added in OpenSSL 0.9.6d.
6009 As the countermeasure turned out to be incompatible with some
6010 broken SSL implementations, the new option is part of SSL_OP_ALL.
6011 SSL_OP_ALL is usually employed when compatibility with weird SSL
6012 implementations is desired (e.g. '-bugs' option to 's_client' and
6013 's_server'), so the new option is automatically set in many
6017 *) Changes in security patch:
6019 Changes marked "(CHATS)" were sponsored by the Defense Advanced
6020 Research Projects Agency (DARPA) and Air Force Research Laboratory,
6021 Air Force Materiel Command, USAF, under agreement number
6024 *) Add various sanity checks to asn1_get_length() to reject
6025 the ASN1 length bytes if they exceed sizeof(long), will appear
6026 negative or the content length exceeds the length of the
6027 supplied buffer. (CVE-2002-0659)
6028 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6030 *) Assertions for various potential buffer overflows, not known to
6032 [Ben Laurie (CHATS)]
6034 *) Various temporary buffers to hold ASCII versions of integers were
6035 too small for 64 bit platforms. (CVE-2002-0655)
6036 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6038 *) Remote buffer overflow in SSL3 protocol - an attacker could
6039 supply an oversized session ID to a client. (CVE-2002-0656)
6040 [Ben Laurie (CHATS)]
6042 *) Remote buffer overflow in SSL2 protocol - an attacker could
6043 supply an oversized client master key. (CVE-2002-0656)
6044 [Ben Laurie (CHATS)]
6046 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6048 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6049 encoded as NULL) with id-dsa-with-sha1.
6050 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6052 *) Check various X509_...() return values in apps/req.c.
6053 [Nils Larsch <nla@trustcenter.de>]
6055 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6056 an end-of-file condition would erronously be flagged, when the CRLF
6057 was just at the end of a processed block. The bug was discovered when
6058 processing data through a buffering memory BIO handing the data to a
6059 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6060 <ptsekov@syntrex.com> and Nedelcho Stanev.
6063 *) Implement a countermeasure against a vulnerability recently found
6064 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6065 before application data chunks to avoid the use of known IVs
6066 with data potentially chosen by the attacker.
6069 *) Fix length checks in ssl3_get_client_hello().
6072 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6073 to prevent ssl3_read_internal() from incorrectly assuming that
6074 ssl3_read_bytes() found application data while handshake
6075 processing was enabled when in fact s->s3->in_read_app_data was
6076 merely automatically cleared during the initial handshake.
6077 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6079 *) Fix object definitions for Private and Enterprise: they were not
6080 recognized in their shortname (=lowercase) representation. Extend
6081 obj_dat.pl to issue an error when using undefined keywords instead
6082 of silently ignoring the problem (Svenning Sorensen
6083 <sss@sss.dnsalias.net>).
6086 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6087 generators, i.e. generators other than 2 and 5. (Previously, the
6088 code did not properly initialise the 'add' and 'rem' values to
6089 BN_generate_prime().)
6091 In the new general case, we do not insist that 'generator' is
6092 actually a primitive root: This requirement is rather pointless;
6093 a generator of the order-q subgroup is just as good, if not
6097 *) Map new X509 verification errors to alerts. Discovered and submitted by
6098 Tom Wu <tom@arcot.com>.
6101 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6102 returning non-zero before the data has been completely received
6103 when using non-blocking I/O.
6104 [Bodo Moeller; problem pointed out by John Hughes]
6106 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6107 [Ben Laurie, Lutz Jaenicke]
6109 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6110 Yoram Zahavi <YoramZ@gilian.com>).
6113 *) Add information about CygWin 1.3 and on, and preserve proper
6114 configuration for the versions before that.
6115 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6117 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6118 check whether we deal with a copy of a session and do not delete from
6119 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6120 <izhar@checkpoint.com>.
6123 *) Do not store session data into the internal session cache, if it
6124 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6125 flag is set). Proposed by Aslam <aslam@funk.com>.
6128 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6132 *) [In 0.9.6d-engine release:]
6133 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6134 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6136 *) Add the configuration target linux-s390x.
6137 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6139 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6140 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6141 variable as an indication that a ClientHello message has been
6142 received. As the flag value will be lost between multiple
6143 invocations of ssl3_accept when using non-blocking I/O, the
6144 function may not be aware that a handshake has actually taken
6145 place, thus preventing a new session from being added to the
6148 To avoid this problem, we now set s->new_session to 2 instead of
6149 using a local variable.
6150 [Lutz Jaenicke, Bodo Moeller]
6152 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6153 if the SSL_R_LENGTH_MISMATCH error is detected.
6154 [Geoff Thorpe, Bodo Moeller]
6156 *) New 'shared_ldflag' column in Configure platform table.
6159 *) Fix EVP_CIPHER_mode macro.
6160 ["Dan S. Camper" <dan@bti.net>]
6162 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
6163 type, we must throw them away by setting rr->length to 0.
6164 [D P Chang <dpc@qualys.com>]
6166 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
6168 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
6169 <Dominikus.Scherkl@biodata.com>. (The previous implementation
6170 worked incorrectly for those cases where range = 10..._2 and
6171 3*range is two bits longer than range.)
6174 *) Only add signing time to PKCS7 structures if it is not already
6178 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
6179 OBJ_ld_ce should be OBJ_id_ce.
6180 Also some ip-pda OIDs in crypto/objects/objects.txt were
6181 incorrect (cf. RFC 3039).
6182 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
6184 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
6185 returns early because it has nothing to do.
6186 [Andy Schneider <andy.schneider@bjss.co.uk>]
6188 *) [In 0.9.6c-engine release:]
6189 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
6190 [Andy Schneider <andy.schneider@bjss.co.uk>]
6192 *) [In 0.9.6c-engine release:]
6193 Add support for Cryptographic Appliance's keyserver technology.
6194 (Use engine 'keyclient')
6195 [Cryptographic Appliances and Geoff Thorpe]
6197 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
6198 is called via tools/c89.sh because arguments have to be
6199 rearranged (all '-L' options must appear before the first object
6201 [Richard Shapiro <rshapiro@abinitio.com>]
6203 *) [In 0.9.6c-engine release:]
6204 Add support for Broadcom crypto accelerator cards, backported
6206 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
6208 *) [In 0.9.6c-engine release:]
6209 Add support for SureWare crypto accelerator cards from
6210 Baltimore Technologies. (Use engine 'sureware')
6211 [Baltimore Technologies and Mark Cox]
6213 *) [In 0.9.6c-engine release:]
6214 Add support for crypto accelerator cards from Accelerated
6215 Encryption Processing, www.aep.ie. (Use engine 'aep')
6216 [AEP Inc. and Mark Cox]
6218 *) Add a configuration entry for gcc on UnixWare.
6219 [Gary Benson <gbenson@redhat.com>]
6221 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
6222 messages are stored in a single piece (fixed-length part and
6223 variable-length part combined) and fix various bugs found on the way.
6226 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
6227 instead. BIO_gethostbyname() does not know what timeouts are
6228 appropriate, so entries would stay in cache even when they have
6230 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
6232 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
6233 faced with a pathologically small ClientHello fragment that does
6234 not contain client_version: Instead of aborting with an error,
6235 simply choose the highest available protocol version (i.e.,
6236 TLS 1.0 unless it is disabled). In practice, ClientHello
6237 messages are never sent like this, but this change gives us
6238 strictly correct behaviour at least for TLS.
6241 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
6242 never resets s->method to s->ctx->method when called from within
6243 one of the SSL handshake functions.
6244 [Bodo Moeller; problem pointed out by Niko Baric]
6246 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
6247 (sent using the client's version number) if client_version is
6248 smaller than the protocol version in use. Also change
6249 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
6250 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
6251 the client will at least see that alert.
6254 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
6258 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
6259 client receives HelloRequest while in a handshake.
6260 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
6262 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
6263 should end in 'break', not 'goto end' which circuments various
6264 cleanups done in state SSL_ST_OK. But session related stuff
6265 must be disabled for SSL_ST_OK in the case that we just sent a
6268 Also avoid some overhead by not calling ssl_init_wbio_buffer()
6269 before just sending a HelloRequest.
6270 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
6272 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
6273 reveal whether illegal block cipher padding was found or a MAC
6274 verification error occured. (Neither SSLerr() codes nor alerts
6275 are directly visible to potential attackers, but the information
6276 may leak via logfiles.)
6278 Similar changes are not required for the SSL 2.0 implementation
6279 because the number of padding bytes is sent in clear for SSL 2.0,
6280 and the extra bytes are just ignored. However ssl/s2_pkt.c
6281 failed to verify that the purported number of padding bytes is in
6285 *) Add OpenUNIX-8 support including shared libraries
6286 (Boyd Lynn Gerber <gerberb@zenez.com>).
6289 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
6290 'wristwatch attack' using huge encoding parameters (cf.
6291 James H. Manger's CRYPTO 2001 paper). Note that the
6292 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
6293 encoding parameters and hence was not vulnerable.
6296 *) BN_sqr() bug fix.
6297 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
6299 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
6300 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
6301 followed by modular reduction.
6302 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
6304 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
6305 equivalent based on BN_pseudo_rand() instead of BN_rand().
6308 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
6309 This function was broken, as the check for a new client hello message
6310 to handle SGC did not allow these large messages.
6311 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
6314 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
6317 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
6318 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
6321 *) Rework the configuration and shared library support for Tru64 Unix.
6322 The configuration part makes use of modern compiler features and
6323 still retains old compiler behavior for those that run older versions
6324 of the OS. The shared library support part includes a variant that
6325 uses the RPATH feature, and is available through the special
6326 configuration target "alpha-cc-rpath", which will never be selected
6328 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
6330 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
6331 with the same message size as in ssl3_get_certificate_request().
6332 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
6333 messages might inadvertently be reject as too long.
6334 [Petr Lampa <lampa@fee.vutbr.cz>]
6336 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6339 *) Modified SSL library such that the verify_callback that has been set
6340 specificly for an SSL object with SSL_set_verify() is actually being
6341 used. Before the change, a verify_callback set with this function was
6342 ignored and the verify_callback() set in the SSL_CTX at the time of
6343 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6344 to allow the necessary settings.
6347 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6348 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6349 done automatically (in contradiction to the requirements of the C
6350 standard). This made problems when used from OpenSSH.
6353 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6354 dh->length and always used
6356 BN_rand_range(priv_key, dh->p).
6358 BN_rand_range() is not necessary for Diffie-Hellman, and this
6359 specific range makes Diffie-Hellman unnecessarily inefficient if
6360 dh->length (recommended exponent length) is much smaller than the
6361 length of dh->p. We could use BN_rand_range() if the order of
6362 the subgroup was stored in the DH structure, but we only have
6367 BN_rand(priv_key, l, ...)
6369 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6375 RSA_eay_public_encrypt
6376 RSA_eay_private_decrypt
6377 RSA_eay_private_encrypt (signing)
6378 RSA_eay_public_decrypt (signature verification)
6380 (default implementations for RSA_public_encrypt,
6381 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6382 always reject numbers >= n.
6385 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6386 to synchronize access to 'locking_thread'. This is necessary on
6387 systems where access to 'locking_thread' (an 'unsigned long'
6388 variable) is not atomic.
6391 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6392 *before* setting the 'crypto_lock_rand' flag. The previous code had
6393 a race condition if 0 is a valid thread ID.
6394 [Travis Vitek <vitek@roguewave.com>]
6396 *) Add support for shared libraries under Irix.
6397 [Albert Chin-A-Young <china@thewrittenword.com>]
6399 *) Add configuration option to build on Linux on both big-endian and
6401 [Ralf Baechle <ralf@uni-koblenz.de>]
6403 *) Add the possibility to create shared libraries on HP-UX.
6406 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6408 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6409 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6410 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6411 PRNG state recovery was possible based on the output of
6412 one PRNG request appropriately sized to gain knowledge on
6413 'md' followed by enough consecutive 1-byte PRNG requests
6414 to traverse all of 'state'.
6416 1. When updating 'md_local' (the current thread's copy of 'md')
6417 during PRNG output generation, hash all of the previous
6418 'md_local' value, not just the half used for PRNG output.
6420 2. Make the number of bytes from 'state' included into the hash
6421 independent from the number of PRNG bytes requested.
6423 The first measure alone would be sufficient to avoid
6424 Markku-Juhani's attack. (Actually it had never occurred
6425 to me that the half of 'md_local' used for chaining was the
6426 half from which PRNG output bytes were taken -- I had always
6427 assumed that the secret half would be used.) The second
6428 measure makes sure that additional data from 'state' is never
6429 mixed into 'md_local' in small portions; this heuristically
6430 further strengthens the PRNG.
6433 *) Fix crypto/bn/asm/mips3.s.
6436 *) When only the key is given to "enc", the IV is undefined. Print out
6437 an error message in this case.
6440 *) Handle special case when X509_NAME is empty in X509 printing routines.
6443 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6444 positive and less than q.
6447 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6448 used: it isn't thread safe and the add_lock_callback should handle
6450 [Paul Rose <Paul.Rose@bridge.com>]
6452 *) Verify that incoming data obeys the block size in
6453 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6457 [Ulf Möller, Bodo Möller]
6459 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6460 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6461 when fixing the server behaviour for backwards-compatible 'client
6462 hello' messages. (Note that the attack is impractical against
6463 SSL 3.0 and TLS 1.0 anyway because length and version checking
6464 means that the probability of guessing a valid ciphertext is
6465 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6468 Before 0.9.5, the countermeasure (hide the error by generating a
6469 random 'decryption result') did not work properly because
6470 ERR_clear_error() was missing, meaning that SSL_get_error() would
6471 detect the supposedly ignored error.
6473 Both problems are now fixed.
6476 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6477 (previously it was 1024).
6480 *) Fix for compatibility mode trust settings: ignore trust settings
6481 unless some valid trust or reject settings are present.
6484 *) Fix for blowfish EVP: its a variable length cipher.
6487 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6488 parameters in DSA public key structures and return an error in the
6489 DSA routines if parameters are absent.
6492 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6493 in the current directory if neither $RANDFILE nor $HOME was set.
6494 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6495 caused some confusion to Windows users who haven't defined $HOME.
6496 Thus RAND_file_name() is changed again: e_os.h can define a
6497 DEFAULT_HOME, which will be used if $HOME is not set.
6498 For Windows, we use "C:"; on other platforms, we still require
6499 environment variables.
6501 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6502 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6503 having multiple threads call RAND_poll() concurrently.
6506 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6507 combination of a flag and a thread ID variable.
6508 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6509 flag), *other* threads can enter ssleay_add_bytes without obeying
6510 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6511 that they do not hold after the first thread unsets add_do_not_lock).
6514 *) Change bctest again: '-x' expressions are not available in all
6518 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6520 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6521 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6523 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6524 the default extension for executables, if any. Also, make the perl
6525 scripts that use symlink() to test if it really exists and use "cp"
6526 if it doesn't. All this made OpenSSL compilable and installable in
6530 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6531 If SEQUENCE is length is indefinite just set c->slen to the total
6532 amount of data available.
6533 [Steve Henson, reported by shige@FreeBSD.org]
6534 [This change does not apply to 0.9.7.]
6536 *) Change bctest to avoid here-documents inside command substitution
6537 (workaround for FreeBSD /bin/sh bug).
6538 For compatibility with Ultrix, avoid shell functions (introduced
6539 in the bctest version that searches along $PATH).
