5 Changes between 1.0.1 and 1.1.0 [xx XXX xxxx]
7 *) Print out deprecated issuer and subject unique ID fields in
12 [Emilia Kasper, Adam Langley, Ben Laurie (Google)]
14 *) Support for automatic EC temporary key parameter selection. If enabled
15 the most preferred EC parameters are automatically used instead of
16 hardcoded fixed parameters. Now a server just has to call:
17 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
18 support ECDH and use the most appropriate parameters.
21 *) Enhance and tidy EC curve and point format TLS extension code. Use
22 static structures instead of allocation if default values are used.
23 New ctrls to set curves we wish to support and to retrieve shared curves.
24 Print out shared curves in s_server. New options to s_server and s_client
25 to set list of supported curves.
28 *) New ctrls to retrieve supported signature algorithms and
29 supported curve values as an array of NIDs. Extend openssl utility
30 to print out received values.
33 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
34 between NIDs and the more common NIST names such as "P-256". Enhance
35 ecparam utility and ECC method to recognise the NIST names for curves.
38 *) Enhance SSL/TLS certificate chain handling to support different
39 chains for each certificate instead of one chain in the parent SSL_CTX.
42 *) Support for fixed DH ciphersuite client authentication: where both
43 server and client use DH certificates with common parameters.
46 *) Support for fixed DH ciphersuites: those requiring DH server
50 *) Transparently support X9.42 DH parameters when calling
51 PEM_read_bio_DHparameters. This means existing applications can handle
52 the new parameter format automatically.
55 *) Initial experimental support for X9.42 DH parameter format: mainly
56 to support use of 'q' parameter for RFC5114 parameters.
59 *) Add DH parameters from RFC5114 including test data to dhtest.
62 *) Update fips_test_suite to support multiple command line options. New
63 test to induce all self test errors in sequence and check expected
67 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
68 sign or verify all in one operation.
71 *) Add fips_algvs: a multicall fips utility incorporaing all the algorithm
72 test programs and fips_test_suite. Includes functionality to parse
73 the minimal script output of fipsalgest.pl directly.
76 *) Add authorisation parameter to FIPS_module_mode_set().
79 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
82 *) Use separate DRBG fields for internal and external flags. New function
83 FIPS_drbg_health_check() to perform on demand health checking. Add
84 generation tests to fips_test_suite with reduced health check interval to
85 demonstrate periodic health checking. Add "nodh" option to
86 fips_test_suite to skip very slow DH test.
89 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
93 *) More extensive health check for DRBG checking many more failure modes.
94 New function FIPS_selftest_drbg_all() to handle every possible DRBG
95 combination: call this in fips_test_suite.
98 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
99 and POST to handle Dual EC cases.
102 *) Add support for canonical generation of DSA parameter 'g'. See
105 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
106 POST to handle HMAC cases.
109 *) Add functions FIPS_module_version() and FIPS_module_version_text()
110 to return numberical and string versions of the FIPS module number.
113 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
114 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implmeneted
115 outside the validated module in the FIPS capable OpenSSL.
118 *) Minor change to DRBG entropy callback semantics. In some cases
119 there is no mutiple of the block length between min_len and
120 max_len. Allow the callback to return more than max_len bytes
121 of entropy but discard any extra: it is the callback's responsibility
122 to ensure that the extra data discarded does not impact the
123 requested amount of entropy.
126 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
127 information in FIPS186-3, SP800-57 and SP800-131A.
130 *) CCM support via EVP. Interface is very similar to GCM case except we
131 must supply all data in one chunk (i.e. no update, final) and the
132 message length must be supplied if AAD is used. Add algorithm test
136 *) Initial version of POST overhaul. Add POST callback to allow the status
137 of POST to be monitored and/or failures induced. Modify fips_test_suite
138 to use callback. Always run all selftests even if one fails.
141 *) XTS support including algorithm test driver in the fips_gcmtest program.
142 Note: this does increase the maximum key length from 32 to 64 bytes but
143 there should be no binary compatibility issues as existing applications
144 will never use XTS mode.
147 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
148 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
149 performs algorithm blocking for unapproved PRNG types. Also do not
150 set PRNG type in FIPS_mode_set(): leave this to the application.
151 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
152 the standard OpenSSL PRNG: set additional data to a date time vector.
155 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
156 This shouldn't present any incompatibility problems because applications
157 shouldn't be using these directly and any that are will need to rethink
158 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
161 *) Extensive self tests and health checking required by SP800-90 DRBG.
162 Remove strength parameter from FIPS_drbg_instantiate and always
163 instantiate at maximum supported strength.
166 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
169 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
172 *) New function DH_compute_key_padded() to compute a DH key and pad with
173 leading zeroes if needed: this complies with SP800-56A et al.
176 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
177 anything, incomplete, subject to change and largely untested at present.
180 *) Modify fipscanisteronly build option to only build the necessary object
181 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
184 *) Add experimental option FIPSSYMS to give all symbols in
185 fipscanister.o and FIPS or fips prefix. This will avoid
186 conflicts with future versions of OpenSSL. Add perl script
187 util/fipsas.pl to preprocess assembly language source files
188 and rename any affected symbols.
191 *) Add selftest checks and algorithm block of non-fips algorithms in
192 FIPS mode. Remove DES2 from selftests.
195 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
196 return internal method without any ENGINE dependencies. Add new
197 tiny fips sign and verify functions.
200 *) New build option no-ec2m to disable characteristic 2 code.
203 *) New build option "fipscanisteronly". This only builds fipscanister.o
204 and (currently) associated fips utilities. Uses the file Makefile.fips
205 instead of Makefile.org as the prototype.
208 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
209 Update fips_gcmtest to use IV generator.
212 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
213 setting output buffer to NULL. The *Final function must be
214 called although it will not retrieve any additional data. The tag
215 can be set or retrieved with a ctrl. The IV length is by default 12
216 bytes (96 bits) but can be set to an alternative value. If the IV
217 length exceeds the maximum IV length (currently 16 bytes) it cannot be
221 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
222 underlying do_cipher function handles all cipher semantics itself
223 including padding and finalisation. This is useful if (for example)
224 an ENGINE cipher handles block padding itself. The behaviour of
225 do_cipher is subtly changed if this flag is set: the return value
226 is the number of characters written to the output buffer (zero is
227 no longer an error code) or a negative error code. Also if the
228 input buffer is NULL and length 0 finalisation should be performed.
231 *) If a candidate issuer certificate is already part of the constructed
232 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
235 *) Improve forward-security support: add functions
237 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
238 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
240 for use by SSL/TLS servers; the callback function will be called whenever a
241 new session is created, and gets to decide whether the session may be
242 cached to make it resumable (return 0) or not (return 1). (As by the
243 SSL/TLS protocol specifications, the session_id sent by the server will be
244 empty to indicate that the session is not resumable; also, the server will
245 not generate RFC 4507 (RFC 5077) session tickets.)
247 A simple reasonable callback implementation is to return is_forward_secure.
248 This parameter will be set to 1 or 0 depending on the ciphersuite selected
249 by the SSL/TLS server library, indicating whether it can provide forward
251 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
253 *) New function OPENSSL_gmtime_diff to find the difference in days
254 and seconds between two tm structures. This will be used to provide
255 additional functionality for ASN1_TIME.
258 *) Add -trusted_first option which attempts to find certificates in the
259 trusted store even if an untrusted chain is also supplied.
262 *) Initial experimental support for explicitly trusted non-root CAs.
263 OpenSSL still tries to build a complete chain to a root but if an
264 intermediate CA has a trust setting included that is used. The first
265 setting is used: whether to trust or reject.
268 *) New -verify_name option in command line utilities to set verification
272 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
273 Add CMAC pkey methods.
276 *) Experimental regnegotiation in s_server -www mode. If the client
277 browses /reneg connection is renegotiated. If /renegcert it is
278 renegotiated requesting a certificate.
281 *) Add an "external" session cache for debugging purposes to s_server. This
282 should help trace issues which normally are only apparent in deployed
283 multi-process servers.
286 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
287 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
288 BIO_set_cipher() and some obscure PEM functions were changed so they
289 can now return an error. The RAND changes required a change to the
290 RAND_METHOD structure.
293 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
294 a gcc attribute to warn if the result of a function is ignored. This
295 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
296 whose return value is often ignored.
299 Changes between 1.0.1c and 1.0.1d [xx XXX xxxx]
301 *) Fix possible deadlock when decoding public keys.
304 *) Don't use TLS 1.0 record version number in initial client hello
308 Changes between 1.0.1b and 1.0.1c [10 May 2012]
310 *) Sanity check record length before skipping explicit IV in TLS
311 1.2, 1.1 and DTLS to avoid DoS attack.
313 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
314 fuzzing as a service testing platform.
318 *) Initialise tkeylen properly when encrypting CMS messages.
319 Thanks to Solar Designer of Openwall for reporting this issue.
322 *) In FIPS mode don't try to use composite ciphers as they are not
326 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
328 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
329 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
330 mean any application compiled against OpenSSL 1.0.0 headers setting
331 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
332 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
333 0x10000000L Any application which was previously compiled against
334 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
335 will need to be recompiled as a result. Letting be results in
336 inability to disable specifically TLS 1.1 and in client context,
337 in unlike event, limit maximum offered version to TLS 1.0 [see below].
340 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
341 disable just protocol X, but all protocols above X *if* there are
342 protocols *below* X still enabled. In more practical terms it means
343 that if application wants to disable TLS1.0 in favor of TLS1.1 and
344 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
345 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
349 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
351 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
352 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
353 in CRYPTO_realloc_clean.
355 Thanks to Tavis Ormandy, Google Security Team, for discovering this
356 issue and to Adam Langley <agl@chromium.org> for fixing it.
358 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
360 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
363 *) Workarounds for some broken servers that "hang" if a client hello
364 record length exceeds 255 bytes:
366 1. Do not use record version number > TLS 1.0 in initial client
367 hello: some (but not all) hanging servers will now work.
368 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
369 the number of ciphers sent in the client hello. This should be
370 set to an even number, such as 50, for example by passing:
371 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
372 Most broken servers should now work.
373 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
374 TLS 1.2 client support entirely.
377 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
380 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
382 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
383 STRING form instead of a DigestInfo.
386 *) The format used for MDC2 RSA signatures is inconsistent between EVP
387 and the RSA_sign/RSA_verify functions. This was made more apparent when
388 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
389 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
390 the correct format in RSA_verify so both forms transparently work.
393 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
394 STRING form instead of a DigestInfo.
397 *) Some servers which support TLS 1.0 can choke if we initially indicate
398 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
399 encrypted premaster secret. As a workaround use the maximum pemitted
400 client version in client hello, this should keep such servers happy
401 and still work with previous versions of OpenSSL.
404 *) Add support for TLS/DTLS heartbeats.
405 [Robin Seggelmann <seggelmann@fh-muenster.de>]
407 *) Add support for SCTP.
408 [Robin Seggelmann <seggelmann@fh-muenster.de>]
410 *) Improved PRNG seeding for VOS.
411 [Paul Green <Paul.Green@stratus.com>]
413 *) Extensive assembler packs updates, most notably:
415 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
416 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
417 - x86_64: bit-sliced AES implementation;
418 - ARM: NEON support, contemporary platforms optimizations;
419 - s390x: z196 support;
420 - *: GHASH and GF(2^m) multiplication implementations;
424 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
425 (removal of unnecessary code)
426 [Peter Sylvester <peter.sylvester@edelweb.fr>]
428 *) Add TLS key material exporter from RFC 5705.
431 *) Add DTLS-SRTP negotiation from RFC 5764.
434 *) Add Next Protocol Negotiation,
435 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
436 disabled with a no-npn flag to config or Configure. Code donated
438 [Adam Langley <agl@google.com> and Ben Laurie]
440 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
441 NIST-P256, NIST-P521, with constant-time single point multiplication on
442 typical inputs. Compiler support for the nonstandard type __uint128_t is
443 required to use this (present in gcc 4.4 and later, for 64-bit builds).
444 Code made available under Apache License version 2.0.
446 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
447 line to include this in your build of OpenSSL, and run "make depend" (or
448 "make update"). This enables the following EC_METHODs:
450 EC_GFp_nistp224_method()
451 EC_GFp_nistp256_method()
452 EC_GFp_nistp521_method()
454 EC_GROUP_new_by_curve_name() will automatically use these (while
455 EC_GROUP_new_curve_GFp() currently prefers the more flexible
457 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
459 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
460 all platforms. Move ssize_t definition from e_os.h to the public
461 header file e_os2.h as it now appears in public header file cms.h
464 *) New -sigopt option to the ca, req and x509 utilities. Additional
465 signature parameters can be passed using this option and in
469 *) Add RSA PSS signing function. This will generate and set the
470 appropriate AlgorithmIdentifiers for PSS based on those in the
471 corresponding EVP_MD_CTX structure. No application support yet.
474 *) Support for companion algorithm specific ASN1 signing routines.
475 New function ASN1_item_sign_ctx() signs a pre-initialised
476 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
477 the appropriate parameters.
480 *) Add new algorithm specific ASN1 verification initialisation function
481 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
482 handling will be the same no matter what EVP_PKEY_METHOD is used.
483 Add a PSS handler to support verification of PSS signatures: checked
484 against a number of sample certificates.
487 *) Add signature printing for PSS. Add PSS OIDs.
488 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
490 *) Add algorithm specific signature printing. An individual ASN1 method
491 can now print out signatures instead of the standard hex dump.
493 More complex signatures (e.g. PSS) can print out more meaningful
494 information. Include DSA version that prints out the signature
498 *) Password based recipient info support for CMS library: implementing
502 *) Split password based encryption into PBES2 and PBKDF2 functions. This
503 neatly separates the code into cipher and PBE sections and is required
504 for some algorithms that split PBES2 into separate pieces (such as
508 *) Session-handling fixes:
509 - Fix handling of connections that are resuming with a session ID,
510 but also support Session Tickets.
511 - Fix a bug that suppressed issuing of a new ticket if the client
512 presented a ticket with an expired session.
513 - Try to set the ticket lifetime hint to something reasonable.
514 - Make tickets shorter by excluding irrelevant information.
515 - On the client side, don't ignore renewed tickets.
516 [Adam Langley, Bodo Moeller (Google)]
518 *) Fix PSK session representation.
521 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
523 This work was sponsored by Intel.
526 *) Add GCM support to TLS library. Some custom code is needed to split
527 the IV between the fixed (from PRF) and explicit (from TLS record)
528 portions. This adds all GCM ciphersuites supported by RFC5288 and
529 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
530 add a special AESGCM string for GCM only.
533 *) Expand range of ctrls for AES GCM. Permit setting invocation
534 field on decrypt and retrieval of invocation field only on encrypt.
537 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
538 As required by RFC5289 these ciphersuites cannot be used if for
539 versions of TLS earlier than 1.2.
542 *) For FIPS capable OpenSSL interpret a NULL default public key method
543 as unset and return the appopriate default but do *not* set the default.
544 This means we can return the appopriate method in applications that
545 swicth between FIPS and non-FIPS modes.
548 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
549 ENGINE is used then we cannot handle that in the FIPS module so we
550 keep original code iff non-FIPS operations are allowed.
553 *) Add -attime option to openssl utilities.
554 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
556 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
559 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
560 FIPS EC methods unconditionally for now.
563 *) New build option no-ec2m to disable characteristic 2 code.
566 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
567 all cases can be covered as some introduce binary incompatibilities.
570 *) Redirect RSA operations to FIPS module including keygen,
571 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
574 *) Add similar low level API blocking to ciphers.
577 *) Low level digest APIs are not approved in FIPS mode: any attempt
578 to use these will cause a fatal error. Applications that *really* want
579 to use them can use the private_* version instead.
582 *) Redirect cipher operations to FIPS module for FIPS builds.
585 *) Redirect digest operations to FIPS module for FIPS builds.
588 *) Update build system to add "fips" flag which will link in fipscanister.o
589 for static and shared library builds embedding a signature if needed.
592 *) Output TLS supported curves in preference order instead of numerical
593 order. This is currently hardcoded for the highest order curves first.
594 This should be configurable so applications can judge speed vs strength.
597 *) Add TLS v1.2 server support for client authentication.
600 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
604 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
605 FIPS modules versions.
608 *) Add TLS v1.2 client side support for client authentication. Keep cache
609 of handshake records longer as we don't know the hash algorithm to use
610 until after the certificate request message is received.
613 *) Initial TLS v1.2 client support. Add a default signature algorithms
614 extension including all the algorithms we support. Parse new signature
615 format in client key exchange. Relax some ECC signing restrictions for
616 TLS v1.2 as indicated in RFC5246.
619 *) Add server support for TLS v1.2 signature algorithms extension. Switch
620 to new signature format when needed using client digest preference.
621 All server ciphersuites should now work correctly in TLS v1.2. No client
622 support yet and no support for client certificates.
625 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
626 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
627 ciphersuites. At present only RSA key exchange ciphersuites work with
628 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
629 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
630 and version checking.
633 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
634 with this defined it will not be affected by any changes to ssl internal
635 structures. Add several utility functions to allow openssl application
636 to work with OPENSSL_NO_SSL_INTERN defined.
640 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
642 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
645 *) Permit abbreviated handshakes when renegotiating using the function
646 SSL_renegotiate_abbreviated().
647 [Robin Seggelmann <seggelmann@fh-muenster.de>]
649 *) Add call to ENGINE_register_all_complete() to
650 ENGINE_load_builtin_engines(), so some implementations get used
651 automatically instead of needing explicit application support.
654 *) Add support for TLS key exporter as described in RFC5705.