6542 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6543 with des_encrypt() defined on some operating systems, like Solaris
6547 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6548 On the Importance of Eliminating Errors in Cryptographic
6549 Computations, J. Cryptology 14 (2001) 2, 101-119,
6550 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6553 *) MIPS assembler BIGNUM division bug fix.
6556 *) Disabled incorrect Alpha assembler code.
6559 *) Fix PKCS#7 decode routines so they correctly update the length
6560 after reading an EOC for the EXPLICIT tag.
6562 [This change does not apply to 0.9.7.]
6564 *) Fix bug in PKCS#12 key generation routines. This was triggered
6565 if a 3DES key was generated with a 0 initial byte. Include
6566 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6567 (but broken) behaviour.
6570 *) Enhance bctest to search for a working bc along $PATH and print
6572 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6574 *) Fix memory leaks in err.c: free err_data string if necessary;
6575 don't write to the wrong index in ERR_set_error_data.
6578 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6582 *) Replace rdtsc with _emit statements for VC++ version 5.
6583 [Jeremy Cooper <jeremy@baymoo.org>]
6585 *) Make it possible to reuse SSLv2 sessions.
6588 *) In copy_email() check for >= 0 as a return value for
6589 X509_NAME_get_index_by_NID() since 0 is a valid index.
6590 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6592 *) Avoid coredump with unsupported or invalid public keys by checking if
6593 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6594 PKCS7_verify() fails with non detached data.
6597 *) Don't use getenv in library functions when run as setuid/setgid.
6598 New function OPENSSL_issetugid().
6601 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6602 due to incorrect handling of multi-threading:
6604 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6606 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6608 3. Count how many times MemCheck_off() has been called so that
6609 nested use can be treated correctly. This also avoids
6610 inband-signalling in the previous code (which relied on the
6611 assumption that thread ID 0 is impossible).
6614 *) Add "-rand" option also to s_client and s_server.
6617 *) Fix CPU detection on Irix 6.x.
6618 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6619 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6621 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6624 [This change does not apply to 0.9.7.]
6626 *) Use the cached encoding of an X509_NAME structure rather than
6627 copying it. This is apparently the reason for the libsafe "errors"
6628 but the code is actually correct.
6631 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6632 Bleichenbacher's DSA attack.
6633 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6634 to be set and top=0 forces the highest bit to be set; top=-1 is new
6635 and leaves the highest bit random.
6636 [Ulf Moeller, Bodo Moeller]
6638 *) In the NCONF_...-based implementations for CONF_... queries
6639 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6640 a temporary CONF structure with the data component set to NULL
6641 (which gives segmentation faults in lh_retrieve).
6642 Instead, use NULL for the CONF pointer in CONF_get_string and
6643 CONF_get_number (which may use environment variables) and directly
6644 return NULL from CONF_get_section.
6647 *) Fix potential buffer overrun for EBCDIC.
6650 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6651 keyUsage if basicConstraints absent for a CA.
6654 *) Make SMIME_write_PKCS7() write mail header values with a format that
6655 is more generally accepted (no spaces before the semicolon), since
6656 some programs can't parse those values properly otherwise. Also make
6657 sure BIO's that break lines after each write do not create invalid
6661 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6662 macros previously used would not encode an empty SEQUENCE OF
6663 and break the signature.
6665 [This change does not apply to 0.9.7.]
6667 *) Zero the premaster secret after deriving the master secret in
6671 *) Add some EVP_add_digest_alias registrations (as found in
6672 OpenSSL_add_all_digests()) to SSL_library_init()
6673 aka OpenSSL_add_ssl_algorithms(). This provides improved
6674 compatibility with peers using X.509 certificates
6675 with unconventional AlgorithmIdentifier OIDs.
6678 *) Fix for Irix with NO_ASM.
6679 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6681 *) ./config script fixes.
6682 [Ulf Moeller, Richard Levitte]
6684 *) Fix 'openssl passwd -1'.
6687 *) Change PKCS12_key_gen_asc() so it can cope with non null
6688 terminated strings whose length is passed in the passlen
6689 parameter, for example from PEM callbacks. This was done
6690 by adding an extra length parameter to asc2uni().
6691 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6693 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6694 call failed, free the DSA structure.
6697 *) Fix to uni2asc() to cope with zero length Unicode strings.
6698 These are present in some PKCS#12 files.
6701 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6702 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6703 when writing a 32767 byte record.
6704 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6706 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6707 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6709 (RSA objects have a reference count access to which is protected
6710 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6711 so they are meant to be shared between threads.)
6712 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6713 "Reddie, Steven" <Steven.Reddie@ca.com>]
6715 *) Fix a deadlock in CRYPTO_mem_leaks().
6718 *) Use better test patterns in bntest.
6721 *) rand_win.c fix for Borland C.
6724 *) BN_rshift bugfix for n == 0.
6727 *) Add a 'bctest' script that checks for some known 'bc' bugs
6728 so that 'make test' does not abort just because 'bc' is broken.
6731 *) Store verify_result within SSL_SESSION also for client side to
6732 avoid potential security hole. (Re-used sessions on the client side
6733 always resulted in verify_result==X509_V_OK, not using the original
6734 result of the server certificate verification.)
6737 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6738 SSL3_RT_APPLICATION_DATA, return 0.
6739 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6743 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6744 releases, have been re-implemented by renaming the previous
6745 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6746 and ssl3_read_internal, respectively, and adding 'peek' parameters
6747 to them. The new ssl[23]_{read,peek} functions are calls to
6748 ssl[23]_read_internal with the 'peek' flag set appropriately.
6749 A 'peek' parameter has also been added to ssl3_read_bytes, which
6750 does the actual work for ssl3_read_internal.
6753 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6754 the method-specific "init()" handler. Also clean up ex_data after
6755 calling the method-specific "finish()" handler. Previously, this was
6756 happening the other way round.
6759 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6760 The previous value, 12, was not always sufficient for BN_mod_exp().
6763 *) Make sure that shared libraries get the internal name engine with
6764 the full version number and not just 0. This should mark the
6765 shared libraries as not backward compatible. Of course, this should
6766 be changed again when we can guarantee backward binary compatibility.
6769 *) Fix typo in get_cert_by_subject() in by_dir.c
6770 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6772 *) Rework the system to generate shared libraries:
6774 - Make note of the expected extension for the shared libraries and
6775 if there is a need for symbolic links from for example libcrypto.so.0
6776 to libcrypto.so.0.9.7. There is extended info in Configure for
6779 - Make as few rebuilds of the shared libraries as possible.
6781 - Still avoid linking the OpenSSL programs with the shared libraries.
6783 - When installing, install the shared libraries separately from the
6787 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6789 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6790 and not in SSL_clear because the latter is also used by the
6791 accept/connect functions; previously, the settings made by
6792 SSL_set_read_ahead would be lost during the handshake.
6793 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6795 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6796 Previously, it would create entries for disableed algorithms no
6800 *) Added several new manual pages for SSL_* function.
6803 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6805 *) In ssl23_get_client_hello, generate an error message when faced
6806 with an initial SSL 3.0/TLS record that is too small to contain the
6807 first two bytes of the ClientHello message, i.e. client_version.
6808 (Note that this is a pathologic case that probably has never happened
6809 in real life.) The previous approach was to use the version number
6810 from the record header as a substitute; but our protocol choice
6811 should not depend on that one because it is not authenticated
6812 by the Finished messages.
6815 *) More robust randomness gathering functions for Windows.
6816 [Jeffrey Altman <jaltman@columbia.edu>]
6818 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6819 not set then we don't setup the error code for issuer check errors
6820 to avoid possibly overwriting other errors which the callback does
6821 handle. If an application does set the flag then we assume it knows
6822 what it is doing and can handle the new informational codes
6826 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6827 a general "ANY" type, as such it should be able to decode anything
6828 including tagged types. However it didn't check the class so it would
6829 wrongly interpret tagged types in the same way as their universal
6830 counterpart and unknown types were just rejected. Changed so that the
6831 tagged and unknown types are handled in the same way as a SEQUENCE:
6832 that is the encoding is stored intact. There is also a new type
6833 "V_ASN1_OTHER" which is used when the class is not universal, in this
6834 case we have no idea what the actual type is so we just lump them all
6838 *) On VMS, stdout may very well lead to a file that is written to
6839 in a record-oriented fashion. That means that every write() will
6840 write a separate record, which will be read separately by the
6841 programs trying to read from it. This can be very confusing.
6843 The solution is to put a BIO filter in the way that will buffer
6844 text until a linefeed is reached, and then write everything a
6845 line at a time, so every record written will be an actual line,
6846 not chunks of lines and not (usually doesn't happen, but I've
6847 seen it once) several lines in one record. BIO_f_linebuffer() is
6850 Currently, it's a VMS-only method, because that's where it has
6851 been tested well enough.
6854 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6855 it can return incorrect results.
6856 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6857 but it was in 0.9.6-beta[12].)
6860 *) Disable the check for content being present when verifying detached
6861 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6862 include zero length content when signing messages.
6865 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6866 BIO_ctrl (for BIO pairs).
6869 *) Add DSO method for VMS.
6872 *) Bug fix: Montgomery multiplication could produce results with the
6876 *) Add RPM specification openssl.spec and modify it to build three
6877 packages. The default package contains applications, application
6878 documentation and run-time libraries. The devel package contains
6879 include files, static libraries and function documentation. The
6880 doc package contains the contents of the doc directory. The original
6881 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6884 *) Add a large number of documentation files for many SSL routines.
6885 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6887 *) Add a configuration entry for Sony News 4.
6888 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6890 *) Don't set the two most significant bits to one when generating a
6891 random number < q in the DSA library.
6894 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6895 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6896 the underlying transport is blocking) if a handshake took place.
6897 (The default behaviour is needed by applications such as s_client
6898 and s_server that use select() to determine when to use SSL_read;
6899 but for applications that know in advance when to expect data, it
6900 just makes things more complicated.)
6903 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
6907 *) Add a few more EBCDIC conditionals that make `req' and `x509'
6908 work better on such systems.
6909 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6911 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
6912 Update PKCS12_parse() so it copies the friendlyName and the
6913 keyid to the certificates aux info.
6916 *) Fix bug in PKCS7_verify() which caused an infinite loop
6917 if there was more than one signature.
6918 [Sven Uszpelkat <su@celocom.de>]
6920 *) Major change in util/mkdef.pl to include extra information
6921 about each symbol, as well as presentig variables as well
6922 as functions. This change means that there's n more need
6923 to rebuild the .num files when some algorithms are excluded.
6926 *) Allow the verify time to be set by an application,
6927 rather than always using the current time.
6930 *) Phase 2 verify code reorganisation. The certificate
6931 verify code now looks up an issuer certificate by a
6932 number of criteria: subject name, authority key id
6933 and key usage. It also verifies self signed certificates
6934 by the same criteria. The main comparison function is
6935 X509_check_issued() which performs these checks.
6937 Lot of changes were necessary in order to support this
6938 without completely rewriting the lookup code.
6940 Authority and subject key identifier are now cached.
6942 The LHASH 'certs' is X509_STORE has now been replaced
6943 by a STACK_OF(X509_OBJECT). This is mainly because an
6944 LHASH can't store or retrieve multiple objects with
6945 the same hash value.
6947 As a result various functions (which were all internal
6948 use only) have changed to handle the new X509_STORE
6949 structure. This will break anything that messed round
6950 with X509_STORE internally.
6952 The functions X509_STORE_add_cert() now checks for an
6953 exact match, rather than just subject name.
6955 The X509_STORE API doesn't directly support the retrieval
6956 of multiple certificates matching a given criteria, however
6957 this can be worked round by performing a lookup first
6958 (which will fill the cache with candidate certificates)
6959 and then examining the cache for matches. This is probably
6960 the best we can do without throwing out X509_LOOKUP
6961 entirely (maybe later...).
6963 The X509_VERIFY_CTX structure has been enhanced considerably.
6965 All certificate lookup operations now go via a get_issuer()
6966 callback. Although this currently uses an X509_STORE it
6967 can be replaced by custom lookups. This is a simple way
6968 to bypass the X509_STORE hackery necessary to make this
6969 work and makes it possible to use more efficient techniques
6970 in future. A very simple version which uses a simple
6971 STACK for its trusted certificate store is also provided
6972 using X509_STORE_CTX_trusted_stack().
6974 The verify_cb() and verify() callbacks now have equivalents
6975 in the X509_STORE_CTX structure.
6977 X509_STORE_CTX also has a 'flags' field which can be used
6978 to customise the verify behaviour.
6981 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
6982 excludes S/MIME capabilities.
6985 *) When a certificate request is read in keep a copy of the
6986 original encoding of the signed data and use it when outputing
6987 again. Signatures then use the original encoding rather than
6988 a decoded, encoded version which may cause problems if the
6989 request is improperly encoded.
6992 *) For consistency with other BIO_puts implementations, call
6993 buffer_write(b, ...) directly in buffer_puts instead of calling
6996 In BIO_puts, increment b->num_write as in BIO_write.
6997 [Peter.Sylvester@EdelWeb.fr]
6999 *) Fix BN_mul_word for the case where the word is 0. (We have to use
7000 BN_zero, we may not return a BIGNUM with an array consisting of
7004 *) Avoid calling abort() from within the library when problems are
7005 detected, except if preprocessor symbols have been defined
7006 (such as REF_CHECK, BN_DEBUG etc.).
7009 *) New openssl application 'rsautl'. This utility can be
7010 used for low level RSA operations. DER public key
7011 BIO/fp routines also added.
7014 *) New Configure entry and patches for compiling on QNX 4.
7015 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
7017 *) A demo state-machine implementation was sponsored by
7018 Nuron (http://www.nuron.com/) and is now available in
7019 demos/state_machine.
7022 *) New options added to the 'dgst' utility for signature
7023 generation and verification.
7026 *) Unrecognized PKCS#7 content types are now handled via a
7027 catch all ASN1_TYPE structure. This allows unsupported
7028 types to be stored as a "blob" and an application can
7029 encode and decode it manually.
7032 *) Fix various signed/unsigned issues to make a_strex.c
7034 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7036 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7037 length if passed a buffer. ASN1_INTEGER_to_BN failed
7038 if passed a NULL BN and its argument was negative.
7039 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7041 *) Modification to PKCS#7 encoding routines to output definite
7042 length encoding. Since currently the whole structures are in
7043 memory there's not real point in using indefinite length
7044 constructed encoding. However if OpenSSL is compiled with
7045 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7048 *) Added BIO_vprintf() and BIO_vsnprintf().
7051 *) Added more prefixes to parse for in the the strings written
7052 through a logging bio, to cover all the levels that are available
7053 through syslog. The prefixes are now:
7055 PANIC, EMERG, EMR => LOG_EMERG
7056 ALERT, ALR => LOG_ALERT
7057 CRIT, CRI => LOG_CRIT
7058 ERROR, ERR => LOG_ERR
7059 WARNING, WARN, WAR => LOG_WARNING
7060 NOTICE, NOTE, NOT => LOG_NOTICE
7061 INFO, INF => LOG_INFO
7062 DEBUG, DBG => LOG_DEBUG
7064 and as before, if none of those prefixes are present at the
7065 beginning of the string, LOG_ERR is chosen.
7067 On Win32, the LOG_* levels are mapped according to this:
7069 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7070 LOG_WARNING => EVENTLOG_WARNING_TYPE
7071 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7075 *) Made it possible to reconfigure with just the configuration
7076 argument "reconf" or "reconfigure". The command line arguments
7077 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7078 and are retrieved from there when reconfiguring.