655 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
657 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
658 a few changes are required:
660 Add SSL_OP_NO_TLSv1_1 flag.
662 Update version checking logic to handle version 1.1.
663 Add explicit IV handling (ported from DTLS code).
664 Add command line options to s_client/s_server.
667 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
669 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
670 in CMS and PKCS7 code. When RSA decryption fails use a random key for
671 content decryption and always return the same error. Note: this attack
672 needs on average 2^20 messages so it only affects automated senders. The
673 old behaviour can be reenabled in the CMS code by setting the
674 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
675 an MMA defence is not necessary.
676 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
677 this issue. (CVE-2012-0884)
680 *) Fix CVE-2011-4619: make sure we really are receiving a
681 client hello before rejecting multiple SGC restarts. Thanks to
682 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
685 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
687 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
688 Thanks to Antonio Martin, Enterprise Secure Access Research and
689 Development, Cisco Systems, Inc. for discovering this bug and
690 preparing a fix. (CVE-2012-0050)
693 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
695 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
696 of the Vaudenay padding oracle attack on CBC mode encryption
697 which enables an efficient plaintext recovery attack against
698 the OpenSSL implementation of DTLS. Their attack exploits timing
699 differences arising during decryption processing. A research
700 paper describing this attack can be found at:
701 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
702 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
703 Security Group at Royal Holloway, University of London
704 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
705 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
706 for preparing the fix. (CVE-2011-4108)
707 [Robin Seggelmann, Michael Tuexen]
709 *) Clear bytes used for block padding of SSL 3.0 records.
711 [Adam Langley (Google)]
713 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
714 Kadianakis <desnacked@gmail.com> for discovering this issue and
715 Adam Langley for preparing the fix. (CVE-2011-4619)
716 [Adam Langley (Google)]
718 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
719 [Andrey Kulikov <amdeich@gmail.com>]
721 *) Prevent malformed RFC3779 data triggering an assertion failure.
722 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
723 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
724 [Rob Austein <sra@hactrn.net>]
726 *) Improved PRNG seeding for VOS.
727 [Paul Green <Paul.Green@stratus.com>]
729 *) Fix ssl_ciph.c set-up race.
730 [Adam Langley (Google)]
732 *) Fix spurious failures in ecdsatest.c.
733 [Emilia Käsper (Google)]
735 *) Fix the BIO_f_buffer() implementation (which was mixing different
736 interpretations of the '..._len' fields).
737 [Adam Langley (Google)]
739 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
740 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
741 threads won't reuse the same blinding coefficients.
743 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
744 lock to call BN_BLINDING_invert_ex, and avoids one use of
745 BN_BLINDING_update for each BN_BLINDING structure (previously,
746 the last update always remained unused).
747 [Emilia Käsper (Google)]
749 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
750 [Bob Buckholz (Google)]
752 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
754 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
755 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
756 [Kaspar Brand <ossl@velox.ch>]
758 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
759 for multi-threaded use of ECDH. (CVE-2011-3210)
760 [Adam Langley (Google)]
762 *) Fix x509_name_ex_d2i memory leak on bad inputs.
765 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
766 signature public key algorithm by using OID xref utilities instead.
767 Before this you could only use some ECC ciphersuites with SHA1 only.
770 *) Add protection against ECDSA timing attacks as mentioned in the paper
771 by Billy Bob Brumley and Nicola Tuveri, see:
773 http://eprint.iacr.org/2011/232.pdf
775 [Billy Bob Brumley and Nicola Tuveri]
777 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
779 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
780 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
782 *) Fix bug in string printing code: if *any* escaping is enabled we must
783 escape the escape character (backslash) or the resulting string is
787 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
789 *) Disable code workaround for ancient and obsolete Netscape browsers
790 and servers: an attacker can use it in a ciphersuite downgrade attack.
791 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
794 *) Fixed J-PAKE implementation error, originally discovered by
795 Sebastien Martini, further info and confirmation from Stefan
796 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
799 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
801 *) Fix extension code to avoid race conditions which can result in a buffer
802 overrun vulnerability: resumed sessions must not be modified as they can
803 be shared by multiple threads. CVE-2010-3864
806 *) Fix WIN32 build system to correctly link an ENGINE directory into
810 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
812 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
814 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
816 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
818 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
819 context. The operation can be customised via the ctrl mechanism in
820 case ENGINEs want to include additional functionality.
823 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
826 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
827 output hashes compatible with older versions of OpenSSL.
828 [Willy Weisz <weisz@vcpc.univie.ac.at>]
830 *) Fix compression algorithm handling: if resuming a session use the
831 compression algorithm of the resumed session instead of determining
832 it from client hello again. Don't allow server to change algorithm.
835 *) Add load_crls() function to apps tidying load_certs() too. Add option
836 to verify utility to allow additional CRLs to be included.
839 *) Update OCSP request code to permit adding custom headers to the request:
840 some responders need this.
843 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
845 [Julia Lawall <julia@diku.dk>]
847 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
848 needlessly dereferenced structures, used obsolete functions and
849 didn't handle all updated verify codes correctly.
852 *) Disable MD2 in the default configuration.
855 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
856 indicate the initial BIO being pushed or popped. This makes it possible
857 to determine whether the BIO is the one explicitly called or as a result
858 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
859 it handles reference counts correctly and doesn't zero out the I/O bio
860 when it is not being explicitly popped. WARNING: applications which
861 included workarounds for the old buggy behaviour will need to be modified
862 or they could free up already freed BIOs.
865 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
866 renaming to all platforms (within the 0.9.8 branch, this was
867 done conditionally on Netware platforms to avoid a name clash).
868 [Guenter <lists@gknw.net>]
870 *) Add ECDHE and PSK support to DTLS.
871 [Michael Tuexen <tuexen@fh-muenster.de>]
873 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
877 *) Add "missing" function EVP_MD_flags() (without this the only way to
878 retrieve a digest flags is by accessing the structure directly. Update
879 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
880 or cipher is registered as in the "from" argument. Print out all
881 registered digests in the dgst usage message instead of manually
882 attempting to work them out.
885 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
886 this allows the use of compression and extensions. Change default cipher
887 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
888 by default unless an application cipher string requests it.
891 *) Alter match criteria in PKCS12_parse(). It used to try to use local
892 key ids to find matching certificates and keys but some PKCS#12 files
893 don't follow the (somewhat unwritten) rules and this strategy fails.
894 Now just gather all certificates together and the first private key
895 then look for the first certificate that matches the key.
898 *) Support use of registered digest and cipher names for dgst and cipher
899 commands instead of having to add each one as a special case. So now
906 openssl dgst -sha256 foo
908 and this works for ENGINE based algorithms too.
912 *) Update Gost ENGINE to support parameter files.
913 [Victor B. Wagner <vitus@cryptocom.ru>]
915 *) Support GeneralizedTime in ca utility.
916 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
918 *) Enhance the hash format used for certificate directory links. The new
919 form uses the canonical encoding (meaning equivalent names will work
920 even if they aren't identical) and uses SHA1 instead of MD5. This form
921 is incompatible with the older format and as a result c_rehash should
922 be used to rebuild symbolic links.
925 *) Make PKCS#8 the default write format for private keys, replacing the
926 traditional format. This form is standardised, more secure and doesn't
927 include an implicit MD5 dependency.
930 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
931 committed to OpenSSL should pass this lot as a minimum.
934 *) Add session ticket override functionality for use by EAP-FAST.
935 [Jouni Malinen <j@w1.fi>]
937 *) Modify HMAC functions to return a value. Since these can be implemented
938 in an ENGINE errors can occur.
941 *) Type-checked OBJ_bsearch_ex.
944 *) Type-checked OBJ_bsearch. Also some constification necessitated
945 by type-checking. Still to come: TXT_DB, bsearch(?),
946 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
950 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
951 seconds to a tm structure directly, instead of going through OS
952 specific date routines. This avoids any issues with OS routines such
953 as the year 2038 bug. New *_adj() functions for ASN1 time structures
954 and X509_time_adj_ex() to cover the extended range. The existing
955 X509_time_adj() is still usable and will no longer have any date issues.
958 *) Delta CRL support. New use deltas option which will attempt to locate
959 and search any appropriate delta CRLs available.
961 This work was sponsored by Google.
964 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
965 code and add additional score elements. Validate alternate CRL paths
966 as part of the CRL checking and indicate a new error "CRL path validation
967 error" in this case. Applications wanting additional details can use
968 the verify callback and check the new "parent" field. If this is not
969 NULL CRL path validation is taking place. Existing applications wont
970 see this because it requires extended CRL support which is off by
973 This work was sponsored by Google.
976 *) Support for freshest CRL extension.
978 This work was sponsored by Google.
981 *) Initial indirect CRL support. Currently only supported in the CRLs
982 passed directly and not via lookup. Process certificate issuer
983 CRL entry extension and lookup CRL entries by bother issuer name
984 and serial number. Check and process CRL issuer entry in IDP extension.
986 This work was sponsored by Google.
989 *) Add support for distinct certificate and CRL paths. The CRL issuer
990 certificate is validated separately in this case. Only enabled if
991 an extended CRL support flag is set: this flag will enable additional
992 CRL functionality in future.
994 This work was sponsored by Google.
997 *) Add support for policy mappings extension.
999 This work was sponsored by Google.
1002 *) Fixes to pathlength constraint, self issued certificate handling,
1003 policy processing to align with RFC3280 and PKITS tests.
1005 This work was sponsored by Google.
1008 *) Support for name constraints certificate extension. DN, email, DNS
1009 and URI types are currently supported.
1011 This work was sponsored by Google.
1014 *) To cater for systems that provide a pointer-based thread ID rather
1015 than numeric, deprecate the current numeric thread ID mechanism and
1016 replace it with a structure and associated callback type. This
1017 mechanism allows a numeric "hash" to be extracted from a thread ID in
1018 either case, and on platforms where pointers are larger than 'long',
1019 mixing is done to help ensure the numeric 'hash' is usable even if it
1020 can't be guaranteed unique. The default mechanism is to use "&errno"
1021 as a pointer-based thread ID to distinguish between threads.
1023 Applications that want to provide their own thread IDs should now use
1024 CRYPTO_THREADID_set_callback() to register a callback that will call
1025 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
1027 Note that ERR_remove_state() is now deprecated, because it is tied
1028 to the assumption that thread IDs are numeric. ERR_remove_state(0)
1029 to free the current thread's error state should be replaced by
1030 ERR_remove_thread_state(NULL).
1032 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
1033 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
1034 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
1035 application was previously providing a numeric thread callback that
1036 was inappropriate for distinguishing threads, then uniqueness might
1037 have been obtained with &errno that happened immediately in the
1038 intermediate development versions of OpenSSL; this is no longer the
1039 case, the numeric thread callback will now override the automatic use
1041 [Geoff Thorpe, with help from Bodo Moeller]
1043 *) Initial support for different CRL issuing certificates. This covers a
1044 simple case where the self issued certificates in the chain exist and
1045 the real CRL issuer is higher in the existing chain.
1047 This work was sponsored by Google.
1050 *) Removed effectively defunct crypto/store from the build.
1053 *) Revamp of STACK to provide stronger type-checking. Still to come:
1054 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
1055 ASN1_STRING, CONF_VALUE.
1058 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
1059 RAM on SSL connections. This option can save about 34k per idle SSL.
1062 *) Revamp of LHASH to provide stronger type-checking. Still to come:
1063 STACK, TXT_DB, bsearch, qsort.
1066 *) Initial support for Cryptographic Message Syntax (aka CMS) based
1067 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
1068 support for data, signedData, compressedData, digestedData and
1069 encryptedData, envelopedData types included. Scripts to check against
1070 RFC4134 examples draft and interop and consistency checks of many
1071 content types and variants.
1074 *) Add options to enc utility to support use of zlib compression BIO.
1077 *) Extend mk1mf to support importing of options and assembly language
1078 files from Configure script, currently only included in VC-WIN32.
1079 The assembly language rules can now optionally generate the source
1080 files from the associated perl scripts.
1083 *) Implement remaining functionality needed to support GOST ciphersuites.
1084 Interop testing has been performed using CryptoPro implementations.
1085 [Victor B. Wagner <vitus@cryptocom.ru>]
1087 *) s390x assembler pack.
1090 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
1094 *) Implement Opaque PRF Input TLS extension as specified in
1095 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
1096 official specification yet and no extension type assignment by
1097 IANA exists, this extension (for now) will have to be explicitly
1098 enabled when building OpenSSL by providing the extension number
1099 to use. For example, specify an option
1101 -DTLSEXT_TYPE_opaque_prf_input=0x9527
1103 to the "config" or "Configure" script to enable the extension,
1104 assuming extension number 0x9527 (which is a completely arbitrary
1105 and unofficial assignment based on the MD5 hash of the Internet
1106 Draft). Note that by doing so, you potentially lose
1107 interoperability with other TLS implementations since these might
1108 be using the same extension number for other purposes.
1110 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
1111 opaque PRF input value to use in the handshake. This will create
1112 an interal copy of the length-'len' string at 'src', and will
1113 return non-zero for success.
1115 To get more control and flexibility, provide a callback function
1118 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
1119 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
1123 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
1126 Callback function 'cb' will be called in handshakes, and is
1127 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
1128 Argument 'arg' is for application purposes (the value as given to
1129 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
1130 be provided to the callback function). The callback function
1131 has to return non-zero to report success: usually 1 to use opaque
1132 PRF input just if possible, or 2 to enforce use of the opaque PRF
1133 input. In the latter case, the library will abort the handshake
1134 if opaque PRF input is not successfully negotiated.
1136 Arguments 'peerinput' and 'len' given to the callback function
1137 will always be NULL and 0 in the case of a client. A server will
1138 see the client's opaque PRF input through these variables if
1139 available (NULL and 0 otherwise). Note that if the server
1140 provides an opaque PRF input, the length must be the same as the
1141 length of the client's opaque PRF input.
1143 Note that the callback function will only be called when creating
1144 a new session (session resumption can resume whatever was
1145 previously negotiated), and will not be called in SSL 2.0
1146 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
1147 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
1148 for applications that need to enforce opaque PRF input.
1152 *) Update ssl code to support digests other than SHA1+MD5 for handshake
1155 [Victor B. Wagner <vitus@cryptocom.ru>]
1157 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1158 RFC4507bis. The encrypted ticket format is an encrypted encoded
1159 SSL_SESSION structure, that way new session features are automatically
1162 If a client application caches session in an SSL_SESSION structure
1163 support is transparent because tickets are now stored in the encoded
1166 The SSL_CTX structure automatically generates keys for ticket
1167 protection in servers so again support should be possible
1168 with no application modification.
1170 If a client or server wishes to disable RFC4507 support then the option
1171 SSL_OP_NO_TICKET can be set.
1173 Add a TLS extension debugging callback to allow the contents of any client
1174 or server extensions to be examined.
1176 This work was sponsored by Google.
1179 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1180 OpenSSL should now compile cleanly on gcc 4.2
1181 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1183 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1184 support including streaming MAC support: this is required for GOST
1185 ciphersuite support.
1186 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1188 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1189 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1190 to output in BER and PEM format.
1193 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1194 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1195 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1196 ENGINE support for HMAC keys which are unextractable. New -mac and
1197 -macopt options to dgst utility.
1200 *) New option -sigopt to dgst utility. Update dgst to use
1201 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1202 alternative signing paramaters such as X9.31 or PSS in the dgst
1206 *) Change ssl_cipher_apply_rule(), the internal function that does
1207 the work each time a ciphersuite string requests enabling
1208 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1209 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1210 the order of disabled ciphersuites such that those ciphersuites
1211 that most recently went from enabled to disabled not only stay
1212 in order with respect to each other, but also have higher priority
1213 than other disabled ciphersuites the next time ciphersuites are
1216 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1217 the same ciphersuites as with "HIGH" alone, but in a specific
1218 order where the PSK ciphersuites come first (since they are the
1219 most recently disabled ciphersuites when "HIGH" is parsed).
1221 Also, change ssl_create_cipher_list() (using this new
1222 funcionality) such that between otherwise identical
1223 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
1227 *) Change ssl_create_cipher_list() so that it automatically
1228 arranges the ciphersuites in reasonable order before starting
1229 to process the rule string. Thus, the definition for "DEFAULT"
1230 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
1231 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
1232 This makes it much easier to arrive at a reasonable default order
1233 in applications for which anonymous ciphers are OK (meaning
1234 that you can't actually use DEFAULT).
1235 [Bodo Moeller; suggested by Victor Duchovni]
1237 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
1238 processing) into multiple integers instead of setting
1239 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
1240 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
1241 (These masks as well as the individual bit definitions are hidden
1242 away into the non-exported interface ssl/ssl_locl.h, so this
1243 change to the definition of the SSL_CIPHER structure shouldn't
1244 affect applications.) This give us more bits for each of these
1245 categories, so there is no longer a need to coagulate AES128 and
1246 AES256 into a single algorithm bit, and to coagulate Camellia128
1247 and Camellia256 into a single algorithm bit, which has led to all
1250 Thus, among other things, the kludge introduced in 0.9.7m and
1251 0.9.8e for masking out AES256 independently of AES128 or masking
1252 out Camellia256 independently of AES256 is not needed here in 0.9.9.
1254 With the change, we also introduce new ciphersuite aliases that
1255 so far were missing: "AES128", "AES256", "CAMELLIA128", and
1259 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
1260 Use the leftmost N bytes of the signature input if the input is
1261 larger than the prime q (with N being the size in bytes of q).
1264 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1265 it yet and it is largely untested.
1268 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1271 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1272 some compilers (gcc 4.2 and later) reject their use. Safestack is
1273 reimplemented. Update ASN1 to avoid use of legacy functions.