7082 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7084 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7087 *) The obj_dat.pl script was messing up the sorting of object
7088 names. The reason was that it compared the quoted version
7089 of strings as a result "OCSP" > "OCSP Signing" because
7090 " > SPACE. Changed script to store unquoted versions of
7091 names and add quotes on output. It was also omitting some
7092 names from the lookup table if they were given a default
7093 value (that is if SN is missing it is given the same
7094 value as LN and vice versa), these are now added on the
7095 grounds that if an object has a name we should be able to
7096 look it up. Finally added warning output when duplicate
7097 short or long names are found.
7100 *) Changes needed for Tandem NSK.
7101 [Scott Uroff <scott@xypro.com>]
7103 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7104 RSA_padding_check_SSLv23(), special padding was never detected
7105 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7106 version rollback attacks was not effective.
7108 In s23_clnt.c, don't use special rollback-attack detection padding
7109 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7110 client; similarly, in s23_srvr.c, don't do the rollback check if
7111 SSL 2.0 is the only protocol enabled in the server.
7114 *) Make it possible to get hexdumps of unprintable data with 'openssl
7115 asn1parse'. By implication, the functions ASN1_parse_dump() and
7116 BIO_dump_indent() are added.
7119 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7120 these print out strings and name structures based on various
7121 flags including RFC2253 support and proper handling of
7122 multibyte characters. Added options to the 'x509' utility
7123 to allow the various flags to be set.
7126 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7127 Also change the functions X509_cmp_current_time() and
7128 X509_gmtime_adj() work with an ASN1_TIME structure,
7129 this will enable certificates using GeneralizedTime in validity
7130 dates to be checked.
7133 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7134 negative public key encodings) on by default,
7135 NO_NEG_PUBKEY_BUG can be set to disable it.
7138 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7139 content octets. An i2c_ASN1_OBJECT is unnecessary because
7140 the encoding can be trivially obtained from the structure.
7143 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7144 not read locks (CRYPTO_r_[un]lock).
7147 *) A first attempt at creating official support for shared
7148 libraries through configuration. I've kept it so the
7149 default is static libraries only, and the OpenSSL programs
7150 are always statically linked for now, but there are
7151 preparations for dynamic linking in place.
7152 This has been tested on Linux and Tru64.
7155 *) Randomness polling function for Win9x, as described in:
7156 Peter Gutmann, Software Generation of Practically Strong
7160 *) Fix so PRNG is seeded in req if using an already existing
7164 *) New options to smime application. -inform and -outform
7165 allow alternative formats for the S/MIME message including
7166 PEM and DER. The -content option allows the content to be
7167 specified separately. This should allow things like Netscape
7168 form signing output easier to verify.
7171 *) Fix the ASN1 encoding of tags using the 'long form'.
7174 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
7175 STRING types. These convert content octets to and from the
7176 underlying type. The actual tag and length octets are
7177 already assumed to have been read in and checked. These
7178 are needed because all other string types have virtually
7179 identical handling apart from the tag. By having versions
7180 of the ASN1 functions that just operate on content octets
7181 IMPLICIT tagging can be handled properly. It also allows
7182 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
7183 and ASN1_INTEGER are identical apart from the tag.
7186 *) Change the handling of OID objects as follows:
7188 - New object identifiers are inserted in objects.txt, following
7189 the syntax given in objects.README.
7190 - objects.pl is used to process obj_mac.num and create a new
7192 - obj_dat.pl is used to create a new obj_dat.h, using the data in
7195 This is currently kind of a hack, and the perl code in objects.pl
7196 isn't very elegant, but it works as I intended. The simplest way
7197 to check that it worked correctly is to look in obj_dat.h and
7198 check the array nid_objs and make sure the objects haven't moved
7199 around (this is important!). Additions are OK, as well as
7200 consistent name changes.
7203 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
7206 *) Addition of the command line parameter '-rand file' to 'openssl req'.
7207 The given file adds to whatever has already been seeded into the
7208 random pool through the RANDFILE configuration file option or
7209 environment variable, or the default random state file.
7212 *) mkstack.pl now sorts each macro group into lexical order.
7213 Previously the output order depended on the order the files
7214 appeared in the directory, resulting in needless rewriting
7218 *) Patches to make OpenSSL compile under Win32 again. Mostly
7219 work arounds for the VC++ problem that it treats func() as
7220 func(void). Also stripped out the parts of mkdef.pl that
7221 added extra typesafe functions: these no longer exist.
7224 *) Reorganisation of the stack code. The macros are now all
7225 collected in safestack.h . Each macro is defined in terms of
7226 a "stack macro" of the form SKM_<name>(type, a, b). The
7227 DEBUG_SAFESTACK is now handled in terms of function casts,
7228 this has the advantage of retaining type safety without the
7229 use of additional functions. If DEBUG_SAFESTACK is not defined
7230 then the non typesafe macros are used instead. Also modified the
7231 mkstack.pl script to handle the new form. Needs testing to see
7232 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
7233 the default if no major problems. Similar behaviour for ASN1_SET_OF
7234 and PKCS12_STACK_OF.
7237 *) When some versions of IIS use the 'NET' form of private key the
7238 key derivation algorithm is different. Normally MD5(password) is
7239 used as a 128 bit RC4 key. In the modified case
7240 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
7241 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
7242 as the old Netscape_RSA functions except they have an additional
7243 'sgckey' parameter which uses the modified algorithm. Also added
7244 an -sgckey command line option to the rsa utility. Thanks to
7245 Adrian Peck <bertie@ncipher.com> for posting details of the modified
7246 algorithm to openssl-dev.
7249 *) The evp_local.h macros were using 'c.##kname' which resulted in
7250 invalid expansion on some systems (SCO 5.0.5 for example).
7251 Corrected to 'c.kname'.
7252 [Phillip Porch <root@theporch.com>]
7254 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
7255 a STACK of email addresses from a certificate or request, these look
7256 in the subject name and the subject alternative name extensions and
7257 omit any duplicate addresses.
7260 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
7261 This makes DSA verification about 2 % faster.
7264 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
7265 (meaning that now 2^5 values will be precomputed, which is only 4 KB
7266 plus overhead for 1024 bit moduli).
7267 This makes exponentiations about 0.5 % faster for 1024 bit
7268 exponents (as measured by "openssl speed rsa2048").
7271 *) Rename memory handling macros to avoid conflicts with other
7273 Malloc => OPENSSL_malloc
7274 Malloc_locked => OPENSSL_malloc_locked
7275 Realloc => OPENSSL_realloc
7276 Free => OPENSSL_free
7279 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
7280 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
7283 *) CygWin32 support.
7284 [John Jarvie <jjarvie@newsguy.com>]
7286 *) The type-safe stack code has been rejigged. It is now only compiled
7287 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
7288 by default all type-specific stack functions are "#define"d back to
7289 standard stack functions. This results in more streamlined output
7290 but retains the type-safety checking possibilities of the original
7294 *) The STACK code has been cleaned up, and certain type declarations
7295 that didn't make a lot of sense have been brought in line. This has
7296 also involved a cleanup of sorts in safestack.h to more correctly
7297 map type-safe stack functions onto their plain stack counterparts.
7298 This work has also resulted in a variety of "const"ifications of
7299 lots of the code, especially "_cmp" operations which should normally
7300 be prototyped with "const" parameters anyway.
7303 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
7304 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
7305 (The PRNG state consists of two parts, the large pool 'state' and 'md',
7306 where all of 'md' is used each time the PRNG is used, but 'state'
7307 is used only indexed by a cyclic counter. As entropy may not be
7308 well distributed from the beginning, 'md' is important as a
7309 chaining variable. However, the output function chains only half
7310 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
7311 all of 'md', and seeding with STATE_SIZE dummy bytes will result
7312 in all of 'state' being rewritten, with the new values depending
7313 on virtually all of 'md'. This overcomes the 80 bit limitation.)
7316 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
7317 the handshake is continued after ssl_verify_cert_chain();
7318 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
7319 can lead to 'unexplainable' connection aborts later.
7320 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
7322 *) Major EVP API cipher revision.
7323 Add hooks for extra EVP features. This allows various cipher
7324 parameters to be set in the EVP interface. Support added for variable
7325 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
7326 setting of RC2 and RC5 parameters.
7328 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
7331 Remove lots of duplicated code from the EVP library. For example *every*
7332 cipher init() function handles the 'iv' in the same way according to the
7333 cipher mode. They also all do nothing if the 'key' parameter is NULL and
7334 for CFB and OFB modes they zero ctx->num.
7336 New functionality allows removal of S/MIME code RC2 hack.
7338 Most of the routines have the same form and so can be declared in terms
7341 By shifting this to the top level EVP_CipherInit() it can be removed from
7342 all individual ciphers. If the cipher wants to handle IVs or keys
7343 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7346 Change lots of functions like EVP_EncryptUpdate() to now return a
7347 value: although software versions of the algorithms cannot fail
7348 any installed hardware versions can.
7351 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7352 this option is set, tolerate broken clients that send the negotiated
7353 protocol version number instead of the requested protocol version
7357 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7358 i.e. non-zero for export ciphersuites, zero otherwise.
7359 Previous versions had this flag inverted, inconsistent with
7360 rsa_tmp_cb (..._TMP_RSA_CB).
7361 [Bodo Moeller; problem reported by Amit Chopra]
7363 *) Add missing DSA library text string. Work around for some IIS
7364 key files with invalid SEQUENCE encoding.
7367 *) Add a document (doc/standards.txt) that list all kinds of standards
7368 and so on that are implemented in OpenSSL.
7371 *) Enhance c_rehash script. Old version would mishandle certificates
7372 with the same subject name hash and wouldn't handle CRLs at all.
7373 Added -fingerprint option to crl utility, to support new c_rehash
7377 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7380 *) Fix for SSL server purpose checking. Server checking was
7381 rejecting certificates which had extended key usage present
7382 but no ssl client purpose.
7383 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7385 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7386 is a little unclear about how a blank password is handled.
7387 Since the password in encoded as a BMPString with terminating
7388 double NULL a zero length password would end up as just the
7389 double NULL. However no password at all is different and is
7390 handled differently in the PKCS#12 key generation code. NS
7391 treats a blank password as zero length. MSIE treats it as no
7392 password on export: but it will try both on import. We now do
7393 the same: PKCS12_parse() tries zero length and no password if
7394 the password is set to "" or NULL (NULL is now a valid password:
7395 it wasn't before) as does the pkcs12 application.
7398 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7399 perror when PEM_read_bio_X509_REQ fails, the error message must
7400 be obtained from the error queue.
7403 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7404 it in ERR_remove_state if appropriate, and change ERR_get_state
7405 accordingly to avoid race conditions (this is necessary because
7406 thread_hash is no longer constant once set).
7409 *) Bugfix for linux-elf makefile.one.
7412 *) RSA_get_default_method() will now cause a default
7413 RSA_METHOD to be chosen if one doesn't exist already.
7414 Previously this was only set during a call to RSA_new()
7415 or RSA_new_method(NULL) meaning it was possible for
7416 RSA_get_default_method() to return NULL.
7419 *) Added native name translation to the existing DSO code
7420 that will convert (if the flag to do so is set) filenames
7421 that are sufficiently small and have no path information
7422 into a canonical native form. Eg. "blah" converted to
7423 "libblah.so" or "blah.dll" etc.
7426 *) New function ERR_error_string_n(e, buf, len) which is like
7427 ERR_error_string(e, buf), but writes at most 'len' bytes
7428 including the 0 terminator. For ERR_error_string_n, 'buf'
7430 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7432 *) CONF library reworked to become more general. A new CONF
7433 configuration file reader "class" is implemented as well as a
7434 new functions (NCONF_*, for "New CONF") to handle it. The now
7435 old CONF_* functions are still there, but are reimplemented to
7436 work in terms of the new functions. Also, a set of functions
7437 to handle the internal storage of the configuration data is
7438 provided to make it easier to write new configuration file
7439 reader "classes" (I can definitely see something reading a
7440 configuration file in XML format, for example), called _CONF_*,
7441 or "the configuration storage API"...
7443 The new configuration file reading functions are:
7445 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7446 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7448 NCONF_default, NCONF_WIN32
7450 NCONF_dump_fp, NCONF_dump_bio
7452 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7453 NCONF_new creates a new CONF object. This works in the same way
7454 as other interfaces in OpenSSL, like the BIO interface.
7455 NCONF_dump_* dump the internal storage of the configuration file,
7456 which is useful for debugging. All other functions take the same
7457 arguments as the old CONF_* functions wth the exception of the
7458 first that must be a `CONF *' instead of a `LHASH *'.
7460 To make it easer to use the new classes with the old CONF_* functions,
7461 the function CONF_set_default_method is provided.
7464 *) Add '-tls1' option to 'openssl ciphers', which was already
7465 mentioned in the documentation but had not been implemented.
7466 (This option is not yet really useful because even the additional
7467 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7470 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7471 OpenSSL-based applications) load shared libraries and bind to
7472 them in a portable way.
7473 [Geoff Thorpe, with contributions from Richard Levitte]
7475 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7477 *) Make sure _lrotl and _lrotr are only used with MSVC.
7479 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7480 (the default implementation of RAND_status).
7482 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7483 to '-clrext' (= clear extensions), as intended and documented.
7484 [Bodo Moeller; inconsistency pointed out by Michael Attili
7487 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7488 was larger than the MD block size.
7489 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7491 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7492 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7493 using the passed key: if the passed key was a private key the result
7494 of X509_print(), for example, would be to print out all the private key
7498 *) des_quad_cksum() byte order bug fix.
7499 [Ulf Möller, using the problem description in krb4-0.9.7, where
7500 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7502 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7504 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7506 *) For easily testing in shell scripts whether some command
7507 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7508 returns with exit code 0 iff no command of the given name is available.
7509 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7510 the output goes to stdout and nothing is printed to stderr.
7511 Additional arguments are always ignored.
7513 Since for each cipher there is a command of the same name,
7514 the 'no-cipher' compilation switches can be tested this way.
7516 ('openssl no-XXX' is not able to detect pseudo-commands such
7517 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7520 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7523 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7524 is set; it will be thrown away anyway because each handshake creates
7526 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7527 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7528 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7529 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7532 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7533 'Q' and 'R' lose their special meanings (quit/renegotiate).
7534 This is part of what -quiet does; unlike -quiet, -ign_eof
7535 does not suppress any output.
7538 *) Add compatibility options to the purpose and trust code. The
7539 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7540 accepts a certificate or CA, this was the previous behaviour,
7541 with all the associated security issues.
7543 X509_TRUST_COMPAT is the old trust behaviour: only and
7544 automatically trust self signed roots in certificate store. A
7545 new trust setting X509_TRUST_DEFAULT is used to specify that
7546 a purpose has no associated trust setting and it should instead
7547 use the value in the default purpose.
7550 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7551 and fix a memory leak.
7554 *) In util/mkerr.pl (which implements 'make errors'), preserve
7555 reason strings from the previous version of the .c file, as
7556 the default to have only downcase letters (and digits) in
7557 automatically generated reasons codes is not always appropriate.
7560 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7561 using strerror. Previously, ERR_reason_error_string() returned
7562 library names as reason strings for SYSerr; but SYSerr is a special
7563 case where small numbers are errno values, not library numbers.
7566 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7567 converts DSA parameters into DH parameters. (When creating parameters,
7568 DSA_generate_parameters is used.)
7571 *) Include 'length' (recommended exponent length) in C code generated
7572 by 'openssl dhparam -C'.
7575 *) The second argument to set_label in perlasm was already being used
7576 so couldn't be used as a "file scope" flag. Moved to third argument
7580 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7581 instead of RAND_bytes for encryption IVs and salts.
7584 *) Include RAND_status() into RAND_METHOD instead of implementing
7585 it only for md_rand.c Otherwise replacing the PRNG by calling
7586 RAND_set_rand_method would be impossible.
7589 *) Don't let DSA_generate_key() enter an infinite loop if the random
7590 number generation fails.
7593 *) New 'rand' application for creating pseudo-random output.
7596 *) Added configuration support for Linux/IA64
7597 [Rolf Haberrecker <rolf@suse.de>]
7599 *) Assembler module support for Mingw32.
7602 *) Shared library support for HPUX (in shlib/).
7603 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7605 *) Shared library support for Solaris gcc.
7606 [Lutz Behnke <behnke@trustcenter.de>]
7608 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7610 *) PKCS7_encrypt() was adding text MIME headers twice because they
7611 were added manually and by SMIME_crlf_copy().
7614 *) In bntest.c don't call BN_rand with zero bits argument.
7615 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7617 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7618 case was implemented. This caused BN_div_recp() to fail occasionally.
7621 *) Add an optional second argument to the set_label() in the perl
7622 assembly language builder. If this argument exists and is set
7623 to 1 it signals that the assembler should use a symbol whose
7624 scope is the entire file, not just the current function. This
7625 is needed with MASM which uses the format label:: for this scope.
7626 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7628 *) Change the ASN1 types so they are typedefs by default. Before
7629 almost all types were #define'd to ASN1_STRING which was causing
7630 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7634 *) Change names of new functions to the new get1/get0 naming
7635 convention: After 'get1', the caller owns a reference count
7636 and has to call ..._free; 'get0' returns a pointer to some
7637 data structure without incrementing reference counters.
7638 (Some of the existing 'get' functions increment a reference
7639 counter, some don't.)
7640 Similarly, 'set1' and 'add1' functions increase reference
7641 counters or duplicate objects.
7644 *) Allow for the possibility of temp RSA key generation failure:
7645 the code used to assume it always worked and crashed on failure.
7648 *) Fix potential buffer overrun problem in BIO_printf().
7649 [Ulf Möller, using public domain code by Patrick Powell; problem
7650 pointed out by David Sacerdote <das33@cornell.edu>]
7652 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7653 RAND_egd() and RAND_status(). In the command line application,
7654 the EGD socket can be specified like a seed file using RANDFILE
7658 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7659 Some CAs (e.g. Verisign) distribute certificates in this form.
7662 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7663 list to exclude them. This means that no special compilation option
7664 is needed to use anonymous DH: it just needs to be included in the
7668 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7669 EVP_MD_type. The old functionality is available in a new macro called
7670 EVP_MD_md(). Change code that uses it and update docs.
7673 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7674 where the 'void *' argument is replaced by a function pointer argument.
7675 Previously 'void *' was abused to point to functions, which works on
7676 many platforms, but is not correct. As these functions are usually
7677 called by macros defined in OpenSSL header files, most source code
7678 should work without changes.
7681 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7682 sections with information on -D... compiler switches used for
7683 compiling the library so that applications can see them. To enable
7684 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7685 must be defined. E.g.,
7686 #define OPENSSL_ALGORITHM_DEFINES
7687 #include <openssl/opensslconf.h>
7688 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7689 [Richard Levitte, Ulf and Bodo Möller]
7691 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7695 *) Change the 'other' type in certificate aux info to a STACK_OF
7696 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7697 the required ASN1 format: arbitrary types determined by an OID.
7700 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7701 argument to 'req'. This is not because the function is newer or
7702 better than others it just uses the work 'NEW' in the certificate
7703 request header lines. Some software needs this.
7706 *) Reorganise password command line arguments: now passwords can be
7707 obtained from various sources. Delete the PEM_cb function and make
7708 it the default behaviour: i.e. if the callback is NULL and the
7709 usrdata argument is not NULL interpret it as a null terminated pass
7710 phrase. If usrdata and the callback are NULL then the pass phrase
7711 is prompted for as usual.
7714 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7715 the support is automatically enabled. The resulting binaries will
7716 autodetect the card and use it if present.
7717 [Ben Laurie and Compaq Inc.]
7719 *) Work around for Netscape hang bug. This sends certificate request
7720 and server done in one record. Since this is perfectly legal in the
7721 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7722 the bugs/SSLv3 entry for more info.
7725 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7728 *) Add -rand argument to smime and pkcs12 applications and read/write
7732 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7735 *) Add command line password options to the remaining applications.
7738 *) Bug fix for BN_div_recp() for numerators with an even number of
7742 *) More tests in bntest.c, and changed test_bn output.
7745 *) ./config recognizes MacOS X now.
7748 *) Bug fix for BN_div() when the first words of num and divsor are
7749 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7752 *) Add support for various broken PKCS#8 formats, and command line
7753 options to produce them.
7756 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7757 get temporary BIGNUMs from a BN_CTX.
7760 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7764 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7765 include a #define from the old name to the new. The original intent
7766 was that statically linked binaries could for example just call
7767 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7768 link with digests. This never worked becayse SSLeay_add_all_digests()
7769 and SSLeay_add_all_ciphers() were in the same source file so calling
7770 one would link with the other. They are now in separate source files.
7773 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7776 *) Use a less unusual form of the Miller-Rabin primality test (it used
7777 a binary algorithm for exponentiation integrated into the Miller-Rabin
7778 loop, our standard modexp algorithms are faster).
7781 *) Support for the EBCDIC character set completed.
7782 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7784 *) Source code cleanups: use const where appropriate, eliminate casts,
7785 use void * instead of char * in lhash.
7788 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7789 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7790 this the server could overwrite ephemeral keys that the client
7794 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7795 using 50 iterations of the Rabin-Miller test.
7797 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7798 iterations of the Rabin-Miller test as required by the appendix
7799 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7800 As BN_is_prime_fasttest includes trial division, DSA parameter
7801 generation becomes much faster.
7803 This implies a change for the callback functions in DSA_is_prime
7804 and DSA_generate_parameters: The callback function is called once
7805 for each positive witness in the Rabin-Miller test, not just
7806 occasionally in the inner loop; and the parameters to the
7807 callback function now provide an iteration count for the outer
7808 loop rather than for the current invocation of the inner loop.
7809 DSA_generate_parameters additionally can call the callback
7810 function with an 'iteration count' of -1, meaning that a
7811 candidate has passed the trial division test (when q is generated
7812 from an application-provided seed, trial division is skipped).
7815 *) New function BN_is_prime_fasttest that optionally does trial
7816 division before starting the Rabin-Miller test and has
7817 an additional BN_CTX * argument (whereas BN_is_prime always
7818 has to allocate at least one BN_CTX).
7819 'callback(1, -1, cb_arg)' is called when a number has passed the
7820 trial division stage.
7823 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7827 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7830 *) New function BN_pseudo_rand().
7833 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7834 bignum version of BN_from_montgomery() with the working code from
7835 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7839 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7840 made it impossible to use the same SSL_SESSION data structure in
7841 SSL2 clients in multiple threads.
7844 *) The return value of RAND_load_file() no longer counts bytes obtained
7845 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7846 to seed the PRNG (previously an explicit byte count was required).
7847 [Ulf Möller, Bodo Möller]
7849 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7850 used (char *) instead of (void *) and had casts all over the place.
7853 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7856 *) Retain source code compatibility for BN_prime_checks macro:
7857 BN_is_prime(..., BN_prime_checks, ...) now uses
7858 BN_prime_checks_for_size to determine the appropriate number of
7859 Rabin-Miller iterations.
7862 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7863 DH_CHECK_P_NOT_SAFE_PRIME.
7864 (Check if this is true? OpenPGP calls them "strong".)
7867 *) Merge the functionality of "dh" and "gendh" programs into a new program
7868 "dhparam". The old programs are retained for now but will handle DH keys
7869 (instead of parameters) in future.
7872 *) Make the ciphers, s_server and s_client programs check the return values
7873 when a new cipher list is set.
7876 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7877 ciphers. Before when the 56bit ciphers were enabled the sorting was
7880 The syntax for the cipher sorting has been extended to support sorting by
7881 cipher-strength (using the strength_bits hard coded in the tables).
7882 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7884 Fix a bug in the cipher-command parser: when supplying a cipher command
7885 string with an "undefined" symbol (neither command nor alphanumeric
7886 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7887 an error is flagged.
7889 Due to the strength-sorting extension, the code of the
7890 ssl_create_cipher_list() function was completely rearranged. I hope that
7891 the readability was also increased :-)
7892 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7894 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7895 for the first serial number and places 2 in the serial number file. This
7896 avoids problems when the root CA is created with serial number zero and
7897 the first user certificate has the same issuer name and serial number
7901 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7902 the new code. Add documentation for this stuff.
7905 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
7906 X509_*() to X509at_*() on the grounds that they don't handle X509
7907 structures and behave in an analagous way to the X509v3 functions:
7908 they shouldn't be called directly but wrapper functions should be used
7911 So we also now have some wrapper functions that call the X509at functions
7912 when passed certificate requests. (TO DO: similar things can be done with
7913 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
7914 things. Some of these need some d2i or i2d and print functionality
7915 because they handle more complex structures.)
7918 *) Add missing #ifndefs that caused missing symbols when building libssl
7919 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
7920 NO_RSA in ssl/s2*.c.
7921 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
7923 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
7924 has a return value which indicates the quality of the random data
7925 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
7926 error queue. New function RAND_pseudo_bytes() generates output that is
7927 guaranteed to be unique but not unpredictable. RAND_add is like
7928 RAND_seed, but takes an extra argument for an entropy estimate
7929 (RAND_seed always assumes full entropy).
7932 *) Do more iterations of Rabin-Miller probable prime test (specifically,
7933 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
7934 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
7935 in crypto/bn/bn_prime.c for the complete table). This guarantees a
7936 false-positive rate of at most 2^-80 for random input.
7939 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
7942 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
7943 in the 0.9.5 release), this returns the chain
7944 from an X509_CTX structure with a dup of the stack and all
7945 the X509 reference counts upped: so the stack will exist
7946 after X509_CTX_cleanup() has been called. Modify pkcs12.c
7949 Also make SSL_SESSION_print() print out the verify return
7953 *) Add manpage for the pkcs12 command. Also change the default
7954 behaviour so MAC iteration counts are used unless the new
7955 -nomaciter option is used. This improves file security and
7956 only older versions of MSIE (4.0 for example) need it.
7959 *) Honor the no-xxx Configure options when creating .DEF files.
7962 *) Add PKCS#10 attributes to field table: challengePassword,
7963 unstructuredName and unstructuredAddress. These are taken from
7964 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
7965 international characters are used.
7967 More changes to X509_ATTRIBUTE code: allow the setting of types
7968 based on strings. Remove the 'loc' parameter when adding
7969 attributes because these will be a SET OF encoding which is sorted
7973 *) Initial changes to the 'req' utility to allow request generation
7974 automation. This will allow an application to just generate a template
7975 file containing all the field values and have req construct the
7978 Initial support for X509_ATTRIBUTE handling. Stacks of these are
7979 used all over the place including certificate requests and PKCS#7
7980 structures. They are currently handled manually where necessary with
7981 some primitive wrappers for PKCS#7. The new functions behave in a
7982 manner analogous to the X509 extension functions: they allow
7983 attributes to be looked up by NID and added.
7985 Later something similar to the X509V3 code would be desirable to
7986 automatically handle the encoding, decoding and printing of the
7987 more complex types. The string types like challengePassword can
7988 be handled by the string table functions.
7990 Also modified the multi byte string table handling. Now there is
7991 a 'global mask' which masks out certain types. The table itself
7992 can use the flag STABLE_NO_MASK to ignore the mask setting: this
7993 is useful when for example there is only one permissible type
7994 (as in countryName) and using the mask might result in no valid
7998 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
7999 SSL_get_peer_finished to allow applications to obtain the latest
8000 Finished messages sent to the peer or expected from the peer,
8001 respectively. (SSL_get_peer_finished is usually the Finished message
8002 actually received from the peer, otherwise the protocol will be aborted.)
8004 As the Finished message are message digests of the complete handshake
8005 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
8006 be used for external authentication procedures when the authentication
8007 provided by SSL/TLS is not desired or is not enough.
8010 *) Enhanced support for Alpha Linux is added. Now ./config checks if
8011 the host supports BWX extension and if Compaq C is present on the
8012 $PATH. Just exploiting of the BWX extension results in 20-30%
8013 performance kick for some algorithms, e.g. DES and RC4 to mention
8014 a couple. Compaq C in turn generates ~20% faster code for MD5 and
8018 *) Add support for MS "fast SGC". This is arguably a violation of the
8019 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
8020 weak crypto and after checking the certificate is SGC a second one
8021 with strong crypto. MS SGC stops the first handshake after receiving
8022 the server certificate message and sends a second client hello. Since
8023 a server will typically do all the time consuming operations before
8024 expecting any further messages from the client (server key exchange
8025 is the most expensive) there is little difference between the two.
8027 To get OpenSSL to support MS SGC we have to permit a second client
8028 hello message after we have sent server done. In addition we have to
8029 reset the MAC if we do get this second client hello.
8032 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
8033 if a DER encoded private key is RSA or DSA traditional format. Changed
8034 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
8035 format DER encoded private key. Newer code should use PKCS#8 format which
8036 has the key type encoded in the ASN1 structure. Added DER private key
8037 support to pkcs8 application.
8040 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8041 ciphersuites has been selected (as required by the SSL 3/TLS 1
8042 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8043 is set, we interpret this as a request to violate the specification
8044 (the worst that can happen is a handshake failure, and 'correct'
8045 behaviour would result in a handshake failure anyway).
8048 *) In SSL_CTX_add_session, take into account that there might be multiple
8049 SSL_SESSION structures with the same session ID (e.g. when two threads
8050 concurrently obtain them from an external cache).
8051 The internal cache can handle only one SSL_SESSION with a given ID,
8052 so if there's a conflict, we now throw out the old one to achieve
8056 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8057 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8058 some routines that use cipher OIDs: some ciphers do not have OIDs
8059 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8063 *) Simplify the trust setting structure and code. Now we just have
8064 two sequences of OIDs for trusted and rejected settings. These will
8065 typically have values the same as the extended key usage extension
8066 and any application specific purposes.
8068 The trust checking code now has a default behaviour: it will just
8069 check for an object with the same NID as the passed id. Functions can
8070 be provided to override either the default behaviour or the behaviour
8071 for a given id. SSL client, server and email already have functions
8072 in place for compatibility: they check the NID and also return "trusted"
8073 if the certificate is self signed.
8076 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8077 traditional format into an EVP_PKEY structure.
8080 *) Add a password callback function PEM_cb() which either prompts for
8081 a password if usr_data is NULL or otherwise assumes it is a null
8082 terminated password. Allow passwords to be passed on command line
8083 environment or config files in a few more utilities.
8086 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8087 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8088 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8089 Update documentation.
8092 *) Support for ASN1 "NULL" type. This could be handled before by using
8093 ASN1_TYPE but there wasn't any function that would try to read a NULL
8094 and produce an error if it couldn't. For compatibility we also have
8095 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8096 don't allocate anything because they don't need to.
8099 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8101 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8103 *) Rebuild of the memory allocation routines used by OpenSSL code and
8104 possibly others as well. The purpose is to make an interface that
8105 provide hooks so anyone can build a separate set of allocation and
8106 deallocation routines to be used by OpenSSL, for example memory
8107 pool implementations, or something else, which was previously hard
8108 since Malloc(), Realloc() and Free() were defined as macros having
8109 the values malloc, realloc and free, respectively (except for Win32
8110 compilations). The same is provided for memory debugging code.