1276 *) Win32/64 targets are linked with Winsock2.
1279 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1280 to external functions. This can be used to increase CRL handling
1281 efficiency especially when CRLs are very large by (for example) storing
1282 the CRL revoked certificates in a database.
1285 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1286 new CRLs added to a directory can be used. New command line option
1287 -verify_return_error to s_client and s_server. This causes real errors
1288 to be returned by the verify callback instead of carrying on no matter
1289 what. This reflects the way a "real world" verify callback would behave.
1292 *) GOST engine, supporting several GOST algorithms and public key formats.
1293 Kindly donated by Cryptocom.
1296 *) Partial support for Issuing Distribution Point CRL extension. CRLs
1297 partitioned by DP are handled but no indirect CRL or reason partitioning
1298 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
1299 selected via a scoring technique which handles IDP and AKID in CRLs.
1302 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1303 will ultimately be used for all verify operations: this will remove the
1304 X509_STORE dependency on certificate verification and allow alternative
1305 lookup methods. X509_STORE based implementations of these two callbacks.
1308 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1309 Modify get_crl() to find a valid (unexpired) CRL if possible.
1312 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1313 this would be called X509_CRL_cmp() but that name is already used by
1314 a function that just compares CRL issuer names. Cache several CRL
1315 extensions in X509_CRL structure and cache CRLDP in X509.
1318 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1319 this maps equivalent X509_NAME structures into a consistent structure.
1320 Name comparison can then be performed rapidly using memcmp().
1323 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1327 *) Allow digests to supply their own micalg string for S/MIME type using
1328 the ctrl EVP_MD_CTRL_MICALG.
1331 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1332 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1333 ctrl. It can then customise the structure before and/or after signing
1337 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1338 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1339 to free up any added signature OIDs.
1342 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1343 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1344 digest and cipher tables. New options added to openssl utility:
1345 list-message-digest-algorithms and list-cipher-algorithms.
1348 *) Change the array representation of binary polynomials: the list
1349 of degrees of non-zero coefficients is now terminated with -1.
1350 Previously it was terminated with 0, which was also part of the
1351 value; thus, the array representation was not applicable to
1352 polynomials where t^0 has coefficient zero. This change makes
1353 the array representation useful in a more general context.
1356 *) Various modifications and fixes to SSL/TLS cipher string
1357 handling. For ECC, the code now distinguishes between fixed ECDH
1358 with RSA certificates on the one hand and with ECDSA certificates
1359 on the other hand, since these are separate ciphersuites. The
1360 unused code for Fortezza ciphersuites has been removed.
1362 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1363 (not "ECDHE"). For consistency with the code for DH
1364 certificates, use of ECDH certificates is now considered ECDH
1365 authentication, not RSA or ECDSA authentication (the latter is
1366 merely the CA's signing algorithm and not actively used in the
1369 The temporary ciphersuite alias "ECCdraft" is no longer
1370 available, and ECC ciphersuites are no longer excluded from "ALL"
1371 and "DEFAULT". The following aliases now exist for RFC 4492
1372 ciphersuites, most of these by analogy with the DH case:
1374 kECDHr - ECDH cert, signed with RSA
1375 kECDHe - ECDH cert, signed with ECDSA
1376 kECDH - ECDH cert (signed with either RSA or ECDSA)
1377 kEECDH - ephemeral ECDH
1378 ECDH - ECDH cert or ephemeral ECDH
1384 AECDH - anonymous ECDH
1385 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1389 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1390 Use correct micalg parameters depending on digest(s) in signed message.
1393 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1394 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1397 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1398 an engine to register a method. Add ENGINE lookups for methods and
1399 functional reference processing.
1402 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1403 EVP_{Sign,Verify}* which allow an application to customise the signature
1407 *) New -resign option to smime utility. This adds one or more signers
1408 to an existing PKCS#7 signedData structure. Also -md option to use an
1409 alternative message digest algorithm for signing.
1412 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1413 create PKCS7 structures containing multiple signers. Update smime
1414 application to support multiple signers.
1417 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1421 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1422 Reorganize PBE internals to lookup from a static table using NIDs,
1423 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1424 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1425 PRF which will be automatically used with PBES2.
1428 *) Replace the algorithm specific calls to generate keys in "req" with the
1432 *) Update PKCS#7 enveloped data routines to use new API. This is now
1433 supported by any public key method supporting the encrypt operation. A
1434 ctrl is added to allow the public key algorithm to examine or modify
1435 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1439 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1440 a default digest type to use. In most cases this will be SHA1 but some
1441 algorithms (such as GOST) need to specify an alternative digest. The
1442 return value indicates how strong the prefernce is 1 means optional and
1443 2 is mandatory (that is it is the only supported type). Modify
1444 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1445 use the default md. Update openssl utilities to use the default digest
1446 type for signing if it is not explicitly indicated.
1449 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1450 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1451 signing method from the key type. This effectively removes the link
1452 between digests and public key types.
1455 *) Add an OID cross reference table and utility functions. Its purpose is to
1456 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1457 rsaEncryption. This will allow some of the algorithm specific hackery
1458 needed to use the correct OID to be removed.
1461 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1462 structures for PKCS7_sign(). They are now set up by the relevant public
1466 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1469 *) Add support for key derivation (agreement) in the API, DH method and
1473 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1474 public and private key formats. As a side effect these add additional
1475 command line functionality not previously available: DSA signatures can be
1476 generated and verified using pkeyutl and DH key support and generation in
1481 [Oliver Tappe <zooey@hirschkaefer.de>]
1483 *) New make target "install_html_docs" installs HTML renditions of the
1485 [Oliver Tappe <zooey@hirschkaefer.de>]
1487 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1488 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1489 support key and parameter generation and add initial key generation
1490 functionality for RSA.
1493 *) Add functions for main EVP_PKEY_method operations. The undocumented
1494 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1495 EVP_PKEY_{encrypt,decrypt}_old.
1498 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1499 key API, doesn't do much yet.
1502 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1503 public key algorithms. New option to openssl utility:
1504 "list-public-key-algorithms" to print out info.
1507 *) Implement the Supported Elliptic Curves Extension for
1508 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1511 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1512 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1515 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1516 utilities such as rsa, dsa, dsaparam etc except they process any key
1520 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1521 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1522 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1526 *) Initial support for pluggable public key ASN1.
1527 De-spaghettify the public key ASN1 handling. Move public and private
1528 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1529 algorithm specific handling to a single module within the relevant
1530 algorithm directory. Add functions to allow (near) opaque processing
1531 of public and private key structures.
1534 *) Implement the Supported Point Formats Extension for
1535 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1538 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1539 for the psk identity [hint] and the psk callback functions to the
1540 SSL_SESSION, SSL and SSL_CTX structure.
1543 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1547 SSL_CTX_use_psk_identity_hint
1548 SSL_get_psk_identity_hint
1549 SSL_get_psk_identity
1550 SSL_use_psk_identity_hint
1552 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1554 *) Add RFC 3161 compliant time stamp request creation, response generation
1555 and response verification functionality.
1556 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1558 *) Add initial support for TLS extensions, specifically for the server_name
1559 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1560 have new members for a host name. The SSL data structure has an
1561 additional member SSL_CTX *initial_ctx so that new sessions can be
1562 stored in that context to allow for session resumption, even after the
1563 SSL has been switched to a new SSL_CTX in reaction to a client's
1564 server_name extension.
1566 New functions (subject to change):
1568 SSL_get_servername()
1569 SSL_get_servername_type()
1572 New CTRL codes and macros (subject to change):
1574 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1575 - SSL_CTX_set_tlsext_servername_callback()
1576 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1577 - SSL_CTX_set_tlsext_servername_arg()
1578 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1580 openssl s_client has a new '-servername ...' option.
1582 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1583 '-key2 ...', '-servername_fatal' (subject to change). This allows
1584 testing the HostName extension for a specific single host name ('-cert'
1585 and '-key' remain fallbacks for handshakes without HostName
1586 negotiation). If the unrecogninzed_name alert has to be sent, this by
1587 default is a warning; it becomes fatal with the '-servername_fatal'
1590 [Peter Sylvester, Remy Allais, Christophe Renou]
1592 *) Whirlpool hash implementation is added.
1595 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1596 bn(64,32). Because of instruction set limitations it doesn't have
1597 any negative impact on performance. This was done mostly in order
1598 to make it possible to share assembler modules, such as bn_mul_mont
1599 implementations, between 32- and 64-bit builds without hassle.
1602 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1603 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1607 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1608 dedicated Montgomery multiplication procedure, is introduced.
1609 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1610 "64-bit" performance on certain 32-bit targets.
1613 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1614 in SSL structures. New SSL ctrl to set maximum send fragment size.
1615 Save memory by seeting the I/O buffer sizes dynamically instead of
1616 using the maximum available value.
1619 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1620 in addition to the text details.
1623 *) Very, very preliminary EXPERIMENTAL support for printing of general
1624 ASN1 structures. This currently produces rather ugly output and doesn't
1625 handle several customised structures at all.
1628 *) Integrated support for PVK file format and some related formats such
1629 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1630 these in the 'rsa' and 'dsa' utilities.
1633 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1636 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1637 place for the (very old) "NETSCAPE" format certificates which are now
1638 handled using new ASN1 code equivalents.
1641 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1642 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1643 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1646 *) Modify CRL distribution points extension code to print out previously
1647 unsupported fields. Enhance extension setting code to allow setting of
1651 *) Add print and set support for Issuing Distribution Point CRL extension.
1654 *) Change 'Configure' script to enable Camellia by default.
1657 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
1659 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1660 Thanks to Antonio Martin, Enterprise Secure Access Research and
1661 Development, Cisco Systems, Inc. for discovering this bug and
1662 preparing a fix. (CVE-2012-0050)
1665 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
1667 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1668 of the Vaudenay padding oracle attack on CBC mode encryption
1669 which enables an efficient plaintext recovery attack against
1670 the OpenSSL implementation of DTLS. Their attack exploits timing
1671 differences arising during decryption processing. A research
1672 paper describing this attack can be found at:
1673 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1674 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1675 Security Group at Royal Holloway, University of London
1676 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1677 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1678 for preparing the fix. (CVE-2011-4108)
1679 [Robin Seggelmann, Michael Tuexen]
1681 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
1682 [Ben Laurie, Kasper <ekasper@google.com>]
1684 *) Clear bytes used for block padding of SSL 3.0 records.
1686 [Adam Langley (Google)]
1688 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1689 Kadianakis <desnacked@gmail.com> for discovering this issue and
1690 Adam Langley for preparing the fix. (CVE-2011-4619)
1691 [Adam Langley (Google)]
1693 *) Prevent malformed RFC3779 data triggering an assertion failure.
1694 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1695 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1696 [Rob Austein <sra@hactrn.net>]
1698 *) Fix ssl_ciph.c set-up race.
1699 [Adam Langley (Google)]
1701 *) Fix spurious failures in ecdsatest.c.
1702 [Emilia Käsper (Google)]
1704 *) Fix the BIO_f_buffer() implementation (which was mixing different
1705 interpretations of the '..._len' fields).
1706 [Adam Langley (Google)]
1708 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1709 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1710 threads won't reuse the same blinding coefficients.
1712 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1713 lock to call BN_BLINDING_invert_ex, and avoids one use of
1714 BN_BLINDING_update for each BN_BLINDING structure (previously,
1715 the last update always remained unused).
1716 [Emilia Käsper (Google)]
1718 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1719 for multi-threaded use of ECDH.
1720 [Adam Langley (Google)]
1722 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1725 *) Add protection against ECDSA timing attacks as mentioned in the paper
1726 by Billy Bob Brumley and Nicola Tuveri, see:
1728 http://eprint.iacr.org/2011/232.pdf
1730 [Billy Bob Brumley and Nicola Tuveri]
1732 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
1734 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1735 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1737 *) Fix bug in string printing code: if *any* escaping is enabled we must
1738 escape the escape character (backslash) or the resulting string is
1742 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
1744 *) Disable code workaround for ancient and obsolete Netscape browsers
1745 and servers: an attacker can use it in a ciphersuite downgrade attack.
1746 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1749 *) Fixed J-PAKE implementation error, originally discovered by
1750 Sebastien Martini, further info and confirmation from Stefan
1751 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1754 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
1756 *) Fix extension code to avoid race conditions which can result in a buffer
1757 overrun vulnerability: resumed sessions must not be modified as they can
1758 be shared by multiple threads. CVE-2010-3864
1761 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
1764 *) Don't reencode certificate when calculating signature: cache and use
1765 the original encoding instead. This makes signature verification of
1766 some broken encodings work correctly.
1769 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
1770 is also one of the inputs.
1771 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1773 *) Don't repeatedly append PBE algorithms to table if they already exist.
1774 Sort table on each new add. This effectively makes the table read only
1775 after all algorithms are added and subsequent calls to PKCS12_pbe_add
1779 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
1781 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
1784 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
1785 access or freeing data twice (CVE-2010-0742)
1786 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
1788 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
1789 common in certificates and some applications which only call
1790 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
1794 Reduce copying into .apps and .test in makevms.com
1795 Don't try to use blank CA certificate in CA.com
1796 Allow use of C files from original directories in maketests.com
1797 [Steven M. Schweda" <sms@antinode.info>]
1799 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1801 *) When rejecting SSL/TLS records due to an incorrect version number, never
1802 update s->server with a new major version number. As of
1803 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1804 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1805 the previous behavior could result in a read attempt at NULL when
1806 receiving specific incorrect SSL/TLS records once record payload
1807 protection is active. (CVE-2010-0740)
1808 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1810 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1811 could be crashed if the relevant tables were not present (e.g. chrooted).
1812 [Tomas Hoger <thoger@redhat.com>]
1814 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1816 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1817 [Martin Olsson, Neel Mehta]
1819 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1820 accommodate for stack sorting, always a write lock!).
1823 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1824 excessive delays in the RAND_poll(): over a minute. As a workaround
1825 include a time check in the inner Heap32Next loop too.
1828 *) The code that handled flushing of data in SSL/TLS originally used the
1829 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1830 the problem outlined in PR#1949. The fix suggested there however can
1831 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1832 of Apache). So instead simplify the code to flush unconditionally.
1833 This should be fine since flushing with no data to flush is a no op.
1836 *) Handle TLS versions 2.0 and later properly and correctly use the
1837 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1838 off ancient servers have a habit of sticking around for a while...
1841 *) Modify compression code so it frees up structures without using the
1842 ex_data callbacks. This works around a problem where some applications
1843 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1844 restarting) then use compression (e.g. SSL with compression) later.
1845 This results in significant per-connection memory leaks and
1846 has caused some security issues including CVE-2008-1678 and
1850 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1851 change when encrypting or decrypting.
1854 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1855 connect and renegotiate with servers which do not support RI.
1856 Until RI is more widely deployed this option is enabled by default.
1859 *) Add "missing" ssl ctrls to clear options and mode.
1862 *) If client attempts to renegotiate and doesn't support RI respond with
1863 a no_renegotiation alert as required by RFC5746. Some renegotiating
1864 TLS clients will continue a connection gracefully when they receive
1865 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1866 waiting for a server hello which it will never receive. Now we treat a
1867 received no_renegotiation alert as a fatal error. This is because
1868 applications requesting a renegotiation might well expect it to succeed
1869 and would have no code in place to handle the server denying it so the
1870 only safe thing to do is to terminate the connection.
1873 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1874 peer supports secure renegotiation and 0 otherwise. Print out peer
1875 renegotiation support in s_client/s_server.
1878 *) Replace the highly broken and deprecated SPKAC certification method with
1879 the updated NID creation version. This should correctly handle UTF8.
1882 *) Implement RFC5746. Re-enable renegotiation but require the extension
1883 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
1884 turns out to be a bad idea. It has been replaced by
1885 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
1886 SSL_CTX_set_options(). This is really not recommended unless you
1887 know what you are doing.
1888 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
1890 *) Fixes to stateless session resumption handling. Use initial_ctx when
1891 issuing and attempting to decrypt tickets in case it has changed during
1892 servername handling. Use a non-zero length session ID when attempting
1893 stateless session resumption: this makes it possible to determine if
1894 a resumption has occurred immediately after receiving server hello
1895 (several places in OpenSSL subtly assume this) instead of later in
1899 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1900 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1901 fixes for a few places where the return code is not checked
1903 [Julia Lawall <julia@diku.dk>]
1905 *) Add --strict-warnings option to Configure script to include devteam
1906 warnings in other configurations.
1909 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1910 makes it possible to install openssl libraries in locations which
1911 have names other than "lib", for example "/usr/lib64" which some
1913 [Steve Henson, based on patch from Jeremy Utley]
1915 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1916 X690 8.9.12 and can produce some misleading textual output of OIDs.
1917 [Steve Henson, reported by Dan Kaminsky]
1919 *) Delete MD2 from algorithm tables. This follows the recommendation in
1920 several standards that it is not used in new applications due to
1921 several cryptographic weaknesses. For binary compatibility reasons
1922 the MD2 API is still compiled in by default.
1925 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1929 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1930 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1932 [Guenter <lists@gknw.net>]
1934 *) Fix the server certificate chain building code to use X509_verify_cert(),
1935 it used to have an ad-hoc builder which was unable to cope with anything
1936 other than a simple chain.
1937 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1939 *) Don't check self signed certificate signatures in X509_verify_cert()
1940 by default (a flag can override this): it just wastes time without
1941 adding any security. As a useful side effect self signed root CAs
1942 with non-FIPS digests are now usable in FIPS mode.