8111 OpenSSL already comes with functionality to find memory leaks, but
8112 this gives people a chance to debug other memory problems.
8114 With these changes, a new set of functions and macros have appeared:
8116 CRYPTO_set_mem_debug_functions() [F]
8117 CRYPTO_get_mem_debug_functions() [F]
8118 CRYPTO_dbg_set_options() [F]
8119 CRYPTO_dbg_get_options() [F]
8120 CRYPTO_malloc_debug_init() [M]
8122 The memory debug functions are NULL by default, unless the library
8123 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8124 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8125 gives the standard debugging functions that come with OpenSSL) or
8126 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8127 provided by the library user) must be used. When the standard
8128 debugging functions are used, CRYPTO_dbg_set_options can be used to
8129 request additional information:
8130 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8131 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8133 Also, things like CRYPTO_set_mem_functions will always give the
8134 expected result (the new set of functions is used for allocation
8135 and deallocation) at all times, regardless of platform and compiler
8138 To finish it up, some functions that were never use in any other
8139 way than through macros have a new API and new semantic:
8142 CRYPTO_dbg_realloc()
8145 All macros of value have retained their old syntax.
8146 [Richard Levitte and Bodo Moeller]
8148 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8149 ordering of SMIMECapabilities wasn't in "strength order" and there
8150 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8154 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8155 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8156 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
8158 *) Merge in my S/MIME library for OpenSSL. This provides a simple
8159 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
8160 functionality to handle multipart/signed properly) and a utility
8161 called 'smime' to call all this stuff. This is based on code I
8162 originally wrote for Celo who have kindly allowed it to be
8163 included in OpenSSL.
8166 *) Add variants des_set_key_checked and des_set_key_unchecked of
8167 des_set_key (aka des_key_sched). Global variable des_check_key
8168 decides which of these is called by des_set_key; this way
8169 des_check_key behaves as it always did, but applications and
8170 the library itself, which was buggy for des_check_key == 1,
8171 have a cleaner way to pick the version they need.
8174 *) New function PKCS12_newpass() which changes the password of a
8178 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
8179 dynamic mix. In both cases the ids can be used as an index into the
8180 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
8181 functions so they accept a list of the field values and the
8182 application doesn't need to directly manipulate the X509_TRUST
8186 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
8190 *) Modify the way the V3 extension code looks up extensions. This now
8191 works in a similar way to the object code: we have some "standard"
8192 extensions in a static table which is searched with OBJ_bsearch()
8193 and the application can add dynamic ones if needed. The file
8194 crypto/x509v3/ext_dat.h now has the info: this file needs to be
8195 updated whenever a new extension is added to the core code and kept
8196 in ext_nid order. There is a simple program 'tabtest.c' which checks
8197 this. New extensions are not added too often so this file can readily
8198 be maintained manually.
8200 There are two big advantages in doing things this way. The extensions
8201 can be looked up immediately and no longer need to be "added" using
8202 X509V3_add_standard_extensions(): this function now does nothing.
8203 [Side note: I get *lots* of email saying the extension code doesn't
8204 work because people forget to call this function]
8205 Also no dynamic allocation is done unless new extensions are added:
8206 so if we don't add custom extensions there is no need to call
8207 X509V3_EXT_cleanup().
8210 *) Modify enc utility's salting as follows: make salting the default. Add a
8211 magic header, so unsalted files fail gracefully instead of just decrypting
8212 to garbage. This is because not salting is a big security hole, so people
8213 should be discouraged from doing it.
8216 *) Fixes and enhancements to the 'x509' utility. It allowed a message
8217 digest to be passed on the command line but it only used this
8218 parameter when signing a certificate. Modified so all relevant
8219 operations are affected by the digest parameter including the
8220 -fingerprint and -x509toreq options. Also -x509toreq choked if a
8221 DSA key was used because it didn't fix the digest.
8224 *) Initial certificate chain verify code. Currently tests the untrusted
8225 certificates for consistency with the verify purpose (which is set
8226 when the X509_STORE_CTX structure is set up) and checks the pathlength.
8228 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
8229 this is because it will reject chains with invalid extensions whereas
8230 every previous version of OpenSSL and SSLeay made no checks at all.
8232 Trust code: checks the root CA for the relevant trust settings. Trust
8233 settings have an initial value consistent with the verify purpose: e.g.
8234 if the verify purpose is for SSL client use it expects the CA to be
8235 trusted for SSL client use. However the default value can be changed to
8236 permit custom trust settings: one example of this would be to only trust
8237 certificates from a specific "secure" set of CAs.
8239 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
8240 which should be used for version portability: especially since the
8241 verify structure is likely to change more often now.
8243 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
8244 to set them. If not set then assume SSL clients will verify SSL servers
8247 Two new options to the verify program: -untrusted allows a set of
8248 untrusted certificates to be passed in and -purpose which sets the
8249 intended purpose of the certificate. If a purpose is set then the
8250 new chain verify code is used to check extension consistency.
8253 *) Support for the authority information access extension.
8256 *) Modify RSA and DSA PEM read routines to transparently handle
8257 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
8258 public keys in a format compatible with certificate
8259 SubjectPublicKeyInfo structures. Unfortunately there were already
8260 functions called *_PublicKey_* which used various odd formats so
8261 these are retained for compatibility: however the DSA variants were
8262 never in a public release so they have been deleted. Changed dsa/rsa
8263 utilities to handle the new format: note no releases ever handled public
8264 keys so we should be OK.
8266 The primary motivation for this change is to avoid the same fiasco
8267 that dogs private keys: there are several incompatible private key
8268 formats some of which are standard and some OpenSSL specific and
8269 require various evil hacks to allow partial transparent handling and
8270 even then it doesn't work with DER formats. Given the option anything
8271 other than PKCS#8 should be dumped: but the other formats have to
8272 stay in the name of compatibility.
8274 With public keys and the benefit of hindsight one standard format
8275 is used which works with EVP_PKEY, RSA or DSA structures: though
8276 it clearly returns an error if you try to read the wrong kind of key.
8278 Added a -pubkey option to the 'x509' utility to output the public key.
8279 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
8280 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
8281 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
8282 that do the same as the EVP_PKEY_assign_*() except they up the
8283 reference count of the added key (they don't "swallow" the
8287 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
8288 CRLs would fail if the file contained no certificates or no CRLs:
8289 added a new function to read in both types and return the number
8290 read: this means that if none are read it will be an error. The
8291 DER versions of the certificate and CRL reader would always fail
8292 because it isn't possible to mix certificates and CRLs in DER format
8293 without choking one or the other routine. Changed this to just read
8294 a certificate: this is the best we can do. Also modified the code
8295 in apps/verify.c to take notice of return codes: it was previously
8296 attempting to read in certificates from NULL pointers and ignoring
8297 any errors: this is one reason why the cert and CRL reader seemed
8298 to work. It doesn't check return codes from the default certificate
8299 routines: these may well fail if the certificates aren't installed.
8302 *) Code to support otherName option in GeneralName.
8305 *) First update to verify code. Change the verify utility
8306 so it warns if it is passed a self signed certificate:
8307 for consistency with the normal behaviour. X509_verify
8308 has been modified to it will now verify a self signed
8309 certificate if *exactly* the same certificate appears
8310 in the store: it was previously impossible to trust a
8311 single self signed certificate. This means that:
8312 openssl verify ss.pem
8313 now gives a warning about a self signed certificate but
8314 openssl verify -CAfile ss.pem ss.pem
8318 *) For servers, store verify_result in SSL_SESSION data structure
8319 (and add it to external session representation).
8320 This is needed when client certificate verifications fails,
8321 but an application-provided verification callback (set by
8322 SSL_CTX_set_cert_verify_callback) allows accepting the session
8323 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
8324 but returns 1): When the session is reused, we have to set
8325 ssl->verify_result to the appropriate error code to avoid
8327 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
8329 *) Fix a bug in the new PKCS#7 code: it didn't consider the
8330 case in PKCS7_dataInit() where the signed PKCS7 structure
8331 didn't contain any existing data because it was being created.
8332 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
8334 *) Add a salt to the key derivation routines in enc.c. This
8335 forms the first 8 bytes of the encrypted file. Also add a
8336 -S option to allow a salt to be input on the command line.
8339 *) New function X509_cmp(). Oddly enough there wasn't a function
8340 to compare two certificates. We do this by working out the SHA1
8341 hash and comparing that. X509_cmp() will be needed by the trust
8345 *) SSL_get1_session() is like SSL_get_session(), but increments
8346 the reference count in the SSL_SESSION returned.
8347 [Geoff Thorpe <geoff@eu.c2.net>]
8349 *) Fix for 'req': it was adding a null to request attributes.
8350 Also change the X509_LOOKUP and X509_INFO code to handle
8351 certificate auxiliary information.
8354 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8358 *) Add the possibility to add extra information to the memory leak
8359 detecting output, to form tracebacks, showing from where each
8360 allocation was originated: CRYPTO_push_info("constant string") adds
8361 the string plus current file name and line number to a per-thread
8362 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8363 is like calling CYRPTO_pop_info() until the stack is empty.
8364 Also updated memory leak detection code to be multi-thread-safe.
8367 *) Add options -text and -noout to pkcs7 utility and delete the
8368 encryption options which never did anything. Update docs.
8371 *) Add options to some of the utilities to allow the pass phrase
8372 to be included on either the command line (not recommended on
8373 OSes like Unix) or read from the environment. Update the
8374 manpages and fix a few bugs.
8377 *) Add a few manpages for some of the openssl commands.
8380 *) Fix the -revoke option in ca. It was freeing up memory twice,
8381 leaking and not finding already revoked certificates.
8384 *) Extensive changes to support certificate auxiliary information.
8385 This involves the use of X509_CERT_AUX structure and X509_AUX
8386 functions. An X509_AUX function such as PEM_read_X509_AUX()
8387 can still read in a certificate file in the usual way but it
8388 will also read in any additional "auxiliary information". By
8389 doing things this way a fair degree of compatibility can be
8390 retained: existing certificates can have this information added
8391 using the new 'x509' options.
8393 Current auxiliary information includes an "alias" and some trust
8394 settings. The trust settings will ultimately be used in enhanced
8395 certificate chain verification routines: currently a certificate
8396 can only be trusted if it is self signed and then it is trusted
8400 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8401 The problem was that one of the replacement routines had not been working
8402 since SSLeay releases. For now the offending routine has been replaced
8403 with non-optimised assembler. Even so, this now gives around 95%
8404 performance improvement for 1024 bit RSA signs.
8407 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8408 handling. Most clients have the effective key size in bits equal to
8409 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8410 A few however don't do this and instead use the size of the decrypted key
8411 to determine the RC2 key length and the AlgorithmIdentifier to determine
8412 the effective key length. In this case the effective key length can still
8413 be 40 bits but the key length can be 168 bits for example. This is fixed
8414 by manually forcing an RC2 key into the EVP_PKEY structure because the
8415 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8416 the key length and effective key length are equal.
8419 *) Add a bunch of functions that should simplify the creation of
8420 X509_NAME structures. Now you should be able to do:
8421 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8422 and have it automatically work out the correct field type and fill in
8423 the structures. The more adventurous can try:
8424 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8425 and it will (hopefully) work out the correct multibyte encoding.
8428 *) Change the 'req' utility to use the new field handling and multibyte
8429 copy routines. Before the DN field creation was handled in an ad hoc
8430 way in req, ca, and x509 which was rather broken and didn't support
8431 BMPStrings or UTF8Strings. Since some software doesn't implement
8432 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8433 using the dirstring_type option. See the new comment in the default
8434 openssl.cnf for more info.
8437 *) Make crypto/rand/md_rand.c more robust:
8438 - Assure unique random numbers after fork().
8439 - Make sure that concurrent threads access the global counter and
8440 md serializably so that we never lose entropy in them
8441 or use exactly the same state in multiple threads.
8442 Access to the large state is not always serializable because
8443 the additional locking could be a performance killer, and
8444 md should be large enough anyway.
8447 *) New file apps/app_rand.c with commonly needed functionality
8448 for handling the random seed file.
8450 Use the random seed file in some applications that previously did not:
8452 dsaparam -genkey (which also ignored its '-rand' option),
8455 x509 (when signing).
8456 Except on systems with /dev/urandom, it is crucial to have a random
8457 seed file at least for key creation, DSA signing, and for DH exchanges;
8458 for RSA signatures we could do without one.
8460 gendh and gendsa (unlike genrsa) used to read only the first byte
8461 of each file listed in the '-rand' option. The function as previously
8462 found in genrsa is now in app_rand.c and is used by all programs
8463 that support '-rand'.
8466 *) In RAND_write_file, use mode 0600 for creating files;
8467 don't just chmod when it may be too late.
8470 *) Report an error from X509_STORE_load_locations
8471 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8474 *) New function ASN1_mbstring_copy() this copies a string in either
8475 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8476 into an ASN1_STRING type. A mask of permissible types is passed
8477 and it chooses the "minimal" type to use or an error if not type
8481 *) Add function equivalents to the various macros in asn1.h. The old
8482 macros are retained with an M_ prefix. Code inside the library can
8483 use the M_ macros. External code (including the openssl utility)
8484 should *NOT* in order to be "shared library friendly".
8487 *) Add various functions that can check a certificate's extensions
8488 to see if it usable for various purposes such as SSL client,
8489 server or S/MIME and CAs of these types. This is currently
8490 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8491 verification. Also added a -purpose flag to x509 utility to
8492 print out all the purposes.
8495 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8499 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8500 for, obtain and decode and extension and obtain its critical flag.
8501 This allows all the necessary extension code to be handled in a
8502 single function call.
8505 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8506 platforms. See crypto/rc4/rc4_enc.c for further details.
8509 *) New -noout option to asn1parse. This causes no output to be produced
8510 its main use is when combined with -strparse and -out to extract data
8511 from a file (which may not be in ASN.1 format).
8514 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8515 when producing the local key id.
8516 [Richard Levitte <levitte@stacken.kth.se>]
8518 *) New option -dhparam in s_server. This allows a DH parameter file to be
8519 stated explicitly. If it is not stated then it tries the first server
8520 certificate file. The previous behaviour hard coded the filename
8524 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8525 a public key to be input or output. For example:
8526 openssl rsa -in key.pem -pubout -out pubkey.pem
8527 Also added necessary DSA public key functions to handle this.
8530 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8531 in the message. This was handled by allowing
8532 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8533 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8535 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8536 to the end of the strings whereas this didn't. This would cause problems
8537 if strings read with d2i_ASN1_bytes() were later modified.
8538 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8540 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8541 data and it contains EOF it will end up returning an error. This is
8542 caused by input 46 bytes long. The cause is due to the way base64
8543 BIOs find the start of base64 encoded data. They do this by trying a
8544 trial decode on each line until they find one that works. When they
8545 do a flag is set and it starts again knowing it can pass all the
8546 data directly through the decoder. Unfortunately it doesn't reset
8547 the context it uses. This means that if EOF is reached an attempt
8548 is made to pass two EOFs through the context and this causes the
8549 resulting error. This can also cause other problems as well. As is
8550 usual with these problems it takes *ages* to find and the fix is
8551 trivial: move one line.
8552 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8554 *) Ugly workaround to get s_client and s_server working under Windows. The
8555 old code wouldn't work because it needed to select() on sockets and the
8556 tty (for keypresses and to see if data could be written). Win32 only
8557 supports select() on sockets so we select() with a 1s timeout on the
8558 sockets and then see if any characters are waiting to be read, if none
8559 are present then we retry, we also assume we can always write data to
8560 the tty. This isn't nice because the code then blocks until we've
8561 received a complete line of data and it is effectively polling the
8562 keyboard at 1s intervals: however it's quite a bit better than not
8563 working at all :-) A dedicated Windows application might handle this
8564 with an event loop for example.
8567 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8568 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8569 will be called when RSA_sign() and RSA_verify() are used. This is useful
8570 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8571 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8572 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8573 This necessitated the support of an extra signature type NID_md5_sha1
8574 for SSL signatures and modifications to the SSL library to use it instead
8575 of calling RSA_public_decrypt() and RSA_private_encrypt().
8578 *) Add new -verify -CAfile and -CApath options to the crl program, these
8579 will lookup a CRL issuers certificate and verify the signature in a
8580 similar way to the verify program. Tidy up the crl program so it
8581 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8582 less strict. It will now permit CRL extensions even if it is not
8583 a V2 CRL: this will allow it to tolerate some broken CRLs.
8586 *) Initialize all non-automatic variables each time one of the openssl
8587 sub-programs is started (this is necessary as they may be started
8588 multiple times from the "OpenSSL>" prompt).
8589 [Lennart Bang, Bodo Moeller]
8591 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8592 removing all other RSA functionality (this is what NO_RSA does). This
8593 is so (for example) those in the US can disable those operations covered
8594 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8598 *) Non-copying interface to BIO pairs.
8599 (still largely untested)
8602 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8603 ASCII string. This was handled independently in various places before.
8606 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8607 UTF8 strings a character at a time.
8610 *) Use client_version from client hello to select the protocol
8611 (s23_srvr.c) and for RSA client key exchange verification
8612 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8615 *) Add various utility functions to handle SPKACs, these were previously
8616 handled by poking round in the structure internals. Added new function
8617 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8618 print, verify and generate SPKACs. Based on an original idea from
8619 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8622 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8625 *) Allow the config file extension section to be overwritten on the
8626 command line. Based on an original idea from Massimiliano Pala
8627 <madwolf@comune.modena.it>. The new option is called -extensions
8628 and can be applied to ca, req and x509. Also -reqexts to override
8629 the request extensions in req and -crlexts to override the crl extensions
8633 *) Add new feature to the SPKAC handling in ca. Now you can include
8634 the same field multiple times by preceding it by "XXXX." for example:
8637 this is the same syntax as used in the req config file.
8640 *) Allow certificate extensions to be added to certificate requests. These
8641 are specified in a 'req_extensions' option of the req section of the
8642 config file. They can be printed out with the -text option to req but
8643 are otherwise ignored at present.
8646 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8647 data read consists of only the final block it would not decrypted because
8648 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8649 A misplaced 'break' also meant the decrypted final block might not be
8650 copied until the next read.
8653 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8654 a few extra parameters to the DH structure: these will be useful if
8655 for example we want the value of 'q' or implement X9.42 DH.
8658 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8659 provides hooks that allow the default DSA functions or functions on a
8660 "per key" basis to be replaced. This allows hardware acceleration and
8661 hardware key storage to be handled without major modification to the
8662 library. Also added low level modexp hooks and CRYPTO_EX structure and
8663 associated functions.
8666 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8667 as "read only": it can't be written to and the buffer it points to will
8668 not be freed. Reading from a read only BIO is much more efficient than
8669 a normal memory BIO. This was added because there are several times when
8670 an area of memory needs to be read from a BIO. The previous method was
8671 to create a memory BIO and write the data to it, this results in two
8672 copies of the data and an O(n^2) reading algorithm. There is a new
8673 function BIO_new_mem_buf() which creates a read only memory BIO from
8674 an area of memory. Also modified the PKCS#7 routines to use read only
8678 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8679 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8680 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8681 but a retry condition occured while trying to read the rest.
8684 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8685 NID_pkcs7_encrypted by default: this was wrong since this should almost
8686 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8687 the encrypted data type: this is a more sensible place to put it and it
8688 allows the PKCS#12 code to be tidied up that duplicated this
8692 *) Changed obj_dat.pl script so it takes its input and output files on
8693 the command line. This should avoid shell escape redirection problems
8697 *) Initial support for certificate extension requests, these are included
8698 in things like Xenroll certificate requests. Included functions to allow
8699 extensions to be obtained and added.
8702 *) -crlf option to s_client and s_server for sending newlines as
8703 CRLF (as required by many protocols).
8706 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8708 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8709 [Ralf S. Engelschall]
8711 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8712 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8714 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8718 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8719 DH parameters/keys (q is lost during that conversion, but the resulting
8720 DH parameters contain its length).
8722 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8723 much faster than DH_generate_parameters (which creates parameters
8724 where p = 2*q + 1), and also the smaller q makes DH computations
8725 much more efficient (160-bit exponentiation instead of 1024-bit
8726 exponentiation); so this provides a convenient way to support DHE
8727 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8728 utter importance to use
8729 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8731 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8732 when such DH parameters are used, because otherwise small subgroup
8733 attacks may become possible!
8736 *) Avoid memory leak in i2d_DHparams.
8739 *) Allow the -k option to be used more than once in the enc program:
8740 this allows the same encrypted message to be read by multiple recipients.
8743 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8744 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8745 it will always use the numerical form of the OID, even if it has a short
8749 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8750 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8751 otherwise bn_mod_exp was called. In the case of hardware keys for example
8752 no private key components need be present and it might store extra data
8753 in the RSA structure, which cannot be accessed from bn_mod_exp.
8754 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8755 private key operations.
8758 *) Added support for SPARC Linux.
8761 *) pem_password_cb function type incompatibly changed from
8762 typedef int pem_password_cb(char *buf, int size, int rwflag);
8764 ....(char *buf, int size, int rwflag, void *userdata);
8765 so that applications can pass data to their callbacks:
8766 The PEM[_ASN1]_{read,write}... functions and macros now take an
8767 additional void * argument, which is just handed through whenever
8768 the password callback is called.
8769 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8771 New function SSL_CTX_set_default_passwd_cb_userdata.
8773 Compatibility note: As many C implementations push function arguments
8774 onto the stack in reverse order, the new library version is likely to
8775 interoperate with programs that have been compiled with the old
8776 pem_password_cb definition (PEM_whatever takes some data that
8777 happens to be on the stack as its last argument, and the callback
8778 just ignores this garbage); but there is no guarantee whatsoever that
8781 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8782 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8783 problems not only on Windows, but also on some Unix platforms.
8784 To avoid problematic command lines, these definitions are now in an
8785 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8786 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8789 *) MIPS III/IV assembler module is reimplemented.
8792 *) More DES library cleanups: remove references to srand/rand and
8793 delete an unused file.
8796 *) Add support for the the free Netwide assembler (NASM) under Win32,
8797 since not many people have MASM (ml) and it can be hard to obtain.
8798 This is currently experimental but it seems to work OK and pass all
8799 the tests. Check out INSTALL.W32 for info.
8802 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8803 without temporary keys kept an extra copy of the server key,
8804 and connections with temporary keys did not free everything in case
8808 *) New function RSA_check_key and new openssl rsa option -check
8809 for verifying the consistency of RSA keys.
8810 [Ulf Moeller, Bodo Moeller]
8812 *) Various changes to make Win32 compile work:
8813 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8814 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8815 comparison" warnings.
8816 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8819 *) Add a debugging option to PKCS#5 v2 key generation function: when
8820 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8821 derived keys are printed to stderr.
8824 *) Copy the flags in ASN1_STRING_dup().
8825 [Roman E. Pavlov <pre@mo.msk.ru>]
8827 *) The x509 application mishandled signing requests containing DSA
8828 keys when the signing key was also DSA and the parameters didn't match.
8830 It was supposed to omit the parameters when they matched the signing key:
8831 the verifying software was then supposed to automatically use the CA's
8832 parameters if they were absent from the end user certificate.
8834 Omitting parameters is no longer recommended. The test was also
8835 the wrong way round! This was probably due to unusual behaviour in
8836 EVP_cmp_parameters() which returns 1 if the parameters match.
8837 This meant that parameters were omitted when they *didn't* match and
8838 the certificate was useless. Certificates signed with 'ca' didn't have
8840 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8842 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8843 The interface is as follows:
8844 Applications can use
8845 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8846 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8847 "off" is now the default.
8848 The library internally uses
8849 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8850 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8851 to disable memory-checking temporarily.
8853 Some inconsistent states that previously were possible (and were
8854 even the default) are now avoided.
8856 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8857 with each memory chunk allocated; this is occasionally more helpful
8858 than just having a counter.
8860 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8862 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8866 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8867 which largely parallels "options", but is for changing API behaviour,
8868 whereas "options" are about protocol behaviour.
8869 Initial "mode" flags are:
8871 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8872 a single record has been written.
8873 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8874 retries use the same buffer location.
8875 (But all of the contents must be
8879 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8882 *) Fix problems with no-hmac etc.
8883 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8885 *) New functions RSA_get_default_method(), RSA_set_method() and
8886 RSA_get_method(). These allows replacement of RSA_METHODs without having
8887 to mess around with the internals of an RSA structure.
8890 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8891 Also really enable memory leak checks in openssl.c and in some
8893 [Chad C. Mulligan, Bodo Moeller]
8895 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8896 up the length of negative integers. This has now been simplified to just
8897 store the length when it is first determined and use it later, rather
8898 than trying to keep track of where data is copied and updating it to
8900 [Steve Henson, reported by Brien Wheeler
8901 <bwheeler@authentica-security.com>]
8903 *) Add a new function PKCS7_signatureVerify. This allows the verification
8904 of a PKCS#7 signature but with the signing certificate passed to the
8905 function itself. This contrasts with PKCS7_dataVerify which assumes the
8906 certificate is present in the PKCS#7 structure. This isn't always the
8907 case: certificates can be omitted from a PKCS#7 structure and be
8908 distributed by "out of band" means (such as a certificate database).
8911 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
8912 function prototypes in pem.h, also change util/mkdef.pl to add the
8913 necessary function names.
8916 *) mk1mf.pl (used by Windows builds) did not properly read the
8917 options set by Configure in the top level Makefile, and Configure
8918 was not even able to write more than one option correctly.
8919 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
8922 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
8923 file to be loaded from a BIO or FILE pointer. The BIO version will
8924 for example allow memory BIOs to contain config info.
8927 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
8928 Whoever hopes to achieve shared-library compatibility across versions
8929 must use this, not the compile-time macro.
8930 (Exercise 0.9.4: Which is the minimum library version required by
8932 Note: All this applies only to multi-threaded programs, others don't
8936 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
8937 through a BIO pair triggered the default case, i.e.
8938 SSLerr(...,SSL_R_UNKNOWN_STATE).
8941 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
8942 can use the SSL library even if none of the specific BIOs is
8946 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
8947 for the encoded length.
8948 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
8950 *) Add initial documentation of the X509V3 functions.
8953 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
8954 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
8955 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
8956 secure PKCS#8 private key format with a high iteration count.
8959 *) Fix determination of Perl interpreter: A perl or perl5
8960 _directory_ in $PATH was also accepted as the interpreter.
8961 [Ralf S. Engelschall]
8963 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
8964 wrong with it but it was very old and did things like calling
8965 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
8969 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
8970 to use the new extension code.
8973 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
8974 with macros. This should make it easier to change their form, add extra
8975 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
8979 *) Add to configuration table a new entry that can specify an alternative
8980 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
8981 according to Mark Crispin <MRC@Panda.COM>.
8985 *) DES CBC did not update the IV. Weird.
8988 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
8989 Changing the behaviour of the former might break existing programs --
8990 where IV updating is needed, des_ncbc_encrypt can be used.
8993 *) When bntest is run from "make test" it drives bc to check its
8994 calculations, as well as internally checking them. If an internal check
8995 fails, it needs to cause bc to give a non-zero result or make test carries
8996 on without noticing the failure. Fixed.
8999 *) DES library cleanups.
9002 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
9003 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
9004 ciphers. NOTE: although the key derivation function has been verified
9005 against some published test vectors it has not been extensively tested
9006 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
9010 *) Instead of "mkdir -p", which is not fully portable, use new
9011 Perl script "util/mkdir-p.pl".
9014 *) Rewrite the way password based encryption (PBE) is handled. It used to
9015 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
9016 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
9017 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
9018 the 'parameter' field of the AlgorithmIdentifier is passed to the
9019 underlying key generation function so it must do its own ASN1 parsing.
9020 This has also changed the EVP_PBE_CipherInit() function which now has a
9021 'parameter' argument instead of literal salt and iteration count values
9022 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
9025 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
9026 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
9027 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
9028 KEY" because this clashed with PKCS#8 unencrypted string. Since this
9029 value was just used as a "magic string" and not used directly its
9030 value doesn't matter.
9033 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9037 *) "linux-sparc64" configuration (ultrapenguin).
9038 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9039 "linux-sparc" configuration.
9040 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9042 *) config now generates no-xxx options for missing ciphers.
9045 *) Support the EBCDIC character set (work in progress).
9046 File ebcdic.c not yet included because it has a different license.
9047 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9049 *) Support BS2000/OSD-POSIX.
9050 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9052 *) Make callbacks for key generation use void * instead of char *.
9055 *) Make S/MIME samples compile (not yet tested).
9058 *) Additional typesafe stacks.
9061 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9065 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9067 *) New configuration variant "sco5-gcc".
9069 *) Updated some demos.
9070 [Sean O Riordain, Wade Scholine]
9072 *) Add missing BIO_free at exit of pkcs12 application.
9075 *) Fix memory leak in conf.c.
9078 *) Updates for Win32 to assembler version of MD5.
9081 *) Set #! path to perl in apps/der_chop to where we found it
9082 instead of using a fixed path.
9085 *) SHA library changes for irix64-mips4-cc.
9088 *) Improvements for VMS support.
9092 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9094 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9095 This also avoids the problems with SC4.2 and unpatched SC5.
9096 [Andy Polyakov <appro@fy.chalmers.se>]
9098 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9099 These are required because of the typesafe stack would otherwise break
9100 existing code. If old code used a structure member which used to be STACK
9101 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9102 sk_num or sk_value it would produce an error because the num, data members
9103 are not present in STACK_OF. Now it just produces a warning. sk_set
9104 replaces the old method of assigning a value to sk_value
9105 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9106 that does this will no longer work (and should use sk_set instead) but
9107 this could be regarded as a "questionable" behaviour anyway.
9110 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9111 correctly handle encrypted S/MIME data.
9114 *) Change type of various DES function arguments from des_cblock
9115 (which means, in function argument declarations, pointer to char)
9116 to des_cblock * (meaning pointer to array with 8 char elements),
9117 which allows the compiler to do more typechecking; it was like
9118 that back in SSLeay, but with lots of ugly casts.
9120 Introduce new type const_des_cblock.
9123 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9124 problems: find RecipientInfo structure that matches recipient certificate
9125 and initialise the ASN1 structures properly based on passed cipher.
9128 *) Belatedly make the BN tests actually check the results.
9131 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9132 to and from BNs: it was completely broken. New compilation option
9133 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9134 key elements as negative integers.
9137 *) Reorganize and speed up MD5.
9138 [Andy Polyakov <appro@fy.chalmers.se>]
9141 [Richard Levitte <richard@levitte.org>]
9143 *) New option -out to asn1parse to allow the parsed structure to be
9144 output to a file. This is most useful when combined with the -strparse
9145 option to examine the output of things like OCTET STRINGS.