1945 *) In dtls1_process_out_of_seq_message() the check if the current message
1946 is already buffered was missing. For every new message was memory
1947 allocated, allowing an attacker to perform an denial of service attack
1948 with sending out of seq handshake messages until there is no memory
1949 left. Additionally every future messege was buffered, even if the
1950 sequence number made no sense and would be part of another handshake.
1951 So only messages with sequence numbers less than 10 in advance will be
1952 buffered. (CVE-2009-1378)
1953 [Robin Seggelmann, discovered by Daniel Mentz]
1955 *) Records are buffered if they arrive with a future epoch to be
1956 processed after finishing the corresponding handshake. There is
1957 currently no limitation to this buffer allowing an attacker to perform
1958 a DOS attack with sending records with future epochs until there is no
1959 memory left. This patch adds the pqueue_size() function to detemine
1960 the size of a buffer and limits the record buffer to 100 entries.
1962 [Robin Seggelmann, discovered by Daniel Mentz]
1964 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
1965 parent structure is freed. (CVE-2009-1379)
1968 *) Handle non-blocking I/O properly in SSL_shutdown() call.
1969 [Darryl Miles <darryl-mailinglists@netbauds.net>]
1972 [Ilya O. <vrghost@gmail.com>]
1974 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
1976 *) Disable renegotiation completely - this fixes a severe security
1977 problem (CVE-2009-3555) at the cost of breaking all
1978 renegotiation. Renegotiation can be re-enabled by setting
1979 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
1980 run-time. This is really not recommended unless you know what
1984 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
1986 *) Don't set val to NULL when freeing up structures, it is freed up by
1987 underlying code. If sizeof(void *) > sizeof(long) this can result in
1988 zeroing past the valid field. (CVE-2009-0789)
1989 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
1991 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
1992 checked correctly. This would allow some invalid signed attributes to
1993 appear to verify correctly. (CVE-2009-0591)
1994 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1996 *) Reject UniversalString and BMPString types with invalid lengths. This
1997 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
1998 a legal length. (CVE-2009-0590)
2001 *) Set S/MIME signing as the default purpose rather than setting it
2002 unconditionally. This allows applications to override it at the store
2006 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2007 to handle some structures.
2010 *) Improve efficiency of mem_gets: don't search whole buffer each time
2012 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2014 *) New -hex option for openssl rand.
2017 *) Print out UTF8String and NumericString when parsing ASN1.
2020 *) Support NumericString type for name components.
2023 *) Allow CC in the environment to override the automatically chosen
2024 compiler. Note that nothing is done to ensure flags work with the
2028 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2030 *) Properly check EVP_VerifyFinal() and similar return values
2032 [Ben Laurie, Bodo Moeller, Google Security Team]
2034 *) Enable TLS extensions by default.
2037 *) Allow the CHIL engine to be loaded, whether the application is
2038 multithreaded or not. (This does not release the developer from the
2039 obligation to set up the dynamic locking callbacks.)
2040 [Sander Temme <sander@temme.net>]
2042 *) Use correct exit code if there is an error in dgst command.
2043 [Steve Henson; problem pointed out by Roland Dirlewanger]
2045 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2046 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2049 *) Add experimental JPAKE support, including demo authentication in
2050 s_client and s_server.
2053 *) Set the comparison function in v3_addr_canonize().
2054 [Rob Austein <sra@hactrn.net>]
2056 *) Add support for XMPP STARTTLS in s_client.
2057 [Philip Paeps <philip@freebsd.org>]
2059 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2060 to ensure that even with this option, only ciphersuites in the
2061 server's preference list will be accepted. (Note that the option
2062 applies only when resuming a session, so the earlier behavior was
2063 just about the algorithm choice for symmetric cryptography.)
2066 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2068 *) Fix NULL pointer dereference if a DTLS server received
2069 ChangeCipherSpec as first record (CVE-2009-1386).
2072 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
2073 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2076 *) The fix in 0.9.8c that supposedly got rid of unsafe
2077 double-checked locking was incomplete for RSA blinding,
2078 addressing just one layer of what turns out to have been
2079 doubly unsafe triple-checked locking.
2081 So now fix this for real by retiring the MONT_HELPER macro
2082 in crypto/rsa/rsa_eay.c.
2084 [Bodo Moeller; problem pointed out by Marius Schilder]
2086 *) Various precautionary measures:
2088 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2090 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2091 (NB: This would require knowledge of the secret session ticket key
2092 to exploit, in which case you'd be SOL either way.)
2094 - Change bn_nist.c so that it will properly handle input BIGNUMs
2095 outside the expected range.
2097 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2100 [Neel Mehta, Bodo Moeller]
2102 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2103 the load fails. Useful for distros.
2104 [Ben Laurie and the FreeBSD team]
2106 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2109 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2112 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2114 This work was sponsored by Logica.
2117 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2118 keystores. Support for SSL/TLS client authentication too.
2119 Not compiled unless enable-capieng specified to Configure.
2121 This work was sponsored by Logica.
2124 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2125 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2126 attribute creation routines such as certifcate requests and PKCS#12
2130 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2132 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2133 handshake which could lead to a cilent crash as found using the
2134 Codenomicon TLS test suite (CVE-2008-1672)
2135 [Steve Henson, Mark Cox]
2137 *) Fix double free in TLS server name extensions which could lead to
2138 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
2141 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
2143 Clear the error queue to ensure that error entries left from
2144 older function calls do not interfere with the correct operation.
2145 [Lutz Jaenicke, Erik de Castro Lopo]
2147 *) Remove root CA certificates of commercial CAs:
2149 The OpenSSL project does not recommend any specific CA and does not
2150 have any policy with respect to including or excluding any CA.
2151 Therefore it does not make any sense to ship an arbitrary selection
2152 of root CA certificates with the OpenSSL software.
2155 *) RSA OAEP patches to fix two separate invalid memory reads.
2156 The first one involves inputs when 'lzero' is greater than
2157 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
2158 before the beginning of from). The second one involves inputs where
2159 the 'db' section contains nothing but zeroes (there is a one-byte
2160 invalid read after the end of 'db').
2161 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2163 *) Partial backport from 0.9.9-dev:
2165 Introduce bn_mul_mont (dedicated Montgomery multiplication
2166 procedure) as a candidate for BIGNUM assembler implementation.
2167 While 0.9.9-dev uses assembler for various architectures, only
2168 x86_64 is available by default here in the 0.9.8 branch, and
2169 32-bit x86 is available through a compile-time setting.
2171 To try the 32-bit x86 assembler implementation, use Configure
2172 option "enable-montasm" (which exists only for this backport).
2174 As "enable-montasm" for 32-bit x86 disclaims code stability
2175 anyway, in this constellation we activate additional code
2176 backported from 0.9.9-dev for further performance improvements,
2177 namely BN_from_montgomery_word. (To enable this otherwise,
2178 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
2180 [Andy Polyakov (backport partially by Bodo Moeller)]
2182 *) Add TLS session ticket callback. This allows an application to set
2183 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
2184 values. This is useful for key rollover for example where several key
2185 sets may exist with different names.
2188 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
2189 This was broken until now in 0.9.8 releases, such that the only way
2190 a registered ENGINE could be used (assuming it initialises
2191 successfully on the host) was to explicitly set it as the default
2192 for the relevant algorithms. This is in contradiction with 0.9.7
2193 behaviour and the documentation. With this fix, when an ENGINE is
2194 registered into a given algorithm's table of implementations, the
2195 'uptodate' flag is reset so that auto-discovery will be used next
2196 time a new context for that algorithm attempts to select an
2198 [Ian Lister (tweaked by Geoff Thorpe)]
2200 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
2201 implemention in the following ways:
2203 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
2206 Lack of BER streaming support means one pass streaming processing is
2207 only supported if data is detached: setting the streaming flag is
2208 ignored for embedded content.
2210 CMS support is disabled by default and must be explicitly enabled
2211 with the enable-cms configuration option.
2214 *) Update the GMP engine glue to do direct copies between BIGNUM and
2215 mpz_t when openssl and GMP use the same limb size. Otherwise the
2216 existing "conversion via a text string export" trick is still used.
2217 [Paul Sheer <paulsheer@gmail.com>]
2219 *) Zlib compression BIO. This is a filter BIO which compressed and
2220 uncompresses any data passed through it.
2223 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
2224 RFC3394 compatible AES key wrapping.
2227 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
2228 sets string data without copying. X509_ALGOR_set0() and
2229 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
2230 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
2231 from an X509_ATTRIBUTE structure optionally checking it occurs only
2232 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
2236 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
2237 to get the expected BN_FLG_CONSTTIME behavior.
2238 [Bodo Moeller (Google)]
2242 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
2243 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
2244 - added some more tests to do_tests.pl
2245 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
2246 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
2247 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
2248 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
2249 - various changes to netware.pl to enable gcc-cross builds on Win32
2251 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
2252 - various changes to fix missing prototype warnings
2253 - fixed x86nasm.pl to create correct asm files for NASM COFF output
2254 - added AES, WHIRLPOOL and CPUID assembler code to build files
2255 - added missing AES assembler make rules to mk1mf.pl
2256 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
2257 [Guenter Knauf <eflash@gmx.net>]
2259 *) Implement certificate status request TLS extension defined in RFC3546.
2260 A client can set the appropriate parameters and receive the encoded
2261 OCSP response via a callback. A server can query the supplied parameters
2262 and set the encoded OCSP response in the callback. Add simplified examples
2263 to s_client and s_server.
2266 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
2268 *) Fix various bugs:
2269 + Binary incompatibility of ssl_ctx_st structure
2270 + DTLS interoperation with non-compliant servers
2271 + Don't call get_session_cb() without proposed session
2272 + Fix ia64 assembler code
2273 [Andy Polyakov, Steve Henson]
2275 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
2277 *) DTLS Handshake overhaul. There were longstanding issues with
2278 OpenSSL DTLS implementation, which were making it impossible for
2279 RFC 4347 compliant client to communicate with OpenSSL server.
2280 Unfortunately just fixing these incompatibilities would "cut off"
2281 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
2282 server keeps tolerating non RFC compliant syntax. The opposite is
2283 not true, 0.9.8f client can not communicate with earlier server.
2284 This update even addresses CVE-2007-4995.
2287 *) Changes to avoid need for function casts in OpenSSL: some compilers
2288 (gcc 4.2 and later) reject their use.
2289 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
2292 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2293 RFC4507bis. The encrypted ticket format is an encrypted encoded
2294 SSL_SESSION structure, that way new session features are automatically
2297 If a client application caches session in an SSL_SESSION structure
2298 support is transparent because tickets are now stored in the encoded
2301 The SSL_CTX structure automatically generates keys for ticket
2302 protection in servers so again support should be possible
2303 with no application modification.
2305 If a client or server wishes to disable RFC4507 support then the option
2306 SSL_OP_NO_TICKET can be set.
2308 Add a TLS extension debugging callback to allow the contents of any client
2309 or server extensions to be examined.
2311 This work was sponsored by Google.
2314 *) Add initial support for TLS extensions, specifically for the server_name
2315 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2316 have new members for a host name. The SSL data structure has an
2317 additional member SSL_CTX *initial_ctx so that new sessions can be
2318 stored in that context to allow for session resumption, even after the
2319 SSL has been switched to a new SSL_CTX in reaction to a client's
2320 server_name extension.
2322 New functions (subject to change):
2324 SSL_get_servername()
2325 SSL_get_servername_type()
2328 New CTRL codes and macros (subject to change):
2330 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2331 - SSL_CTX_set_tlsext_servername_callback()
2332 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2333 - SSL_CTX_set_tlsext_servername_arg()
2334 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2336 openssl s_client has a new '-servername ...' option.
2338 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2339 '-key2 ...', '-servername_fatal' (subject to change). This allows
2340 testing the HostName extension for a specific single host name ('-cert'
2341 and '-key' remain fallbacks for handshakes without HostName
2342 negotiation). If the unrecogninzed_name alert has to be sent, this by
2343 default is a warning; it becomes fatal with the '-servername_fatal'
2346 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2348 *) Add AES and SSE2 assembly language support to VC++ build.
2351 *) Mitigate attack on final subtraction in Montgomery reduction.
2354 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2355 (which previously caused an internal error).
2358 *) Squeeze another 10% out of IGE mode when in != out.
2361 *) AES IGE mode speedup.
2362 [Dean Gaudet (Google)]
2364 *) Add the Korean symmetric 128-bit cipher SEED (see
2365 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2366 add SEED ciphersuites from RFC 4162:
2368 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2369 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2370 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2371 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2373 To minimize changes between patchlevels in the OpenSSL 0.9.8
2374 series, SEED remains excluded from compilation unless OpenSSL
2375 is configured with 'enable-seed'.
2376 [KISA, Bodo Moeller]
2378 *) Mitigate branch prediction attacks, which can be practical if a
2379 single processor is shared, allowing a spy process to extract
2380 information. For detailed background information, see
2381 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2382 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2383 and Necessary Software Countermeasures"). The core of the change
2384 are new versions BN_div_no_branch() and
2385 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2386 respectively, which are slower, but avoid the security-relevant
2387 conditional branches. These are automatically called by BN_div()
2388 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2389 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2390 remove a conditional branch.
2392 BN_FLG_CONSTTIME is the new name for the previous
2393 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2394 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2395 in the exponent causes BN_mod_exp_mont() to use the alternative
2396 implementation in BN_mod_exp_mont_consttime().) The old name
2397 remains as a deprecated alias.
2399 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2400 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2401 constant-time implementations for more than just exponentiation.
2402 Here too the old name is kept as a deprecated alias.
2404 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2405 the BN_BLINDING structure gets an independent copy of the
2406 modulus. This means that the previous "BIGNUM *m" argument to
2407 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2408 essentially becomes "const BIGNUM *m", although we can't actually
2409 change this in the header file before 0.9.9. It allows
2410 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2411 enable BN_FLG_CONSTTIME.
2413 [Matthew D Wood (Intel Corp)]
2415 *) In the SSL/TLS server implementation, be strict about session ID
2416 context matching (which matters if an application uses a single
2417 external cache for different purposes). Previously,
2418 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2419 set. This did ensure strict client verification, but meant that,
2420 with applications using a single external cache for quite
2421 different requirements, clients could circumvent ciphersuite
2422 restrictions for a given session ID context by starting a session
2423 in a different context.
2426 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2427 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2428 authentication-only ciphersuites.
2431 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2432 not complete and could lead to a possible single byte overflow
2433 (CVE-2007-5135) [Ben Laurie]
2435 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2437 *) Since AES128 and AES256 (and similarly Camellia128 and
2438 Camellia256) share a single mask bit in the logic of
2439 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2440 kludge to work properly if AES128 is available and AES256 isn't
2441 (or if Camellia128 is available and Camellia256 isn't).
2444 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2445 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2446 When a point or a seed is encoded in a BIT STRING, we need to
2447 prevent the removal of trailing zero bits to get the proper DER
2448 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2449 of a NamedBitList, for which trailing 0 bits need to be removed.)
2452 *) Have SSL/TLS server implementation tolerate "mismatched" record
2453 protocol version while receiving ClientHello even if the
2454 ClientHello is fragmented. (The server can't insist on the
2455 particular protocol version it has chosen before the ServerHello
2456 message has informed the client about his choice.)
2459 *) Add RFC 3779 support.
2460 [Rob Austein for ARIN, Ben Laurie]
2462 *) Load error codes if they are not already present instead of using a
2463 static variable. This allows them to be cleanly unloaded and reloaded.
2464 Improve header file function name parsing.
2467 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2468 or CAPABILITY handshake as required by RFCs.
2471 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2473 *) Introduce limits to prevent malicious keys being able to
2474 cause a denial of service. (CVE-2006-2940)
2475 [Steve Henson, Bodo Moeller]
2477 *) Fix ASN.1 parsing of certain invalid structures that can result
2478 in a denial of service. (CVE-2006-2937) [Steve Henson]
2480 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2481 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2483 *) Fix SSL client code which could crash if connecting to a
2484 malicious SSLv2 server. (CVE-2006-4343)
2485 [Tavis Ormandy and Will Drewry, Google Security Team]
2487 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2488 match only those. Before that, "AES256-SHA" would be interpreted
2489 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2490 the same strength classification in 0.9.7h) as we currently only
2491 have a single AES bit in the ciphersuite description bitmap.
2492 That change, however, also applied to ciphersuite strings such as
2493 "RC4-MD5" that intentionally matched multiple ciphersuites --
2494 namely, SSL 2.0 ciphersuites in addition to the more common ones
2495 from SSL 3.0/TLS 1.0.
2497 So we change the selection algorithm again: Naming an explicit
2498 ciphersuite selects this one ciphersuite, and any other similar
2499 ciphersuite (same bitmap) from *other* protocol versions.
2500 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2501 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2503 Since SSL 2.0 does not have any ciphersuites for which the
2504 128/256 bit distinction would be relevant, this works for now.
2505 The proper fix will be to use different bits for AES128 and
2506 AES256, which would have avoided the problems from the beginning;
2507 however, bits are scarce, so we can only do this in a new release
2508 (not just a patchlevel) when we can change the SSL_CIPHER
2509 definition to split the single 'unsigned long mask' bitmap into
2510 multiple values to extend the available space.
2514 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2516 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2517 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2519 *) Add AES IGE and biIGE modes.
2522 *) Change the Unix randomness entropy gathering to use poll() when
2523 possible instead of select(), since the latter has some
2524 undesirable limitations.
2525 [Darryl Miles via Richard Levitte and Bodo Moeller]
2527 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2528 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2529 cannot be implicitly activated as part of, e.g., the "AES" alias.
2530 However, please upgrade to OpenSSL 0.9.9[-dev] for
2531 non-experimental use of the ECC ciphersuites to get TLS extension
2532 support, which is required for curve and point format negotiation
2533 to avoid potential handshake problems.