9148 *) Make SSL library a little more fool-proof by not requiring any longer
9149 that SSL_set_{accept,connect}_state be called before
9150 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9151 in many applications because usually everything *appeared* to work as
9152 intended anyway -- now it really works as intended).
9155 *) Move openssl.cnf out of lib/.
9158 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
9159 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
9160 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
9161 [Ralf S. Engelschall]
9163 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
9164 handle PKCS#7 enveloped data properly.
9165 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
9167 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
9168 copying pointers. The cert_st handling is changed by this in
9169 various ways (and thus what used to be known as ctx->default_cert
9170 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
9171 any longer when s->cert does not give us what we need).
9172 ssl_cert_instantiate becomes obsolete by this change.
9173 As soon as we've got the new code right (possibly it already is?),
9174 we have solved a couple of bugs of the earlier code where s->cert
9175 was used as if it could not have been shared with other SSL structures.
9177 Note that using the SSL API in certain dirty ways now will result
9178 in different behaviour than observed with earlier library versions:
9179 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
9180 does not influence s as it used to.
9182 In order to clean up things more thoroughly, inside SSL_SESSION
9183 we don't use CERT any longer, but a new structure SESS_CERT
9184 that holds per-session data (if available); currently, this is
9185 the peer's certificate chain and, for clients, the server's certificate
9186 and temporary key. CERT holds only those values that can have
9187 meaningful defaults in an SSL_CTX.
9190 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
9191 from the internal representation. Various PKCS#7 fixes: remove some
9192 evil casts and set the enc_dig_alg field properly based on the signing
9196 *) Allow PKCS#12 password to be set from the command line or the
9197 environment. Let 'ca' get its config file name from the environment
9198 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
9202 *) Allow certificate policies extension to use an IA5STRING for the
9203 organization field. This is contrary to the PKIX definition but
9204 VeriSign uses it and IE5 only recognises this form. Document 'x509'
9208 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
9209 without disallowing inline assembler and the like for non-pedantic builds.
9212 *) Support Borland C++ builder.
9213 [Janez Jere <jj@void.si>, modified by Ulf Möller]
9218 *) SHA-1 cleanups and performance enhancements.
9219 [Andy Polyakov <appro@fy.chalmers.se>]
9221 *) Sparc v8plus assembler for the bignum library.
9222 [Andy Polyakov <appro@fy.chalmers.se>]
9224 *) Accept any -xxx and +xxx compiler options in Configure.
9227 *) Update HPUX configuration.
9230 *) Add missing sk_<type>_unshift() function to safestack.h
9231 [Ralf S. Engelschall]
9233 *) New function SSL_CTX_use_certificate_chain_file that sets the
9234 "extra_cert"s in addition to the certificate. (This makes sense
9235 only for "PEM" format files, as chains as a whole are not
9239 *) Support verify_depth from the SSL API.
9240 x509_vfy.c had what can be considered an off-by-one-error:
9241 Its depth (which was not part of the external interface)
9242 was actually counting the number of certificates in a chain;
9243 now it really counts the depth.
9246 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
9247 instead of X509err, which often resulted in confusing error
9248 messages since the error codes are not globally unique
9249 (e.g. an alleged error in ssl3_accept when a certificate
9250 didn't match the private key).
9252 *) New function SSL_CTX_set_session_id_context that allows to set a default
9253 value (so that you don't need SSL_set_session_id_context for each
9254 connection using the SSL_CTX).
9257 *) OAEP decoding bug fix.
9260 *) Support INSTALL_PREFIX for package builders, as proposed by
9264 *) New Configure options "threads" and "no-threads". For systems
9265 where the proper compiler options are known (currently Solaris
9266 and Linux), "threads" is the default.
9269 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
9272 *) Install various scripts to $(OPENSSLDIR)/misc, not to
9273 $(INSTALLTOP)/bin -- they shouldn't clutter directories
9274 such as /usr/local/bin.
9277 *) "make linux-shared" to build shared libraries.
9278 [Niels Poppe <niels@netbox.org>]
9280 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
9283 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
9284 extension adding in x509 utility.
9287 *) Remove NOPROTO sections and error code comments.
9290 *) Partial rewrite of the DEF file generator to now parse the ANSI
9294 *) New Configure options --prefix=DIR and --openssldir=DIR.
9297 *) Complete rewrite of the error code script(s). It is all now handled
9298 by one script at the top level which handles error code gathering,
9299 header rewriting and C source file generation. It should be much better
9300 than the old method: it now uses a modified version of Ulf's parser to
9301 read the ANSI prototypes in all header files (thus the old K&R definitions
9302 aren't needed for error creation any more) and do a better job of
9303 translating function codes into names. The old 'ASN1 error code imbedded
9304 in a comment' is no longer necessary and it doesn't use .err files which
9305 have now been deleted. Also the error code call doesn't have to appear all
9306 on one line (which resulted in some large lines...).
9309 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
9312 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
9313 0 (which usually indicates a closed connection), but continue reading.
9316 *) Fix some race conditions.
9319 *) Add support for CRL distribution points extension. Add Certificate
9320 Policies and CRL distribution points documentation.
9323 *) Move the autogenerated header file parts to crypto/opensslconf.h.
9326 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
9327 8 of keying material. Merlin has also confirmed interop with this fix
9328 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
9329 [Merlin Hughes <merlin@baltimore.ie>]
9331 *) Fix lots of warnings.
9332 [Richard Levitte <levitte@stacken.kth.se>]
9334 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
9335 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9336 [Richard Levitte <levitte@stacken.kth.se>]
9338 *) Fix problems with sizeof(long) == 8.
9339 [Andy Polyakov <appro@fy.chalmers.se>]
9341 *) Change functions to ANSI C.
9344 *) Fix typos in error codes.
9345 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9347 *) Remove defunct assembler files from Configure.
9350 *) SPARC v8 assembler BIGNUM implementation.
9351 [Andy Polyakov <appro@fy.chalmers.se>]
9353 *) Support for Certificate Policies extension: both print and set.
9354 Various additions to support the r2i method this uses.
9357 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9358 return a const string when you are expecting an allocated buffer.
9361 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9362 types DirectoryString and DisplayText.
9365 *) Add code to allow r2i extensions to access the configuration database,
9366 add an LHASH database driver and add several ctx helper functions.
9369 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9370 fail when they extended the size of a BIGNUM.
9373 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9374 support typesafe stack.
9377 *) Fix typo in SSL_[gs]et_options().
9378 [Nils Frostberg <nils@medcom.se>]
9380 *) Delete various functions and files that belonged to the (now obsolete)
9381 old X509V3 handling code.
9384 *) New Configure option "rsaref".
9387 *) Don't auto-generate pem.h.
9390 *) Introduce type-safe ASN.1 SETs.
9393 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9394 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9396 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9397 that links with OpenSSL (well at least cause lots of warnings), but fear
9398 not: the conversion is trivial, and it eliminates loads of evil casts. A
9399 few STACKed things have been converted already. Feel free to convert more.
9400 In the fullness of time, I'll do away with the STACK type altogether.
9403 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9404 specified in <certfile> by updating the entry in the index.txt file.
9405 This way one no longer has to edit the index.txt file manually for
9406 revoking a certificate. The -revoke option does the gory details now.
9407 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9409 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9410 `-text' option at all and this way the `-noout -text' combination was
9411 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9412 [Ralf S. Engelschall]
9414 *) Make sure a corresponding plain text error message exists for the
9415 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9416 verify callback function determined that a certificate was revoked.
9417 [Ralf S. Engelschall]
9419 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9420 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9421 all available cipers including rc5, which was forgotten until now.
9422 In order to let the testing shell script know which algorithms
9423 are available, a new (up to now undocumented) command
9424 "openssl list-cipher-commands" is used.
9427 *) Bugfix: s_client occasionally would sleep in select() when
9428 it should have checked SSL_pending() first.
9431 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9432 the raw DSA values prior to ASN.1 encoding.
9435 *) Tweaks to Configure
9436 [Niels Poppe <niels@netbox.org>]
9438 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9442 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9445 *) New config option to avoid instructions that are illegal on the 80386.
9446 The default code is faster, but requires at least a 486.
9449 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9450 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9451 same as SSL2_VERSION anyway.
9454 *) New "-showcerts" option for s_client.
9457 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9458 application. Various cleanups and fixes.
9461 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9462 modify error routines to work internally. Add error codes and PBE init
9463 to library startup routines.
9466 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9467 packing functions to asn1 and evp. Changed function names and error
9468 codes along the way.
9471 *) PKCS12 integration: and so it begins... First of several patches to
9472 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9473 objects to objects.h
9476 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9477 and display support for Thawte strong extranet extension.
9480 *) Add LinuxPPC support.
9481 [Jeff Dubrule <igor@pobox.org>]
9483 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9484 bn_div_words in alpha.s.
9485 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9487 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9488 OAEP isn't supported when OpenSSL is built with RSAref.
9489 [Ulf Moeller <ulf@fitug.de>]
9491 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9492 so they no longer are missing under -DNOPROTO.
9493 [Soren S. Jorvang <soren@t.dk>]
9496 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9498 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9499 doesn't work when the session is reused. Coming soon!
9502 *) Fix a security hole, that allows sessions to be reused in the wrong
9503 context thus bypassing client cert protection! All software that uses
9504 client certs and session caches in multiple contexts NEEDS PATCHING to
9505 allow session reuse! A fuller solution is in the works.
9506 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9508 *) Some more source tree cleanups (removed obsolete files
9509 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9510 permission on "config" script to be executable) and a fix for the INSTALL
9512 [Ulf Moeller <ulf@fitug.de>]
9514 *) Remove some legacy and erroneous uses of malloc, free instead of
9516 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9518 *) Make rsa_oaep_test return non-zero on error.
9519 [Ulf Moeller <ulf@fitug.de>]
9521 *) Add support for native Solaris shared libraries. Configure
9522 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9523 if someone would make that last step automatic.
9524 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9526 *) ctx_size was not built with the right compiler during "make links". Fixed.
9529 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9530 except NULL ciphers". This means the default cipher list will no longer
9531 enable NULL ciphers. They need to be specifically enabled e.g. with
9532 the string "DEFAULT:eNULL".
9535 *) Fix to RSA private encryption routines: if p < q then it would
9536 occasionally produce an invalid result. This will only happen with
9537 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9540 *) Be less restrictive and allow also `perl util/perlpath.pl
9541 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9542 because this way one can also use an interpreter named `perl5' (which is
9543 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9544 installed as `perl').
9545 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9547 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9548 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9550 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9551 advapi32.lib to Win32 build and change the pem test comparision
9552 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9553 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9554 and crypto/des/ede_cbcm_enc.c.
9557 *) DES quad checksum was broken on big-endian architectures. Fixed.
9560 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9561 Win32 test batch file so it (might) work again. The Win32 test batch file
9562 is horrible: I feel ill....
9565 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9566 in e_os.h. Audit of header files to check ANSI and non ANSI
9567 sections: 10 functions were absent from non ANSI section and not exported
9568 from Windows DLLs. Fixed up libeay.num for new functions.
9571 *) Make `openssl version' output lines consistent.
9572 [Ralf S. Engelschall]
9574 *) Fix Win32 symbol export lists for BIO functions: Added
9575 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9576 to ms/libeay{16,32}.def.
9577 [Ralf S. Engelschall]
9579 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9580 fine under Unix and passes some trivial tests I've now added. But the
9581 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9582 added to make sure no one expects that this stuff really works in the
9583 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9584 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9586 [Ralf S. Engelschall]
9588 *) Fix the generation of two part addresses in perl.
9589 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9591 *) Add config entry for Linux on MIPS.
9592 [John Tobey <jtobey@channel1.com>]
9594 *) Make links whenever Configure is run, unless we are on Windoze.
9597 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9598 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9602 *) Add a useful kludge to allow package maintainers to specify compiler and
9603 other platforms details on the command line without having to patch the
9604 Configure script everytime: One now can use ``perl Configure
9605 <id>:<details>'', i.e. platform ids are allowed to have details appended
9606 to them (seperated by colons). This is treated as there would be a static
9607 pre-configured entry in Configure's %table under key <id> with value
9608 <details> and ``perl Configure <id>'' is called. So, when you want to
9609 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9610 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9611 now, which overrides the FreeBSD-elf entry on-the-fly.
9612 [Ralf S. Engelschall]
9614 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9617 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9618 on the `perl Configure ...' command line. This way one can compile
9619 OpenSSL libraries with Position Independent Code (PIC) which is needed
9620 for linking it into DSOs.
9621 [Ralf S. Engelschall]
9623 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9627 *) Cleaned up the LICENSE document: The official contact for any license
9628 questions now is the OpenSSL core team under openssl-core@openssl.org.
9629 And add a paragraph about the dual-license situation to make sure people
9630 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9631 to the OpenSSL toolkit.
9632 [Ralf S. Engelschall]
9634 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9635 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9636 Additonally cleaned up the `make links' target: Remove unnecessary
9637 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9638 to speed processing and no longer clutter the display with confusing
9639 stuff. Instead only the actually done links are displayed.
9640 [Ralf S. Engelschall]
9642 *) Permit null encryption ciphersuites, used for authentication only. It used
9643 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9644 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9648 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9649 signed attributes when verifying signatures (this would break them),
9650 the detached data encoding was wrong and public keys obtained using
9651 X509_get_pubkey() weren't freed.
9654 *) Add text documentation for the BUFFER functions. Also added a work around
9655 to a Win95 console bug. This was triggered by the password read stuff: the
9656 last character typed gets carried over to the next fread(). If you were
9657 generating a new cert request using 'req' for example then the last
9658 character of the passphrase would be CR which would then enter the first
9662 *) Added the new `Includes OpenSSL Cryptography Software' button as
9663 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9664 button and can be used by applications based on OpenSSL to show the
9665 relationship to the OpenSSL project.
9666 [Ralf S. Engelschall]
9668 *) Remove confusing variables in function signatures in files
9669 ssl/ssl_lib.c and ssl/ssl.h.
9670 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9672 *) Don't install bss_file.c under PREFIX/include/
9673 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9675 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9676 functions that return function pointers and has support for NT specific
9677 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9678 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9679 unsigned to signed types: this was killing the Win32 compile.
9682 *) Add new certificate file to stack functions,
9683 SSL_add_dir_cert_subjects_to_stack() and
9684 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9685 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9686 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9687 This means that Apache-SSL and similar packages don't have to mess around
9688 to add as many CAs as they want to the preferred list.
9691 *) Experiment with doxygen documentation. Currently only partially applied to
9693 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9694 openssl.doxy as the configuration file.
9697 *) Get rid of remaining C++-style comments which strict C compilers hate.
9698 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9700 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9701 compiled in by default: it has problems with large keys.
9704 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9705 DH private keys and/or callback functions which directly correspond to
9706 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9707 is needed for applications which have to configure certificates on a
9708 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9710 For the RSA certificate situation is makes no difference, but
9711 for the DSA certificate situation this fixes the "no shared cipher"
9712 problem where the OpenSSL cipher selection procedure failed because the
9713 temporary keys were not overtaken from the context and the API provided
9714 no way to reconfigure them.
9715 The new functions now let applications reconfigure the stuff and they
9716 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9717 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9718 non-public-API function ssl_cert_instantiate() is used as a helper
9719 function and also to reduce code redundancy inside ssl_rsa.c.
9720 [Ralf S. Engelschall]
9722 *) Move s_server -dcert and -dkey options out of the undocumented feature
9723 area because they are useful for the DSA situation and should be
9724 recognized by the users.
9725 [Ralf S. Engelschall]
9727 *) Fix the cipher decision scheme for export ciphers: the export bits are
9728 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9729 SSL_EXP_MASK. So, the original variable has to be used instead of the
9730 already masked variable.
9731 [Richard Levitte <levitte@stacken.kth.se>]
9733 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9734 [Richard Levitte <levitte@stacken.kth.se>]
9736 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9737 from `int' to `unsigned int' because it's a length and initialized by
9738 EVP_DigestFinal() which expects an `unsigned int *'.
9739 [Richard Levitte <levitte@stacken.kth.se>]
9741 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9742 script. Instead use the usual Shell->Perl transition trick.
9743 [Ralf S. Engelschall]
9745 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9746 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9747 -noout -modulus' as it's already the case for `openssl rsa -noout
9748 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9749 currently the public key is printed (a decision which was already done by
9750 `openssl dsa -modulus' in the past) which serves a similar purpose.
9751 Additionally the NO_RSA no longer completely removes the whole -modulus
9752 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9754 [Ralf S. Engelschall]
9756 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9757 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9758 [Arne Ansper <arne@ats.cyber.ee>]
9760 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9761 to be added. Now both 'req' and 'ca' can use new objects defined in the
9765 *) Add cool BIO that does syslog (or event log on NT).
9766 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9768 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9769 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9770 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9771 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9774 *) Add preliminary config info for new extension code.
9777 *) Make RSA_NO_PADDING really use no padding.
9778 [Ulf Moeller <ulf@fitug.de>]
9780 *) Generate errors when private/public key check is done.
9783 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9784 for some CRL extensions and new objects added.
9787 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9788 key usage extension and fuller support for authority key id.
9791 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9792 padding method for RSA, which is recommended for new applications in PKCS
9793 #1 v2.0 (RFC 2437, October 1998).
9794 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9795 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9796 against Bleichbacher's attack on RSA.
9797 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9800 *) Updates to the new SSL compression code
9801 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9803 *) Fix so that the version number in the master secret, when passed
9804 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9805 (because the server will not accept higher), that the version number
9806 is 0x03,0x01, not 0x03,0x00
9807 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9809 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9810 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9811 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9814 *) Support for RAW extensions where an arbitrary extension can be
9815 created by including its DER encoding. See apps/openssl.cnf for
9819 *) Make sure latest Perl versions don't interpret some generated C array
9820 code as Perl array code in the crypto/err/err_genc.pl script.
9821 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9823 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9824 not many people have the assembler. Various Win32 compilation fixes and
9825 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9829 *) Modify configure script 'Configure' to automatically create crypto/date.h
9830 file under Win32 and also build pem.h from pem.org. New script
9831 util/mkfiles.pl to create the MINFO file on environments that can't do a
9832 'make files': perl util/mkfiles.pl >MINFO should work.
9835 *) Major rework of DES function declarations, in the pursuit of correctness
9836 and purity. As a result, many evil casts evaporated, and some weirdness,
9837 too. You may find this causes warnings in your code. Zapping your evil
9838 casts will probably fix them. Mostly.
9841 *) Fix for a typo in asn1.h. Bug fix to object creation script
9842 obj_dat.pl. It considered a zero in an object definition to mean
9843 "end of object": none of the objects in objects.h have any zeros
9844 so it wasn't spotted.
9845 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9847 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9848 Masking (CBCM). In the absence of test vectors, the best I have been able
9849 to do is check that the decrypt undoes the encrypt, so far. Send me test
9850 vectors if you have them.
9853 *) Correct calculation of key length for export ciphers (too much space was
9854 allocated for null ciphers). This has not been tested!
9857 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9858 message is now correct (it understands "crypto" and "ssl" on its
9859 command line). There is also now an "update" option. This will update
9860 the util/ssleay.num and util/libeay.num files with any new functions.
9862 perl util/mkdef.pl crypto ssl update
9863 it will update them.
9866 *) Overhauled the Perl interface (perl/*):
9867 - ported BN stuff to OpenSSL's different BN library
9868 - made the perl/ source tree CVS-aware
9869 - renamed the package from SSLeay to OpenSSL (the files still contain
9870 their history because I've copied them in the repository)
9871 - removed obsolete files (the test scripts will be replaced
9872 by better Test::Harness variants in the future)
9873 [Ralf S. Engelschall]
9875 *) First cut for a very conservative source tree cleanup:
9876 1. merge various obsolete readme texts into doc/ssleay.txt
9877 where we collect the old documents and readme texts.
9878 2. remove the first part of files where I'm already sure that we no
9879 longer need them because of three reasons: either they are just temporary
9880 files which were left by Eric or they are preserved original files where
9881 I've verified that the diff is also available in the CVS via "cvs diff
9882 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9883 the crypto/md/ stuff).
9884 [Ralf S. Engelschall]
9886 *) More extension code. Incomplete support for subject and issuer alt
9887 name, issuer and authority key id. Change the i2v function parameters
9888 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9889 what that's for :-) Fix to ASN1 macro which messed up
9890 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9893 *) Preliminary support for ENUMERATED type. This is largely copied from the
9897 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9898 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9900 *) Make sure `make rehash' target really finds the `openssl' program.
9901 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9903 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9904 like to hear about it if this slows down other processors.
9907 *) Add CygWin32 platform information to Configure script.
9908 [Alan Batie <batie@aahz.jf.intel.com>]
9910 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
9911 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
9913 *) New program nseq to manipulate netscape certificate sequences
9916 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
9920 *) Fixes to BN code. Previously the default was to define BN_RECURSION
9921 but the BN code had some problems that would cause failures when
9922 doing certificate verification and some other functions.
9923 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9925 *) Add ASN1 and PEM code to support netscape certificate sequences.
9928 *) Add ASN1 and PEM code to support netscape certificate sequences.
9931 *) Add several PKIX and private extended key usage OIDs.
9934 *) Modify the 'ca' program to handle the new extension code. Modify
9935 openssl.cnf for new extension format, add comments.
9938 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
9939 and add a sample to openssl.cnf so req -x509 now adds appropriate
9943 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
9944 error code, add initial support to X509_print() and x509 application.
9947 *) Takes a deep breath and start addding X509 V3 extension support code. Add
9948 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
9949 stuff is currently isolated and isn't even compiled yet.
9952 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
9953 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
9954 Removed the versions check from X509 routines when loading extensions:
9955 this allows certain broken certificates that don't set the version
9956 properly to be processed.
9959 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
9960 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
9961 can still be regenerated with "make depend".
9964 *) Spelling mistake in C version of CAST-128.
9965 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
9967 *) Changes to the error generation code. The perl script err-code.pl
9968 now reads in the old error codes and retains the old numbers, only
9969 adding new ones if necessary. It also only changes the .err files if new
9970 codes are added. The makefiles have been modified to only insert errors
9971 when needed (to avoid needlessly modifying header files). This is done
9972 by only inserting errors if the .err file is newer than the auto generated
9973 C file. To rebuild all the error codes from scratch (the old behaviour)
9974 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
9975 or delete all the .err files.
9978 *) CAST-128 was incorrectly implemented for short keys. The C version has
9979 been fixed, but is untested. The assembler versions are also fixed, but
9980 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
9981 to regenerate it if needed.
9982 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
9983 Hagino <itojun@kame.net>]
9985 *) File was opened incorrectly in randfile.c.
9986 [Ulf Möller <ulf@fitug.de>]
9988 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
9989 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
9990 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
9991 al: it's just almost always a UTCTime. Note this patch adds new error
9992 codes so do a "make errors" if there are problems.
9995 *) Correct Linux 1 recognition in config.
9996 [Ulf Möller <ulf@fitug.de>]
9998 *) Remove pointless MD5 hash when using DSA keys in ca.
9999 [Anonymous <nobody@replay.com>]
10001 *) Generate an error if given an empty string as a cert directory. Also
10002 generate an error if handed NULL (previously returned 0 to indicate an
10003 error, but didn't set one).
10004 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
10006 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
10009 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
10010 parameters. This was causing a warning which killed off the Win32 compile.
10013 *) Remove C++ style comments from crypto/bn/bn_local.h.
10014 [Neil Costigan <neil.costigan@celocom.com>]
10016 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
10017 based on a text string, looking up short and long names and finally
10018 "dot" format. The "dot" format stuff didn't work. Added new function
10019 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
10020 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
10021 OID is not part of the table.
10024 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
10025 X509_LOOKUP_by_alias().
10028 *) Sort openssl functions by name.
10031 *) Get the gendsa program working (hopefully) and add it to app list. Remove
10032 encryption from sample DSA keys (in case anyone is interested the password
10036 *) Make _all_ *_free functions accept a NULL pointer.
10037 [Frans Heymans <fheymans@isaserver.be>]
10039 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10041 [Anonymous <nobody@replay.com>]
10043 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10044 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10046 *) Don't blow it for numeric -newkey arguments to apps/req.
10047 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10049 *) Temp key "for export" tests were wrong in s3_srvr.c.
10050 [Anonymous <nobody@replay.com>]
10052 *) Add prototype for temp key callback functions
10053 SSL_CTX_set_tmp_{rsa,dh}_callback().
10056 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10057 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10060 *) X509_name_add_entry() freed the wrong thing after an error.
10061 [Arne Ansper <arne@ats.cyber.ee>]
10063 *) rsa_eay.c would attempt to free a NULL context.
10064 [Arne Ansper <arne@ats.cyber.ee>]
10066 *) BIO_s_socket() had a broken should_retry() on Windoze.
10067 [Arne Ansper <arne@ats.cyber.ee>]
10069 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10070 [Arne Ansper <arne@ats.cyber.ee>]
10072 *) Make sure the already existing X509_STORE->depth variable is initialized
10073 in X509_STORE_new(), but document the fact that this variable is still
10074 unused in the certificate verification process.
10075 [Ralf S. Engelschall]
10077 *) Fix the various library and apps files to free up pkeys obtained from
10078 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10081 *) Fix reference counting in X509_PUBKEY_get(). This makes
10082 demos/maurice/example2.c work, amongst others, probably.
10083 [Steve Henson and Ben Laurie]
10085 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10086 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10087 are no longer created. This way we have a single and consistent command
10088 line interface `openssl <command>', similar to `cvs <command>'.
10089 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10091 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10092 BIT STRING wrapper always have zero unused bits.
10095 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10098 *) Make the top-level INSTALL documentation easier to understand.
10101 *) Makefiles updated to exit if an error occurs in a sub-directory
10102 make (including if user presses ^C) [Paul Sutton]
10104 *) Make Montgomery context stuff explicit in RSA data structure.
10107 *) Fix build order of pem and err to allow for generated pem.h.
10110 *) Fix renumbering bug in X509_NAME_delete_entry().
10113 *) Enhanced the err-ins.pl script so it makes the error library number
10114 global and can add a library name. This is needed for external ASN1 and
10115 other error libraries.
10118 *) Fixed sk_insert which never worked properly.
10121 *) Fix ASN1 macros so they can handle indefinite length construted
10122 EXPLICIT tags. Some non standard certificates use these: they can now
10126 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10127 into a single doc/ssleay.txt bundle. This way the information is still
10128 preserved but no longer messes up this directory. Now it's new room for
10129 the new set of documenation files.
10130 [Ralf S. Engelschall]
10132 *) SETs were incorrectly DER encoded. This was a major pain, because they
10133 shared code with SEQUENCEs, which aren't coded the same. This means that
10134 almost everything to do with SETs or SEQUENCEs has either changed name or
10135 number of arguments.
10136 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10138 *) Fix test data to work with the above.
10141 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10142 was already fixed by Eric for 0.9.1 it seems.
10143 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10145 *) Autodetect FreeBSD3.
10148 *) Fix various bugs in Configure. This affects the following platforms:
10152 unixware-2.0-pentium
10156 *) Eliminate generated files from CVS. Reorder tests to regenerate files
10157 before they are needed.
10160 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
10164 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
10166 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
10167 changed SSLeay to OpenSSL in version strings.
10168 [Ralf S. Engelschall]
10170 *) Some fixups to the top-level documents.
10173 *) Fixed the nasty bug where rsaref.h was not found under compile-time
10174 because the symlink to include/ was missing.
10175 [Ralf S. Engelschall]
10177 *) Incorporated the popular no-RSA/DSA-only patches
10178 which allow to compile a RSA-free SSLeay.
10179 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
10181 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
10182 when "ssleay" is still not found.
10183 [Ralf S. Engelschall]
10185 *) Added more platforms to Configure: Cray T3E, HPUX 11,
10186 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
10188 *) Updated the README file.
10189 [Ralf S. Engelschall]
10191 *) Added various .cvsignore files in the CVS repository subdirs
10192 to make a "cvs update" really silent.
10193 [Ralf S. Engelschall]
10195 *) Recompiled the error-definition header files and added
10196 missing symbols to the Win32 linker tables.
10197 [Ralf S. Engelschall]
10199 *) Cleaned up the top-level documents;
10200 o new files: CHANGES and LICENSE
10201 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
10202 o merged COPYRIGHT into LICENSE
10203 o removed obsolete TODO file
10204 o renamed MICROSOFT to INSTALL.W32
10205 [Ralf S. Engelschall]
10207 *) Removed dummy files from the 0.9.1b source tree:
10208 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
10209 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
10210 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
10211 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
10212 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
10213 [Ralf S. Engelschall]
10215 *) Added various platform portability fixes.
10218 *) The Genesis of the OpenSSL rpject:
10219 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
10220 Young and Tim J. Hudson created while they were working for C2Net until
10222 [The OpenSSL Project]
10225 Changes between 0.9.0b and 0.9.1b [not released]
10227 *) Updated a few CA certificates under certs/
10230 *) Changed some BIGNUM api stuff.
10233 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
10234 DGUX x86, Linux Alpha, etc.
10237 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
10238 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
10242 *) Add -strparse option to asn1pars program which parses nested
10244 [Dr Stephen Henson <shenson@bigfoot.com>]
10246 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
10249 *) DSA fix for "ca" program.
10252 *) Added "-genkey" option to "dsaparam" program.
10255 *) Added RIPE MD160 (rmd160) message digest.
10258 *) Added -a (all) option to "ssleay version" command.
10261 *) Added PLATFORM define which is the id given to Configure.
10264 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
10267 *) Extended the ASN.1 parser routines.
10270 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
10273 *) Added a BN_CTX to the BN library.
10276 *) Fixed the weak key values in DES library
10279 *) Changed API in EVP library for cipher aliases.
10282 *) Added support for RC2/64bit cipher.
10285 *) Converted the lhash library to the crypto/mem.c functions.
10288 *) Added more recognized ASN.1 object ids.
10291 *) Added more RSA padding checks for SSL/TLS.
10294 *) Added BIO proxy/filter functionality.
10297 *) Added extra_certs to SSL_CTX which can be used
10298 send extra CA certificates to the client in the CA cert chain sending
10299 process. It can be configured with SSL_CTX_add_extra_chain_cert().
10302 *) Now Fortezza is denied in the authentication phase because
10303 this is key exchange mechanism is not supported by SSLeay at all.
10306 *) Additional PKCS1 checks.
10309 *) Support the string "TLSv1" for all TLS v1 ciphers.
10312 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
10313 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
10316 *) Fixed a few memory leaks.
10319 *) Fixed various code and comment typos.
10322 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
10323 bytes sent in the client random.
10324 [Edward Bishop <ebishop@spyglass.com>]
projects / openssl.git / blob