2536 *) Disable rogue ciphersuites:
2538 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2539 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2540 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2542 The latter two were purportedly from
2543 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2546 Also deactivate the remaining ciphersuites from
2547 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2548 unofficial, and the ID has long expired.
2551 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2552 dual-core machines) and other potential thread-safety issues.
2555 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2556 versions), which is now available for royalty-free use
2557 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2558 Also, add Camellia TLS ciphersuites from RFC 4132.
2560 To minimize changes between patchlevels in the OpenSSL 0.9.8
2561 series, Camellia remains excluded from compilation unless OpenSSL
2562 is configured with 'enable-camellia'.
2565 *) Disable the padding bug check when compression is in use. The padding
2566 bug check assumes the first packet is of even length, this is not
2567 necessarily true if compresssion is enabled and can result in false
2568 positives causing handshake failure. The actual bug test is ancient
2569 code so it is hoped that implementations will either have fixed it by
2570 now or any which still have the bug do not support compression.
2573 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2575 *) When applying a cipher rule check to see if string match is an explicit
2576 cipher suite and only match that one cipher suite if it is.
2579 *) Link in manifests for VC++ if needed.
2580 [Austin Ziegler <halostatue@gmail.com>]
2582 *) Update support for ECC-based TLS ciphersuites according to
2583 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2584 TLS extensions, which are supported starting with the 0.9.9
2585 branch, not in the OpenSSL 0.9.8 branch).
2588 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2589 opaque EVP_CIPHER_CTX handling.
2592 *) Fixes and enhancements to zlib compression code. We now only use
2593 "zlib1.dll" and use the default __cdecl calling convention on Win32
2594 to conform with the standards mentioned here:
2595 http://www.zlib.net/DLL_FAQ.txt
2596 Static zlib linking now works on Windows and the new --with-zlib-include
2597 --with-zlib-lib options to Configure can be used to supply the location
2598 of the headers and library. Gracefully handle case where zlib library
2602 *) Several fixes and enhancements to the OID generation code. The old code
2603 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2604 handle numbers larger than ULONG_MAX, truncated printing and had a
2605 non standard OBJ_obj2txt() behaviour.
2608 *) Add support for building of engines under engine/ as shared libraries
2609 under VC++ build system.
2612 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2613 Hopefully, we will not see any false combination of paths any more.
2616 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2618 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2619 (part of SSL_OP_ALL). This option used to disable the
2620 countermeasure against man-in-the-middle protocol-version
2621 rollback in the SSL 2.0 server implementation, which is a bad
2622 idea. (CVE-2005-2969)
2624 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2625 for Information Security, National Institute of Advanced Industrial
2626 Science and Technology [AIST], Japan)]
2628 *) Add two function to clear and return the verify parameter flags.
2631 *) Keep cipherlists sorted in the source instead of sorting them at
2632 runtime, thus removing the need for a lock.
2635 *) Avoid some small subgroup attacks in Diffie-Hellman.
2636 [Nick Mathewson and Ben Laurie]
2638 *) Add functions for well-known primes.
2641 *) Extended Windows CE support.
2642 [Satoshi Nakamura and Andy Polyakov]
2644 *) Initialize SSL_METHOD structures at compile time instead of during
2645 runtime, thus removing the need for a lock.
2648 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2649 attempting to decrypt each encrypted key in turn. Add support to
2653 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2655 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2658 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2661 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2662 key into the same file any more.
2665 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2668 *) Add -utf8 command line and config file option to 'ca'.
2669 [Stefan <stf@udoma.org]
2671 *) Removed the macro des_crypt(), as it seems to conflict with some
2672 libraries. Use DES_crypt().
2675 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2676 involves renaming the source and generated shared-libs for
2677 both. The engines will accept the corrected or legacy ids
2678 ('ncipher' and '4758_cca' respectively) when binding. NB,
2679 this only applies when building 'shared'.
2680 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2682 *) Add attribute functions to EVP_PKEY structure. Modify
2683 PKCS12_create() to recognize a CSP name attribute and
2684 use it. Make -CSP option work again in pkcs12 utility.
2687 *) Add new functionality to the bn blinding code:
2688 - automatic re-creation of the BN_BLINDING parameters after
2689 a fixed number of uses (currently 32)
2690 - add new function for parameter creation
2691 - introduce flags to control the update behaviour of the
2692 BN_BLINDING parameters
2693 - hide BN_BLINDING structure
2694 Add a second BN_BLINDING slot to the RSA structure to improve
2695 performance when a single RSA object is shared among several
2699 *) Add support for DTLS.
2700 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2702 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2703 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2706 *) Remove buggy and incompletet DH cert support from
2707 ssl/ssl_rsa.c and ssl/s3_both.c
2710 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2711 the apps/openssl applications.
2714 *) Compile clean with "-Wall -Wmissing-prototypes
2715 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2716 DEBUG_SAFESTACK must also be set.
2719 *) Change ./Configure so that certain algorithms can be disabled by default.
2720 The new counterpiece to "no-xxx" is "enable-xxx".
2722 The patented RC5 and MDC2 algorithms will now be disabled unless
2723 "enable-rc5" and "enable-mdc2", respectively, are specified.
2725 (IDEA remains enabled despite being patented. This is because IDEA
2726 is frequently required for interoperability, and there is no license
2727 fee for non-commercial use. As before, "no-idea" can be used to
2728 avoid this algorithm.)
2732 *) Add processing of proxy certificates (see RFC 3820). This work was
2733 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2734 EGEE (Enabling Grids for E-science in Europe).
2737 *) RC4 performance overhaul on modern architectures/implementations, such
2738 as Intel P4, IA-64 and AMD64.
2741 *) New utility extract-section.pl. This can be used specify an alternative
2742 section number in a pod file instead of having to treat each file as
2743 a separate case in Makefile. This can be done by adding two lines to the
2746 =for comment openssl_section:XXX
2748 The blank line is mandatory.
2752 *) New arguments -certform, -keyform and -pass for s_client and s_server
2753 to allow alternative format key and certificate files and passphrase
2757 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2758 update associated structures and add various utility functions.
2760 Add new policy related verify parameters, include policy checking in
2761 standard verify code. Enhance 'smime' application with extra parameters
2762 to support policy checking and print out.
2765 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2766 Nehemiah processors. These extensions support AES encryption in hardware
2767 as well as RNG (though RNG support is currently disabled).
2768 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2770 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2773 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2774 [Andy Polyakov and a number of other people]
2776 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2777 implementation contributed by IBM.
2778 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2780 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2781 exponent rather than 'unsigned long'. There is a corresponding change to
2782 the new 'rsa_keygen' element of the RSA_METHOD structure.
2783 [Jelte Jansen, Geoff Thorpe]
2785 *) Functionality for creating the initial serial number file is now
2786 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2788 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2789 number file to 1, which is bound to cause problems. To avoid
2790 the problems while respecting compatibility between different 0.9.7
2791 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2792 CA.pl for serial number initialization. With the new release 0.9.8,
2793 we can fix the problem directly in the 'ca' utility.)
2796 *) Reduced header interdepencies by declaring more opaque objects in
2797 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2798 give fewer recursive includes, which could break lazy source code - so
2799 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2800 developers should define this symbol when building and using openssl to
2801 ensure they track the recommended behaviour, interfaces, [etc], but
2802 backwards-compatible behaviour prevails when this isn't defined.
2805 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2808 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2809 This will generate a random key of the appropriate length based on the
2810 cipher context. The EVP_CIPHER can provide its own random key generation
2811 routine to support keys of a specific form. This is used in the des and
2812 3des routines to generate a key of the correct parity. Update S/MIME
2813 code to use new functions and hence generate correct parity DES keys.
2814 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2815 valid (weak or incorrect parity).
2818 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2819 as looking them up. This is useful when the verified structure may contain
2820 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2821 present unless the new PKCS7_NO_CRL flag is asserted.
2824 *) Extend ASN1 oid configuration module. It now additionally accepts the
2827 shortName = some long name, 1.2.3.4
2830 *) Reimplemented the BN_CTX implementation. There is now no more static
2831 limitation on the number of variables it can handle nor the depth of the
2832 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2833 information can now expand as required, and rather than having a single
2834 static array of bignums, BN_CTX now uses a linked-list of such arrays
2835 allowing it to expand on demand whilst maintaining the usefulness of
2836 BN_CTX's "bundling".
2839 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2840 to allow all RSA operations to function using a single BN_CTX.
2843 *) Preliminary support for certificate policy evaluation and checking. This
2844 is initially intended to pass the tests outlined in "Conformance Testing
2845 of Relying Party Client Certificate Path Processing Logic" v1.07.
2848 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2849 remained unused and not that useful. A variety of other little bignum
2850 tweaks and fixes have also been made continuing on from the audit (see
2854 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2855 associated ASN1, EVP and SSL functions and old ASN1 macros.
2858 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2859 and this should never fail. So the return value from the use of
2860 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2861 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2864 *) BN_CTX_get() should return zero-valued bignums, providing the same
2865 initialised value as BN_new().
2866 [Geoff Thorpe, suggested by Ulf Möller]
2868 *) Support for inhibitAnyPolicy certificate extension.
2871 *) An audit of the BIGNUM code is underway, for which debugging code is
2872 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2873 is considered valid when processing BIGNUMs, and causes execution to
2874 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2875 further steps are taken to deliberately pollute unused data in BIGNUM
2876 structures to try and expose faulty code further on. For now, openssl will
2877 (in its default mode of operation) continue to tolerate the inconsistent
2878 forms that it has tolerated in the past, but authors and packagers should
2879 consider trying openssl and their own applications when compiled with
2880 these debugging symbols defined. It will help highlight potential bugs in
2881 their own code, and will improve the test coverage for OpenSSL itself. At
2882 some point, these tighter rules will become openssl's default to improve
2883 maintainability, though the assert()s and other overheads will remain only
2884 in debugging configurations. See bn.h for more details.
2885 [Geoff Thorpe, Nils Larsch, Ulf Möller]
2887 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
2888 that can only be obtained through BN_CTX_new() (which implicitly
2889 initialises it). The presence of this function only made it possible
2890 to overwrite an existing structure (and cause memory leaks).
2893 *) Because of the callback-based approach for implementing LHASH as a
2894 template type, lh_insert() adds opaque objects to hash-tables and
2895 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2896 to clean up those corresponding objects before destroying the hash table
2897 (and losing the object pointers). So some over-zealous constifications in
2898 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2899 objects as "const" and the lh_doall[_arg] callback wrappers are not
2900 prototyped to have "const" restrictions on the object pointers they are
2901 given (and so aren't required to cast them away any more).
2904 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2905 (speed) prefers to use its own implementation. The two implementations
2906 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2907 its object type properly exposed (MS_TM) instead of casting to/from "char
2908 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2909 aren't necessarily the greatest nomenclatures - but this is what was used
2910 internally to the implementation so I've used that for now.
2913 *) Ensure that deprecated functions do not get compiled when
2914 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2915 the self-tests were still using deprecated key-generation functions so
2916 these have been updated also.
2919 *) Reorganise PKCS#7 code to separate the digest location functionality
2920 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2921 New function PKCS7_set_digest() to set the digest type for PKCS#7
2922 digestedData type. Add additional code to correctly generate the
2923 digestedData type and add support for this type in PKCS7 initialization
2927 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2928 structure of type "other".
2931 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2932 sure the loop does correctly stop and breaking ("division by zero")
2933 modulus operations are not performed. The (pre-generated) prime
2934 table crypto/bn/bn_prime.h was already correct, but it could not be
2935 re-generated on some platforms because of the "division by zero"
2936 situation in the script.
2937 [Ralf S. Engelschall]
2939 *) Update support for ECC-based TLS ciphersuites according to
2940 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2941 SHA-1 now is only used for "small" curves (where the
2942 representation of a field element takes up to 24 bytes); for
2943 larger curves, the field element resulting from ECDH is directly
2944 used as premaster secret.
2945 [Douglas Stebila (Sun Microsystems Laboratories)]
2947 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
2948 curve secp160r1 to the tests.
2949 [Douglas Stebila (Sun Microsystems Laboratories)]
2951 *) Add the possibility to load symbols globally with DSO.
2952 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
2954 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
2955 control of the error stack.
2958 *) Add support for STORE in ENGINE.
2961 *) Add the STORE type. The intention is to provide a common interface
2962 to certificate and key stores, be they simple file-based stores, or
2963 HSM-type store, or LDAP stores, or...
2964 NOTE: The code is currently UNTESTED and isn't really used anywhere.
2967 *) Add a generic structure called OPENSSL_ITEM. This can be used to
2968 pass a list of arguments to any function as well as provide a way
2969 for a function to pass data back to the caller.
2972 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
2973 works like BUF_strdup() but can be used to duplicate a portion of
2974 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
2978 *) Add the function sk_find_ex() which works like sk_find(), but will
2979 return an index to an element even if an exact match couldn't be
2980 found. The index is guaranteed to point at the element where the
2981 searched-for key would be inserted to preserve sorting order.
2984 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
2985 takes an extra flags argument for optional functionality. Currently,
2986 the following flags are defined:
2988 OBJ_BSEARCH_VALUE_ON_NOMATCH
2989 This one gets OBJ_bsearch_ex() to return a pointer to the first
2990 element where the comparing function returns a negative or zero
2993 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
2994 This one gets OBJ_bsearch_ex() to return a pointer to the first
2995 element where the comparing function returns zero. This is useful
2996 if there are more than one element where the comparing function
3000 *) Make it possible to create self-signed certificates with 'openssl ca'
3001 in such a way that the self-signed certificate becomes part of the
3002 CA database and uses the same mechanisms for serial number generation
3003 as all other certificate signing. The new flag '-selfsign' enables
3004 this functionality. Adapt CA.sh and CA.pl.in.
3007 *) Add functionality to check the public key of a certificate request
3008 against a given private. This is useful to check that a certificate
3009 request can be signed by that key (self-signing).
3012 *) Make it possible to have multiple active certificates with the same
3013 subject in the CA index file. This is done only if the keyword
3014 'unique_subject' is set to 'no' in the main CA section (default
3015 if 'CA_default') of the configuration file. The value is saved
3016 with the database itself in a separate index attribute file,
3017 named like the index file with '.attr' appended to the name.
3020 *) Generate muti valued AVAs using '+' notation in config files for
3024 *) Support for nameConstraints certificate extension.
3027 *) Support for policyConstraints certificate extension.
3030 *) Support for policyMappings certificate extension.
3033 *) Make sure the default DSA_METHOD implementation only uses its
3034 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3035 and change its own handlers to be NULL so as to remove unnecessary
3036 indirection. This lets alternative implementations fallback to the
3037 default implementation more easily.
3040 *) Support for directoryName in GeneralName related extensions
3044 *) Make it possible to link applications using Makefile.shared.
3045 Make that possible even when linking against static libraries!
3048 *) Support for single pass processing for S/MIME signing. This now
3049 means that S/MIME signing can be done from a pipe, in addition
3050 cleartext signing (multipart/signed type) is effectively streaming
3051 and the signed data does not need to be all held in memory.
3053 This is done with a new flag PKCS7_STREAM. When this flag is set
3054 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3055 is done after the data is output (and digests calculated) in
3056 SMIME_write_PKCS7().
3059 *) Add full support for -rpath/-R, both in shared libraries and
3060 applications, at least on the platforms where it's known how
3064 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3065 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3066 will now compute a table of multiples of the generator that
3067 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3068 faster (notably in the case of a single point multiplication,
3069 scalar * generator).
3070 [Nils Larsch, Bodo Moeller]
3072 *) IPv6 support for certificate extensions. The various extensions
3073 which use the IP:a.b.c.d can now take IPv6 addresses using the
3074 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3078 *) Added an ENGINE that implements RSA by performing private key
3079 exponentiations with the GMP library. The conversions to and from
3080 GMP's mpz_t format aren't optimised nor are any montgomery forms
3081 cached, and on x86 it appears OpenSSL's own performance has caught up.
3082 However there are likely to be other architectures where GMP could
3083 provide a boost. This ENGINE is not built in by default, but it can be
3084 specified at Configure time and should be accompanied by the necessary
3085 linker additions, eg;
3086 ./config -DOPENSSL_USE_GMP -lgmp
3089 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3090 testing availability of engines with "-t" - the old behaviour is
3091 produced by increasing the feature's verbosity with "-tt".
3094 *) ECDSA routines: under certain error conditions uninitialized BN objects
3095 could be freed. Solution: make sure initialization is performed early
3096 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3100 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3101 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3102 software implementations. For DSA and DH, parameter generation can
3103 also be overriden by providing the appropriate method callbacks.
3106 *) Change the "progress" mechanism used in key-generation and
3107 primality testing to functions that take a new BN_GENCB pointer in
3108 place of callback/argument pairs. The new API functions have "_ex"
3109 postfixes and the older functions are reimplemented as wrappers for
3110 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3111 declarations of the old functions to help (graceful) attempts to
3112 migrate to the new functions. Also, the new key-generation API
3113 functions operate on a caller-supplied key-structure and return
3114 success/failure rather than returning a key or NULL - this is to
3115 help make "keygen" another member function of RSA_METHOD etc.
3117 Example for using the new callback interface:
3119 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3123 BN_GENCB_set(&my_cb, my_callback, my_arg);
3125 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3126 /* For the meaning of a, b in calls to my_callback(), see the
3127 * documentation of the function that calls the callback.
3128 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3129 * my_callback should return 1 if it wants BN_is_prime_ex()
3130 * to continue, or 0 to stop.
3135 *) Change the ZLIB compression method to be stateful, and make it
3136 available to TLS with the number defined in
3137 draft-ietf-tls-compression-04.txt.
3140 *) Add the ASN.1 structures and functions for CertificatePair, which
3141 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
3143 CertificatePair ::= SEQUENCE {
3144 forward [0] Certificate OPTIONAL,
3145 reverse [1] Certificate OPTIONAL,
3146 -- at least one of the pair shall be present -- }
3148 Also implement the PEM functions to read and write certificate
3149 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
3151 This needed to be defined, mostly for the sake of the LDAP
3152 attribute crossCertificatePair, but may prove useful elsewhere as
3156 *) Make it possible to inhibit symlinking of shared libraries in
3157 Makefile.shared, for Cygwin's sake.
3160 *) Extend the BIGNUM API by creating a function
3161 void BN_set_negative(BIGNUM *a, int neg);
3162 and a macro that behave like
3163 int BN_is_negative(const BIGNUM *a);
3165 to avoid the need to access 'a->neg' directly in applications.
3168 *) Implement fast modular reduction for pseudo-Mersenne primes
3169 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
3170 EC_GROUP_new_curve_GFp() will now automatically use this
3172 [Nils Larsch <nla@trustcenter.de>]
3174 *) Add new lock type (CRYPTO_LOCK_BN).
3177 *) Change the ENGINE framework to automatically load engines
3178 dynamically from specific directories unless they could be
3179 found to already be built in or loaded. Move all the
3180 current engines except for the cryptodev one to a new
3182 The engines in engines/ are built as shared libraries if
3183 the "shared" options was given to ./Configure or ./config.
3184 Otherwise, they are inserted in libcrypto.a.
3185 /usr/local/ssl/engines is the default directory for dynamic
3186 engines, but that can be overriden at configure time through
3187 the usual use of --prefix and/or --openssldir, and at run
3188 time with the environment variable OPENSSL_ENGINES.
3189 [Geoff Thorpe and Richard Levitte]
3191 *) Add Makefile.shared, a helper makefile to build shared
3192 libraries. Addapt Makefile.org.
3195 *) Add version info to Win32 DLLs.
3196 [Peter 'Luna' Runestig" <peter@runestig.com>]
3198 *) Add new 'medium level' PKCS#12 API. Certificates and keys
3199 can be added using this API to created arbitrary PKCS#12
3200 files while avoiding the low level API.
3202 New options to PKCS12_create(), key or cert can be NULL and
3203 will then be omitted from the output file. The encryption
3204 algorithm NIDs can be set to -1 for no encryption, the mac
3205 iteration count can be set to 0 to omit the mac.
3207 Enhance pkcs12 utility by making the -nokeys and -nocerts
3208 options work when creating a PKCS#12 file. New option -nomac
3209 to omit the mac, NONE can be set for an encryption algorithm.
3210 New code is modified to use the enhanced PKCS12_create()
3211 instead of the low level API.
3214 *) Extend ASN1 encoder to support indefinite length constructed
3215 encoding. This can output sequences tags and octet strings in
3216 this form. Modify pk7_asn1.c to support indefinite length
3217 encoding. This is experimental and needs additional code to
3218 be useful, such as an ASN1 bio and some enhanced streaming
3221 Extend template encode functionality so that tagging is passed
3222 down to the template encoder.
3225 *) Let 'openssl req' fail if an argument to '-newkey' is not
3226 recognized instead of using RSA as a default.
3229 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
3230 As these are not official, they are not included in "ALL";
3231 the "ECCdraft" ciphersuite group alias can be used to select them.
3232 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
3234 *) Add ECDH engine support.
3235 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
3237 *) Add ECDH in new directory crypto/ecdh/.
3238 [Douglas Stebila (Sun Microsystems Laboratories)]
3240 *) Let BN_rand_range() abort with an error after 100 iterations
3241 without success (which indicates a broken PRNG).
3244 *) Change BN_mod_sqrt() so that it verifies that the input value
3245 is really the square of the return value. (Previously,
3246 BN_mod_sqrt would show GIGO behaviour.)
3249 *) Add named elliptic curves over binary fields from X9.62, SECG,
3250 and WAP/WTLS; add OIDs that were still missing.
3252 [Sheueling Chang Shantz and Douglas Stebila
3253 (Sun Microsystems Laboratories)]
3255 *) Extend the EC library for elliptic curves over binary fields
3256 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
3259 EC_GF2m_simple_method
3263 EC_GROUP_new_curve_GF2m
3264 EC_GROUP_set_curve_GF2m
3265 EC_GROUP_get_curve_GF2m
3266 EC_POINT_set_affine_coordinates_GF2m
3267 EC_POINT_get_affine_coordinates_GF2m
3268 EC_POINT_set_compressed_coordinates_GF2m
3270 Point compression for binary fields is disabled by default for
3271 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
3274 As binary polynomials are represented as BIGNUMs, various members
3275 of the EC_GROUP and EC_POINT data structures can be shared
3276 between the implementations for prime fields and binary fields;
3277 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
3278 are essentially identical to their ..._GFp counterparts.
3279 (For simplicity, the '..._GFp' prefix has been dropped from
3280 various internal method names.)
3282 An internal 'field_div' method (similar to 'field_mul' and
3283 'field_sqr') has been added; this is used only for binary fields.
3285 [Sheueling Chang Shantz and Douglas Stebila
3286 (Sun Microsystems Laboratories)]
3288 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
3289 through methods ('mul', 'precompute_mult').
3291 The generic implementations (now internally called 'ec_wNAF_mul'
3292 and 'ec_wNAF_precomputed_mult') remain the default if these
3293 methods are undefined.
3295 [Sheueling Chang Shantz and Douglas Stebila
3296 (Sun Microsystems Laboratories)]
3298 *) New function EC_GROUP_get_degree, which is defined through
3299 EC_METHOD. For curves over prime fields, this returns the bit
3300 length of the modulus.
3302 [Sheueling Chang Shantz and Douglas Stebila
3303 (Sun Microsystems Laboratories)]
3305 *) New functions EC_GROUP_dup, EC_POINT_dup.
3306 (These simply call ..._new and ..._copy).
3308 [Sheueling Chang Shantz and Douglas Stebila
3309 (Sun Microsystems Laboratories)]
3311 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3312 Polynomials are represented as BIGNUMs (where the sign bit is not
3313 used) in the following functions [macros]:
3316 BN_GF2m_sub [= BN_GF2m_add]
3317 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3318 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3319 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3321 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3322 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3323 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3324 BN_GF2m_cmp [= BN_ucmp]
3326 (Note that only the 'mod' functions are actually for fields GF(2^m).
3327 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3329 For some functions, an the irreducible polynomial defining a
3330 field can be given as an 'unsigned int[]' with strictly
3331 decreasing elements giving the indices of those bits that are set;
3332 i.e., p[] represents the polynomial
3333 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3335 p[0] > p[1] > ... > p[k] = 0.
3336 This applies to the following functions:
3341 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3342 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3344 BN_GF2m_mod_sqrt_arr
3345 BN_GF2m_mod_solve_quad_arr
3349 Conversion can be performed by the following functions:
3354 bntest.c has additional tests for binary polynomial arithmetic.
3356 Two implementations for BN_GF2m_mod_div() are available.
3357 The default algorithm simply uses BN_GF2m_mod_inv() and
3358 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3359 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3360 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3362 [Sheueling Chang Shantz and Douglas Stebila
3363 (Sun Microsystems Laboratories)]
3365 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3366 functionality is disabled at compile-time.
3367 [Douglas Stebila <douglas.stebila@sun.com>]
3369 *) Change default behaviour of 'openssl asn1parse' so that more
3370 information is visible when viewing, e.g., a certificate:
3372 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3373 mode the content of non-printable OCTET STRINGs is output in a
3374 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3375 avoid the appearance of a printable string.
3376 [Nils Larsch <nla@trustcenter.de>]
3378 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3380 EC_GROUP_set_asn1_flag()
3381 EC_GROUP_get_asn1_flag()
3382 EC_GROUP_set_point_conversion_form()
3383 EC_GROUP_get_point_conversion_form()
3384 These control ASN1 encoding details:
3385 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3386 has been set to OPENSSL_EC_NAMED_CURVE.
3387 - Points are encoded in uncompressed form by default; options for
3388 asn1_for are as for point2oct, namely
3389 POINT_CONVERSION_COMPRESSED
3390 POINT_CONVERSION_UNCOMPRESSED
3391 POINT_CONVERSION_HYBRID
3393 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3396 EC_GROUP_get0_seed()
3397 EC_GROUP_get_seed_len()
3398 This is used only for ASN1 purposes (so far).
3399 [Nils Larsch <nla@trustcenter.de>]
3401 *) Add 'field_type' member to EC_METHOD, which holds the NID
3402 of the appropriate field type OID. The new function
3403 EC_METHOD_get_field_type() returns this value.
3404 [Nils Larsch <nla@trustcenter.de>]
3409 EC_POINT_point2hex()
3410 EC_POINT_hex2point()
3411 providing useful interfaces to EC_POINT_point2oct() and
3412 EC_POINT_oct2point().
3413 [Nils Larsch <nla@trustcenter.de>]
3415 *) Change internals of the EC library so that the functions
3416 EC_GROUP_set_generator()
3417 EC_GROUP_get_generator()
3418 EC_GROUP_get_order()
3419 EC_GROUP_get_cofactor()
3420 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3421 to methods, which would lead to unnecessary code duplication when
3422 adding different types of curves.
3423 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3425 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3426 arithmetic, and such that modified wNAFs are generated
3427 (which avoid length expansion in many cases).
3430 *) Add a function EC_GROUP_check_discriminant() (defined via
3431 EC_METHOD) that verifies that the curve discriminant is non-zero.
3433 Add a function EC_GROUP_check() that makes some sanity tests
3434 on a EC_GROUP, its generator and order. This includes
3435 EC_GROUP_check_discriminant().
3436 [Nils Larsch <nla@trustcenter.de>]
3438 *) Add ECDSA in new directory crypto/ecdsa/.
3440 Add applications 'openssl ecparam' and 'openssl ecdsa'
3441 (these are based on 'openssl dsaparam' and 'openssl dsa').
3443 ECDSA support is also included in various other files across the
3444 library. Most notably,
3445 - 'openssl req' now has a '-newkey ecdsa:file' option;
3446 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3447 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3448 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3449 them suitable for ECDSA where domain parameters must be
3450 extracted before the specific public key;
3451 - ECDSA engine support has been added.
3452 [Nils Larsch <nla@trustcenter.de>]
3454 *) Include some named elliptic curves, and add OIDs from X9.62,
3455 SECG, and WAP/WTLS. Each curve can be obtained from the new
3457 EC_GROUP_new_by_curve_name(),
3458 and the list of available named curves can be obtained with
3459 EC_get_builtin_curves().
3460 Also add a 'curve_name' member to EC_GROUP objects, which can be
3462 EC_GROUP_set_curve_name()
3463 EC_GROUP_get_curve_name()
3464 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3466 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3467 was actually never needed) and in BN_mul(). The removal in BN_mul()
3468 required a small change in bn_mul_part_recursive() and the addition
3469 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3470 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3471 bn_sub_words() and bn_add_words() except they take arrays with
3475 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3477 *) Cleanse PEM buffers before freeing them since they may contain
3479 [Benjamin Bennett <ben@psc.edu>]
3481 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3482 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3483 authentication-only ciphersuites.
3486 *) Since AES128 and AES256 share a single mask bit in the logic of
3487 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3488 kludge to work properly if AES128 is available and AES256 isn't.
3491 *) Expand security boundary to match 1.1.1 module.
3494 *) Remove redundant features: hash file source, editing of test vectors
3495 modify fipsld to use external fips_premain.c signature.
3498 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3499 run algorithm test programs.
3502 *) Make algorithm test programs more tolerant of whitespace.
3505 *) Have SSL/TLS server implementation tolerate "mismatched" record
3506 protocol version while receiving ClientHello even if the
3507 ClientHello is fragmented. (The server can't insist on the
3508 particular protocol version it has chosen before the ServerHello
3509 message has informed the client about his choice.)
3512 *) Load error codes if they are not already present instead of using a
3513 static variable. This allows them to be cleanly unloaded and reloaded.
3516 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3518 *) Introduce limits to prevent malicious keys being able to
3519 cause a denial of service. (CVE-2006-2940)
3520 [Steve Henson, Bodo Moeller]
3522 *) Fix ASN.1 parsing of certain invalid structures that can result
3523 in a denial of service. (CVE-2006-2937) [Steve Henson]
3525 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3526 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3528 *) Fix SSL client code which could crash if connecting to a
3529 malicious SSLv2 server. (CVE-2006-4343)
3530 [Tavis Ormandy and Will Drewry, Google Security Team]
3532 *) Change ciphersuite string processing so that an explicit
3533 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3534 will no longer include "AES128-SHA"), and any other similar
3535 ciphersuite (same bitmap) from *other* protocol versions (so that
3536 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3537 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3538 changes from 0.9.8b and 0.9.8d.
3541 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3543 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3544 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3546 *) Change the Unix randomness entropy gathering to use poll() when
3547 possible instead of select(), since the latter has some
3548 undesirable limitations.
3549 [Darryl Miles via Richard Levitte and Bodo Moeller]
3551 *) Disable rogue ciphersuites:
3553 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3554 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3555 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3557 The latter two were purportedly from
3558 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3561 Also deactive the remaining ciphersuites from
3562 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3563 unofficial, and the ID has long expired.
3566 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3567 dual-core machines) and other potential thread-safety issues.
3570 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3572 *) Adapt fipsld and the build system to link against the validated FIPS
3573 module in FIPS mode.
3576 *) Fixes for VC++ 2005 build under Windows.
3579 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3580 from a Windows bash shell such as MSYS. It is autodetected from the
3581 "config" script when run from a VC++ environment. Modify standard VC++
3582 build to use fipscanister.o from the GNU make build.
3585 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3587 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3588 The value now differs depending on if you build for FIPS or not.
3589 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3590 safely run with a non-FIPSed libcrypto, as it may crash because of
3591 the difference induced by this change.
3594 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3596 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3597 (part of SSL_OP_ALL). This option used to disable the
3598 countermeasure against man-in-the-middle protocol-version
3599 rollback in the SSL 2.0 server implementation, which is a bad
3600 idea. (CVE-2005-2969)
3602 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3603 for Information Security, National Institute of Advanced Industrial
3604 Science and Technology [AIST], Japan)]
3606 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3607 mainly for FIPS compliance and not fully integrated at this stage.
3610 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3611 the exponentiation using a fixed-length exponent. (Otherwise,
3612 the information leaked through timing could expose the secret key
3613 after many signatures; cf. Bleichenbacher's attack on DSA with
3617 *) Make a new fixed-window mod_exp implementation the default for
3618 RSA, DSA, and DH private-key operations so that the sequence of
3619 squares and multiplies and the memory access pattern are
3620 independent of the particular secret key. This will mitigate
3621 cache-timing and potential related attacks.
3623 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3624 and this is automatically used by BN_mod_exp_mont() if the new flag
3625 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3626 will use this BN flag for private exponents unless the flag
3627 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3628 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3630 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3632 *) Change the client implementation for SSLv23_method() and
3633 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3634 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3635 (Previously, the SSL 2.0 backwards compatible Client Hello
3636 message format would be used even with SSL_OP_NO_SSLv2.)
3639 *) Add support for smime-type MIME parameter in S/MIME messages which some
3643 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3644 a threadsafe manner. Modify rsa code to use new function and add calls
3645 to dsa and dh code (which had race conditions before).
3648 *) Include the fixed error library code in the C error file definitions
3649 instead of fixing them up at runtime. This keeps the error code
3650 structures constant.
3653 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3655 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3658 *) Fixes for newer kerberos headers. NB: the casts are needed because
3659 the 'length' field is signed on one version and unsigned on another
3660 with no (?) obvious way to tell the difference, without these VC++
3661 complains. Also the "definition" of FAR (blank) is no longer included
3662 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3663 some needed definitions.
3666 *) Undo Cygwin change.
3669 *) Added support for proxy certificates according to RFC 3820.
3670 Because they may be a security thread to unaware applications,
3671 they must be explicitely allowed in run-time. See
3672 docs/HOWTO/proxy_certificates.txt for further information.
3675 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3677 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3678 server and client random values. Previously
3679 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3680 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3682 This change has negligible security impact because:
3684 1. Server and client random values still have 24 bytes of pseudo random
3687 2. Server and client random values are sent in the clear in the initial
3690 3. The master secret is derived using the premaster secret (48 bytes in
3691 size for static RSA ciphersuites) as well as client server and random
3694 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3697 [Stephen Henson, reported by UK NISCC]
3699 *) Use Windows randomness collection on Cygwin.
3702 *) Fix hang in EGD/PRNGD query when communication socket is closed
3703 prematurely by EGD/PRNGD.
3704 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3706 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3709 *) Back-port of selected performance improvements from development
3710 branch, as well as improved support for PowerPC platforms.
3713 *) Add lots of checks for memory allocation failure, error codes to indicate
3714 failure and freeing up memory if a failure occurs.
3715 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3717 *) Add new -passin argument to dgst.
3720 *) Perform some character comparisons of different types in X509_NAME_cmp:
3721 this is needed for some certificates that reencode DNs into UTF8Strings
3722 (in violation of RFC3280) and can't or wont issue name rollover
3726 *) Make an explicit check during certificate validation to see that
3727 the CA setting in each certificate on the chain is correct. As a
3728 side effect always do the following basic checks on extensions,
3729 not just when there's an associated purpose to the check:
3731 - if there is an unhandled critical extension (unless the user
3732 has chosen to ignore this fault)
3733 - if the path length has been exceeded (if one is set at all)
3734 - that certain extensions fit the associated purpose (if one has
3738 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3740 *) Avoid a race condition when CRLs are checked in a multi threaded
3741 environment. This would happen due to the reordering of the revoked
3742 entries during signature checking and serial number lookup. Now the
3743 encoding is cached and the serial number sort performed under a lock.
3744 Add new STACK function sk_is_sorted().
3747 *) Add Delta CRL to the extension code.
3750 *) Various fixes to s3_pkt.c so alerts are sent properly.
3751 [David Holmes <d.holmes@f5.com>]
3753 *) Reduce the chances of duplicate issuer name and serial numbers (in
3754 violation of RFC3280) using the OpenSSL certificate creation utilities.
3755 This is done by creating a random 64 bit value for the initial serial
3756 number when a serial number file is created or when a self signed
3757 certificate is created using 'openssl req -x509'. The initial serial
3758 number file is created using 'openssl x509 -next_serial' in CA.pl
3759 rather than being initialized to 1.
3762 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3764 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3765 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3766 [Joe Orton, Steve Henson]
3768 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3770 [Joe Orton, Steve Henson]
3772 *) Make it possible to have multiple active certificates with the same
3773 subject in the CA index file. This is done only if the keyword
3774 'unique_subject' is set to 'no' in the main CA section (default
3775 if 'CA_default') of the configuration file. The value is saved
3776 with the database itself in a separate index attribute file,
3777 named like the index file with '.attr' appended to the name.
3780 *) X509 verify fixes. Disable broken certificate workarounds when
3781 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3782 keyUsage extension present. Don't accept CRLs with unhandled critical
3783 extensions: since verify currently doesn't process CRL extensions this
3784 rejects a CRL with *any* critical extensions. Add new verify error codes
3788 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3789 A clarification of RFC2560 will require the use of OCTET STRINGs and
3790 some implementations cannot handle the current raw format. Since OpenSSL
3791 copies and compares OCSP nonces as opaque blobs without any attempt at
3792 parsing them this should not create any compatibility issues.
3795 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3796 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3797 this HMAC (and other) operations are several times slower than OpenSSL
3801 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3802 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3804 *) Use the correct content when signing type "other".
3807 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3809 *) Fix various bugs revealed by running the NISCC test suite:
3811 Stop out of bounds reads in the ASN1 code when presented with
3812 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3814 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3816 If verify callback ignores invalid public key errors don't try to check
3817 certificate signature with the NULL public key.
3821 *) New -ignore_err option in ocsp application to stop the server
3822 exiting on the first error in a request.
3825 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3826 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3830 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3831 extra data after the compression methods not only for TLS 1.0
3832 but also for SSL 3.0 (as required by the specification).
3833 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3835 *) Change X509_certificate_type() to mark the key as exported/exportable
3836 when it's 512 *bits* long, not 512 bytes.
3839 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3840 blocks during encryption.
3843 *) Various fixes to base64 BIO and non blocking I/O. On write
3844 flushes were not handled properly if the BIO retried. On read
3845 data was not being buffered properly and had various logic bugs.
3846 This also affects blocking I/O when the data being decoded is a
3850 *) Various S/MIME bugfixes and compatibility changes:
3851 output correct application/pkcs7 MIME type if
3852 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3853 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3854 of files as .eml work). Correctly handle very long lines in MIME
3858 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3860 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3861 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3862 a protocol version number mismatch like a decryption error
3863 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3866 *) Turn on RSA blinding by default in the default implementation
3867 to avoid a timing attack. Applications that don't want it can call
3868 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3869 They would be ill-advised to do so in most cases.
3870 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3872 *) Change RSA blinding code so that it works when the PRNG is not
3873 seeded (in this case, the secret RSA exponent is abused as
3874 an unpredictable seed -- if it is not unpredictable, there
3875 is no point in blinding anyway). Make RSA blinding thread-safe
3876 by remembering the creator's thread ID in rsa->blinding and
3877 having all other threads use local one-time blinding factors
3878 (this requires more computation than sharing rsa->blinding, but
3879 avoids excessive locking; and if an RSA object is not shared
3880 between threads, blinding will still be very fast).
3883 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
3884 ENGINE as defaults for all supported algorithms irrespective of
3885 the 'flags' parameter. 'flags' is now honoured, so applications
3886 should make sure they are passing it correctly.
3889 *) Target "mingw" now allows native Windows code to be generated in
3890 the Cygwin environment as well as with the MinGW compiler.
3893 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
3895 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3896 via timing by performing a MAC computation even if incorrrect
3897 block cipher padding has been found. This is a countermeasure
3898 against active attacks where the attacker has to distinguish
3899 between bad padding and a MAC verification error. (CVE-2003-0078)
3901 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3902 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3903 Martin Vuagnoux (EPFL, Ilion)]
3905 *) Make the no-err option work as intended. The intention with no-err
3906 is not to have the whole error stack handling routines removed from
3907 libcrypto, it's only intended to remove all the function name and
3908 reason texts, thereby removing some of the footprint that may not
3909 be interesting if those errors aren't displayed anyway.
3911 NOTE: it's still possible for any application or module to have it's
3912 own set of error texts inserted. The routines are there, just not
3913 used by default when no-err is given.
3916 *) Add support for FreeBSD on IA64.
3917 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
3919 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
3920 Kerberos function mit_des_cbc_cksum(). Before this change,
3921 the value returned by DES_cbc_cksum() was like the one from
3922 mit_des_cbc_cksum(), except the bytes were swapped.
3923 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
3925 *) Allow an application to disable the automatic SSL chain building.
3926 Before this a rather primitive chain build was always performed in
3927 ssl3_output_cert_chain(): an application had no way to send the
3928 correct chain if the automatic operation produced an incorrect result.
3930 Now the chain builder is disabled if either:
3932 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
3934 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
3936 The reasoning behind this is that an application would not want the
3937 auto chain building to take place if extra chain certificates are
3938 present and it might also want a means of sending no additional
3939 certificates (for example the chain has two certificates and the
3943 *) Add the possibility to build without the ENGINE framework.
3944 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3946 *) Under Win32 gmtime() can return NULL: check return value in
3947 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
3950 *) DSA routines: under certain error conditions uninitialized BN objects
3951 could be freed. Solution: make sure initialization is performed early
3952 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
3953 Nils Larsch <nla@trustcenter.de> via PR#459)
3956 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
3957 checked on reconnect on the client side, therefore session resumption
3958 could still fail with a "ssl session id is different" error. This
3959 behaviour is masked when SSL_OP_ALL is used due to
3960 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
3961 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3962 followup to PR #377.
3965 *) IA-32 assembler support enhancements: unified ELF targets, support
3966 for SCO/Caldera platforms, fix for Cygwin shared build.
3969 *) Add support for FreeBSD on sparc64. As a consequence, support for
3970 FreeBSD on non-x86 processors is separate from x86 processors on
3971 the config script, much like the NetBSD support.
3972 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
3974 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
3976 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
3979 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
3980 code (06) was taken as the first octet of the session ID and the last
3981 octet was ignored consequently. As a result SSLv2 client side session
3982 caching could not have worked due to the session ID mismatch between
3984 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3988 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
3989 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
3993 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
3994 seems that in spite of existing for more than a year, many application
3995 author have done nothing to provide the necessary callbacks, which
3996 means that this particular engine will not work properly anywhere.
3997 This is a very unfortunate situation which forces us, in the name
3998 of usability, to give the hw_ncipher.c a static lock, which is part
4000 NOTE: This is for the 0.9.7 series ONLY. This hack will never
4001 appear in 0.9.8 or later. We EXPECT application authors to have
4002 dealt properly with this when 0.9.8 is released (unless we actually
4003 make such changes in the libcrypto locking code that changes will
4004 have to be made anyway).
4007 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
4008 octets have been read, EOF or an error occurs. Without this change
4009 some truncated ASN1 structures will not produce an error.
4012 *) Disable Heimdal support, since it hasn't been fully implemented.
4013 Still give the possibility to force the use of Heimdal, but with
4014 warnings and a request that patches get sent to openssl-dev.
4017 *) Add the VC-CE target, introduce the WINCE sysname, and add
4018 INSTALL.WCE and appropriate conditionals to make it build.
4019 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4021 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4022 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4023 edit numbers of the version.
4024 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4026 *) Introduce safe string copy and catenation functions
4027 (BUF_strlcpy() and BUF_strlcat()).
4028 [Ben Laurie (CHATS) and Richard Levitte]
4030 *) Avoid using fixed-size buffers for one-line DNs.
4031 [Ben Laurie (CHATS)]
4033 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4034 resizing buffers containing secrets, and use where appropriate.
4035 [Ben Laurie (CHATS)]
4037 *) Avoid using fixed size buffers for configuration file location.
4038 [Ben Laurie (CHATS)]
4040 *) Avoid filename truncation for various CA files.
4041 [Ben Laurie (CHATS)]
4043 *) Use sizeof in preference to magic numbers.
4044 [Ben Laurie (CHATS)]
4046 *) Avoid filename truncation in cert requests.
4047 [Ben Laurie (CHATS)]
4049 *) Add assertions to check for (supposedly impossible) buffer
4051 [Ben Laurie (CHATS)]
4053 *) Don't cache truncated DNS entries in the local cache (this could
4054 potentially lead to a spoofing attack).
4055 [Ben Laurie (CHATS)]
4057 *) Fix various buffers to be large enough for hex/decimal
4058 representations in a platform independent manner.
4059 [Ben Laurie (CHATS)]
4061 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4062 resizing buffers containing secrets, and use where appropriate.
4063 [Ben Laurie (CHATS)]
4065 *) Add BIO_indent() to avoid much slightly worrying code to do
4067 [Ben Laurie (CHATS)]
4069 *) Convert sprintf()/BIO_puts() to BIO_printf().
4070 [Ben Laurie (CHATS)]
4072 *) buffer_gets() could terminate with the buffer only half
4074 [Ben Laurie (CHATS)]
4076 *) Add assertions to prevent user-supplied crypto functions from
4077 overflowing internal buffers by having large block sizes, etc.
4078 [Ben Laurie (CHATS)]
4080 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4082 [Ben Laurie (CHATS)]
4084 *) Eliminate unused copy of key in RC4.
4085 [Ben Laurie (CHATS)]
4087 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4088 [Ben Laurie (CHATS)]
4090 *) Fix off-by-one error in EGD path.
4091 [Ben Laurie (CHATS)]
4093 *) If RANDFILE path is too long, ignore instead of truncating.
4094 [Ben Laurie (CHATS)]
4096 *) Eliminate unused and incorrectly sized X.509 structure
4098 [Ben Laurie (CHATS)]
4100 *) Eliminate unused and dangerous function knumber().
4101 [Ben Laurie (CHATS)]
4103 *) Eliminate unused and dangerous structure, KSSL_ERR.
4104 [Ben Laurie (CHATS)]
4106 *) Protect against overlong session ID context length in an encoded
4107 session object. Since these are local, this does not appear to be
4109 [Ben Laurie (CHATS)]
4111 *) Change from security patch (see 0.9.6e below) that did not affect
4112 the 0.9.6 release series:
4114 Remote buffer overflow in SSL3 protocol - an attacker could
4115 supply an oversized master key in Kerberos-enabled versions.
4117 [Ben Laurie (CHATS)]
4119 *) Change the SSL kerb5 codes to match RFC 2712.
4122 *) Make -nameopt work fully for req and add -reqopt switch.
4123 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4125 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4126 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4128 *) Make sure tests can be performed even if the corresponding algorithms
4129 have been removed entirely. This was also the last step to make
4130 OpenSSL compilable with DJGPP under all reasonable conditions.
4131 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
4133 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
4134 to allow version independent disabling of normally unselected ciphers,
4135 which may be activated as a side-effect of selecting a single cipher.
4137 (E.g., cipher list string "RSA" enables ciphersuites that are left
4138 out of "ALL" because they do not provide symmetric encryption.
4139 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
4140 [Lutz Jaenicke, Bodo Moeller]
4142 *) Add appropriate support for separate platform-dependent build
4143 directories. The recommended way to make a platform-dependent
4144 build directory is the following (tested on Linux), maybe with
4147 # Place yourself outside of the OpenSSL source tree. In
4148 # this example, the environment variable OPENSSL_SOURCE
4149 # is assumed to contain the absolute OpenSSL source directory.
4150 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
4151 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
4152 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
4153 mkdir -p `dirname $F`
4154 ln -s $OPENSSL_SOURCE/$F $F
4157 To be absolutely sure not to disturb the source tree, a "make clean"
4158 is a good thing. If it isn't successfull, don't worry about it,
4159 it probably means the source directory is very clean.
4162 *) Make sure any ENGINE control commands make local copies of string
4163 pointers passed to them whenever necessary. Otherwise it is possible
4164 the caller may have overwritten (or deallocated) the original string
4165 data when a later ENGINE operation tries to use the stored values.
4166 [Götz Babin-Ebell <babinebell@trustcenter.de>]
4168 *) Improve diagnostics in file reading and command-line digests.
4169 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
4171 *) Add AES modes CFB and OFB to the object database. Correct an
4172 error in AES-CFB decryption.
4175 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
4176 allows existing EVP_CIPHER_CTX structures to be reused after
4177 calling EVP_*Final(). This behaviour is used by encryption
4178 BIOs and some applications. This has the side effect that
4179 applications must explicitly clean up cipher contexts with
4180 EVP_CIPHER_CTX_cleanup() or they will leak memory.
4183 *) Check the values of dna and dnb in bn_mul_recursive before calling
4184 bn_mul_comba (a non zero value means the a or b arrays do not contain
4185 n2 elements) and fallback to bn_mul_normal if either is not zero.
4188 *) Fix escaping of non-ASCII characters when using the -subj option
4189 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
4192 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
4193 form for "surname", serialNumber has no short form.
4194 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
4195 therefore remove "mail" short name for "internet 7".
4196 The OID for unique identifiers in X509 certificates is
4197 x500UniqueIdentifier, not uniqueIdentifier.
4198 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
4201 *) Add an "init" command to the ENGINE config module and auto initialize
4202 ENGINEs. Without any "init" command the ENGINE will be initialized
4203 after all ctrl commands have been executed on it. If init=1 the
4204 ENGINE is initailized at that point (ctrls before that point are run
4205 on the uninitialized ENGINE and after on the initialized one). If
4206 init=0 then the ENGINE will not be iniatialized at all.
4209 *) Fix the 'app_verify_callback' interface so that the user-defined
4210 argument is actually passed to the callback: In the
4211 SSL_CTX_set_cert_verify_callback() prototype, the callback
4212 declaration has been changed from
4215 int (*cb)(X509_STORE_CTX *,void *);
4216 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
4217 i=s->ctx->app_verify_callback(&ctx)
4218 has been changed into
4219 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
4221 To update applications using SSL_CTX_set_cert_verify_callback(),
4222 a dummy argument can be added to their callback functions.
4223 [D. K. Smetters <smetters@parc.xerox.com>]
4225 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
4226 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
4228 *) Add and OPENSSL_LOAD_CONF define which will cause
4229 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
4230 This allows older applications to transparently support certain
4231 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
4232 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
4233 load the config file and OPENSSL_add_all_algorithms_conf() which will
4234 always load it have also been added.
4237 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
4238 Adjust NIDs and EVP layer.
4239 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4241 *) Config modules support in openssl utility.
4243 Most commands now load modules from the config file,
4244 though in a few (such as version) this isn't done
4245 because it couldn't be used for anything.
4247 In the case of ca and req the config file used is
4248 the same as the utility itself: that is the -config
4249 command line option can be used to specify an
4253 *) Move default behaviour from OPENSSL_config(). If appname is NULL
4254 use "openssl_conf" if filename is NULL use default openssl config file.
4257 *) Add an argument to OPENSSL_config() to allow the use of an alternative
4258 config section name. Add a new flag to tolerate a missing config file
4259 and move code to CONF_modules_load_file().
4262 *) Support for crypto accelerator cards from Accelerated Encryption
4263 Processing, www.aep.ie. (Use engine 'aep')
4264 The support was copied from 0.9.6c [engine] and adapted/corrected
4265 to work with the new engine framework.
4266 [AEP Inc. and Richard Levitte]
4268 *) Support for SureWare crypto accelerator cards from Baltimore
4269 Technologies. (Use engine 'sureware')
4270 The support was copied from 0.9.6c [engine] and adapted
4271 to work with the new engine framework.
4274 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
4275 make the newer ENGINE framework commands for the CHIL engine work.
4276 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
4278 *) Make it possible to produce shared libraries on ReliantUNIX.
4279 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
4281 *) Add the configuration target debug-linux-ppro.
4282 Make 'openssl rsa' use the general key loading routines
4283 implemented in apps.c, and make those routines able to
4284 handle the key format FORMAT_NETSCAPE and the variant
4286 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4288 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4289 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4291 *) Add -keyform to rsautl, and document -engine.
4292 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
4294 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
4295 BIO_R_NO_SUCH_FILE error code rather than the generic
4296 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
4299 *) Add new functions
4301 ERR_peek_last_error_line
4302 ERR_peek_last_error_line_data.
4303 These are similar to
4306 ERR_peek_error_line_data,
4307 but report on the latest error recorded rather than the first one
4308 still in the error queue.
4309 [Ben Laurie, Bodo Moeller]
4311 *) default_algorithms option in ENGINE config module. This allows things
4313 default_algorithms = ALL
4314 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4317 *) Prelminary ENGINE config module.
4320 *) New experimental application configuration code.
4323 *) Change the AES code to follow the same name structure as all other
4324 symmetric ciphers, and behave the same way. Move everything to
4325 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4326 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4328 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4329 [Ben Laurie and Theo de Raadt]
4331 *) Add option to output public keys in req command.
4332 [Massimiliano Pala madwolf@openca.org]
4334 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4335 (up to about 10% better than before for P-192 and P-224).
4338 *) New functions/macros
4340 SSL_CTX_set_msg_callback(ctx, cb)
4341 SSL_CTX_set_msg_callback_arg(ctx, arg)
4342 SSL_set_msg_callback(ssl, cb)
4343 SSL_set_msg_callback_arg(ssl, arg)
4345 to request calling a callback function
4347 void cb(int write_p, int version, int content_type,
4348 const void *buf, size_t len, SSL *ssl, void *arg)
4350 whenever a protocol message has been completely received
4351 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4352 protocol version according to which the SSL library interprets
4353 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4354 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4355 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4356 specification (change_cipher_spec(20), alert(21), handshake(22)).
4357 'buf' and 'len' point to the actual message, 'ssl' to the
4358 SSL object, and 'arg' is the application-defined value set by
4359 SSL[_CTX]_set_msg_callback_arg().
4361 'openssl s_client' and 'openssl s_server' have new '-msg' options
4362 to enable a callback that displays all protocol messages.
4365 *) Change the shared library support so shared libraries are built as
4366 soon as the corresponding static library is finished, and thereby get
4367 openssl and the test programs linked against the shared library.
4368 This still only happens when the keyword "shard" has been given to
4369 the configuration scripts.
4371 NOTE: shared library support is still an experimental thing, and
4372 backward binary compatibility is still not guaranteed.
4373 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4375 *) Add support for Subject Information Access extension.
4376 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4378 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4379 additional bytes when new memory had to be allocated, not just
4380 when reusing an existing buffer.
4383 *) New command line and configuration option 'utf8' for the req command.
4384 This allows field values to be specified as UTF8 strings.
4387 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4388 runs for the former and machine-readable output for the latter.
4391 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4392 of the e-mail address in the DN (i.e., it will go into a certificate
4393 extension only). The new configuration file option 'email_in_dn = no'
4394 has the same effect.
4395 [Massimiliano Pala madwolf@openca.org]
4397 *) Change all functions with names starting with des_ to be starting
4398 with DES_ instead. Add wrappers that are compatible with libdes,
4399 but are named _ossl_old_des_*. Finally, add macros that map the
4400 des_* symbols to the corresponding _ossl_old_des_* if libdes
4401 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4402 desired, the des_* symbols will be mapped to DES_*, with one
4405 Since we provide two compatibility mappings, the user needs to
4406 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4407 compatibility is desired. The default (i.e., when that macro
4408 isn't defined) is OpenSSL 0.9.6c compatibility.
4410 There are also macros that enable and disable the support of old
4411 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4412 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4413 are defined, the default will apply: to support the old des routines.
4415 In either case, one must include openssl/des.h to get the correct
4416 definitions. Do not try to just include openssl/des_old.h, that
4419 NOTE: This is a major break of an old API into a new one. Software
4420 authors are encouraged to switch to the DES_ style functions. Some
4421 time in the future, des_old.h and the libdes compatibility functions
4422 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4423 default), and then completely removed.
4426 *) Test for certificates which contain unsupported critical extensions.
4427 If such a certificate is found during a verify operation it is
4428 rejected by default: this behaviour can be overridden by either
4429 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4430 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4431 X509_supported_extension() has also been added which returns 1 if a
4432 particular extension is supported.
4435 *) Modify the behaviour of EVP cipher functions in similar way to digests
4436 to retain compatibility with existing code.
4439 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4440 compatibility with existing code. In particular the 'ctx' parameter does
4441 not have to be to be initialized before the call to EVP_DigestInit() and
4442 it is tidied up after a call to EVP_DigestFinal(). New function
4443 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4444 EVP_MD_CTX_copy() changed to not require the destination to be
4445 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4446 requires the destination to be valid.
4448 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4449 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4452 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4453 so that complete 'Handshake' protocol structures are kept in memory
4454 instead of overwriting 'msg_type' and 'length' with 'body' data.
4457 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4458 [Massimo Santin via Richard Levitte]
4460 *) Major restructuring to the underlying ENGINE code. This includes
4461 reduction of linker bloat, separation of pure "ENGINE" manipulation
4462 (initialisation, etc) from functionality dealing with implementations
4463 of specific crypto iterfaces. This change also introduces integrated
4464 support for symmetric ciphers and digest implementations - so ENGINEs
4465 can now accelerate these by providing EVP_CIPHER and EVP_MD
4466 implementations of their own. This is detailed in crypto/engine/README
4467 as it couldn't be adequately described here. However, there are a few
4468 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4469 were changed in the original introduction of ENGINE code have now
4470 reverted back - the hooking from this code to ENGINE is now a good
4471 deal more passive and at run-time, operations deal directly with
4472 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4473 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4474 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4475 they were not being used by the framework as there is no concept of a
4476 BIGNUM_METHOD and they could not be generalised to the new
4477 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4478 ENGINE_cpy() has been removed as it cannot be consistently defined in
4482 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4485 *) Change mkdef.pl to sort symbols that get the same entry number,
4486 and make sure the automatically generated functions ERR_load_*
4487 become part of libeay.num as well.
4490 *) New function SSL_renegotiate_pending(). This returns true once
4491 renegotiation has been requested (either SSL_renegotiate() call
4492 or HelloRequest/ClientHello receveived from the peer) and becomes
4493 false once a handshake has been completed.
4494 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4495 sends a HelloRequest, but does not ensure that a handshake takes
4496 place. SSL_renegotiate_pending() is useful for checking if the
4497 client has followed the request.)
4500 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4501 By default, clients may request session resumption even during
4502 renegotiation (if session ID contexts permit); with this option,
4503 session resumption is possible only in the first handshake.
4505 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4506 more bits available for options that should not be part of
4507 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4510 *) Add some demos for certificate and certificate request creation.
4513 *) Make maximum certificate chain size accepted from the peer application
4514 settable (SSL*_get/set_max_cert_list()), as proposed by
4515 "Douglas E. Engert" <deengert@anl.gov>.
4518 *) Add support for shared libraries for Unixware-7
4519 (Boyd Lynn Gerber <gerberb@zenez.com>).
4522 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4523 be done prior to destruction. Use this to unload error strings from
4524 ENGINEs that load their own error strings. NB: This adds two new API
4525 functions to "get" and "set" this destroy handler in an ENGINE.
4528 *) Alter all existing ENGINE implementations (except "openssl" and
4529 "openbsd") to dynamically instantiate their own error strings. This
4530 makes them more flexible to be built both as statically-linked ENGINEs
4531 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4532 Also, add stub code to each that makes building them as self-contained
4533 shared-libraries easier (see README.ENGINE).
4536 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4537 implementations into applications that are completely implemented in
4538 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4539 commands that can be used to configure what shared-library to load and
4540 to control aspects of the way it is handled. Also, made an update to
4541 the README.ENGINE file that brings its information up-to-date and
4542 provides some information and instructions on the "dynamic" ENGINE
4543 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4546 *) Make it possible to unload ranges of ERR strings with a new
4547 "ERR_unload_strings" function.
4550 *) Add a copy() function to EVP_MD.
4553 *) Make EVP_MD routines take a context pointer instead of just the
4554 md_data void pointer.
4557 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4558 that the digest can only process a single chunk of data
4559 (typically because it is provided by a piece of
4560 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4561 is only going to provide a single chunk of data, and hence the
4562 framework needn't accumulate the data for oneshot drivers.
4565 *) As with "ERR", make it possible to replace the underlying "ex_data"
4566 functions. This change also alters the storage and management of global
4567 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4568 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4569 index counters. The API functions that use this state have been changed
4570 to take a "class_index" rather than pointers to the class's local STACK
4571 and counter, and there is now an API function to dynamically create new
4572 classes. This centralisation allows us to (a) plug a lot of the
4573 thread-safety problems that existed, and (b) makes it possible to clean
4574 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4575 such data would previously have always leaked in application code and
4576 workarounds were in place to make the memory debugging turn a blind eye
4577 to it. Application code that doesn't use this new function will still
4578 leak as before, but their memory debugging output will announce it now
4579 rather than letting it slide.
4581 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4582 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4583 has a return value to indicate success or failure.
4586 *) Make it possible to replace the underlying "ERR" functions such that the
4587 global state (2 LHASH tables and 2 locks) is only used by the "default"
4588 implementation. This change also adds two functions to "get" and "set"
4589 the implementation prior to it being automatically set the first time
4590 any other ERR function takes place. Ie. an application can call "get",
4591 pass the return value to a module it has just loaded, and that module
4592 can call its own "set" function using that value. This means the
4593 module's "ERR" operations will use (and modify) the error state in the
4594 application and not in its own statically linked copy of OpenSSL code.
4597 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4598 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4599 the operation, and provides a more encapsulated way for external code
4600 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4601 to use these functions rather than manually incrementing the counts.
4603 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4606 *) Add EVP test program.
4609 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4612 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4613 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4614 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4615 These allow a CRL to be built without having to access X509_CRL fields
4616 directly. Modify 'ca' application to use new functions.
4619 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4620 bug workarounds. Rollback attack detection is a security feature.
4621 The problem will only arise on OpenSSL servers when TLSv1 is not
4622 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4623 Software authors not wanting to support TLSv1 will have special reasons
4624 for their choice and can explicitly enable this option.
4625 [Bodo Moeller, Lutz Jaenicke]
4627 *) Rationalise EVP so it can be extended: don't include a union of
4628 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4629 (similar to those existing for EVP_CIPHER_CTX).
4634 EVP_MD_CTX_init(&md); /* new function call */
4635 EVP_DigestInit(&md, EVP_sha1());
4636 EVP_DigestUpdate(&md, in, len);
4637 EVP_DigestFinal(&md, out, NULL);
4638 EVP_MD_CTX_cleanup(&md); /* new function call */
4642 *) Make DES key schedule conform to the usual scheme, as well as
4643 correcting its structure. This means that calls to DES functions
4644 now have to pass a pointer to a des_key_schedule instead of a
4645 plain des_key_schedule (which was actually always a pointer
4648 des_key_schedule ks;
4650 des_set_key_checked(..., &ks);
4651 des_ncbc_encrypt(..., &ks, ...);
4653 (Note that a later change renames 'des_...' into 'DES_...'.)
4656 *) Initial reduction of linker bloat: the use of some functions, such as
4657 PEM causes large amounts of unused functions to be linked in due to
4658 poor organisation. For example pem_all.c contains every PEM function
4659 which has a knock on effect of linking in large amounts of (unused)
4660 ASN1 code. Grouping together similar functions and splitting unrelated
4661 functions prevents this.
4664 *) Cleanup of EVP macros.
4667 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4668 correct _ecb suffix.
4671 *) Add initial OCSP responder support to ocsp application. The
4672 revocation information is handled using the text based index
4673 use by the ca application. The responder can either handle
4674 requests generated internally, supplied in files (for example
4675 via a CGI script) or using an internal minimal server.
4678 *) Add configuration choices to get zlib compression for TLS.
4681 *) Changes to Kerberos SSL for RFC 2712 compliance:
4682 1. Implemented real KerberosWrapper, instead of just using
4683 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4684 2. Implemented optional authenticator field of KerberosWrapper.
4686 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4687 and authenticator structs; see crypto/krb5/.
4689 Generalized Kerberos calls to support multiple Kerberos libraries.
4690 [Vern Staats <staatsvr@asc.hpc.mil>,
4691 Jeffrey Altman <jaltman@columbia.edu>
4692 via Richard Levitte]
4694 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4695 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4696 values for each of the key sizes rather than having just
4697 parameters (and 'speed' generating keys each time).
4700 *) Speed up EVP routines.
4703 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4704 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4705 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4706 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4708 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4709 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4710 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4713 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4715 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4718 *) Added the OS2-EMX target.
4719 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4721 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4722 to support NCONF routines in extension code. New function CONF_set_nconf()
4723 to allow functions which take an NCONF to also handle the old LHASH
4724 structure: this means that the old CONF compatible routines can be
4725 retained (in particular wrt extensions) without having to duplicate the
4726 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4729 *) Enhance the general user interface with mechanisms for inner control
4730 and with possibilities to have yes/no kind of prompts.
4733 *) Change all calls to low level digest routines in the library and
4734 applications to use EVP. Add missing calls to HMAC_cleanup() and
4735 don't assume HMAC_CTX can be copied using memcpy().
4736 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4738 *) Add the possibility to control engines through control names but with
4739 arbitrary arguments instead of just a string.
4740 Change the key loaders to take a UI_METHOD instead of a callback
4741 function pointer. NOTE: this breaks binary compatibility with earlier
4742 versions of OpenSSL [engine].
4743 Adapt the nCipher code for these new conditions and add a card insertion
4747 *) Enhance the general user interface with mechanisms to better support
4748 dialog box interfaces, application-defined prompts, the possibility
4749 to use defaults (for example default passwords from somewhere else)
4750 and interrupts/cancellations.
4753 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4754 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4757 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4758 tidy up some unnecessarily weird code in 'sk_new()').
4759 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4761 *) Change the key loading routines for ENGINEs to use the same kind
4762 callback (pem_password_cb) as all other routines that need this
4766 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4767 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4768 than this minimum value is recommended.
4771 *) New random seeder for OpenVMS, using the system process statistics
4772 that are easily reachable.
4775 *) Windows apparently can't transparently handle global
4776 variables defined in DLLs. Initialisations such as:
4778 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4780 wont compile. This is used by the any applications that need to
4781 declare their own ASN1 modules. This was fixed by adding the option
4782 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4783 needed for static libraries under Win32.
4786 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4787 setting of purpose and trust fields. New X509_STORE trust and
4788 purpose functions and tidy up setting in other SSL functions.
4791 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4792 structure. These are inherited by X509_STORE_CTX when it is
4793 initialised. This allows various defaults to be set in the
4794 X509_STORE structure (such as flags for CRL checking and custom
4795 purpose or trust settings) for functions which only use X509_STORE_CTX
4796 internally such as S/MIME.
4798 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4799 trust settings if they are not set in X509_STORE. This allows X509_STORE
4800 purposes and trust (in S/MIME for example) to override any set by default.
4802 Add command line options for CRL checking to smime, s_client and s_server
4806 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4807 are set then the CRL is looked up in the X509_STORE structure and
4808 its validity and signature checked, then if the certificate is found
4809 in the CRL the verify fails with a revoked error.
4811 Various new CRL related callbacks added to X509_STORE_CTX structure.
4813 Command line options added to 'verify' application to support this.
4815 This needs some additional work, such as being able to handle multiple
4816 CRLs with different times, extension based lookup (rather than just
4817 by subject name) and ultimately more complete V2 CRL extension
4821 *) Add a general user interface API (crypto/ui/). This is designed
4822 to replace things like des_read_password and friends (backward
4823 compatibility functions using this new API are provided).
4824 The purpose is to remove prompting functions from the DES code
4825 section as well as provide for prompting through dialog boxes in
4826 a window system and the like.
4829 *) Add "ex_data" support to ENGINE so implementations can add state at a
4830 per-structure level rather than having to store it globally.
4833 *) Make it possible for ENGINE structures to be copied when retrieved by
4834 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4835 This causes the "original" ENGINE structure to act like a template,
4836 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4837 operational state can be localised to each ENGINE structure, despite the
4838 fact they all share the same "methods". New ENGINE structures returned in
4839 this case have no functional references and the return value is the single
4840 structural reference. This matches the single structural reference returned
4841 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4845 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4846 needs to match any other type at all we need to manually clear the
4850 *) Changes to the "openssl engine" utility to include;
4851 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4852 about an ENGINE's available control commands.
4853 - executing control commands from command line arguments using the
4854 '-pre' and '-post' switches. '-post' is only used if '-t' is
4855 specified and the ENGINE is successfully initialised. The syntax for
4856 the individual commands are colon-separated, for example;
4857 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4860 *) New dynamic control command support for ENGINEs. ENGINEs can now
4861 declare their own commands (numbers), names (strings), descriptions,
4862 and input types for run-time discovery by calling applications. A
4863 subset of these commands are implicitly classed as "executable"
4864 depending on their input type, and only these can be invoked through
4865 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4866 can be based on user input, config files, etc). The distinction is
4867 that "executable" commands cannot return anything other than a boolean
4868 result and can only support numeric or string input, whereas some
4869 discoverable commands may only be for direct use through
4870 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4871 pointers, or other custom uses. The "executable" commands are to
4872 support parameterisations of ENGINE behaviour that can be
4873 unambiguously defined by ENGINEs and used consistently across any
4874 OpenSSL-based application. Commands have been added to all the
4875 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4876 control over shared-library paths without source code alterations.
4879 *) Changed all ENGINE implementations to dynamically allocate their
4880 ENGINEs rather than declaring them statically. Apart from this being
4881 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
4882 this also allows the implementations to compile without using the
4883 internal engine_int.h header.
4886 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
4887 'const' value. Any code that should be able to modify a RAND_METHOD
4888 should already have non-const pointers to it (ie. they should only
4889 modify their own ones).
4892 *) Made a variety of little tweaks to the ENGINE code.
4893 - "atalla" and "ubsec" string definitions were moved from header files
4894 to C code. "nuron" string definitions were placed in variables
4895 rather than hard-coded - allowing parameterisation of these values
4896 later on via ctrl() commands.
4897 - Removed unused "#if 0"'d code.
4898 - Fixed engine list iteration code so it uses ENGINE_free() to release
4899 structural references.
4900 - Constified the RAND_METHOD element of ENGINE structures.
4901 - Constified various get/set functions as appropriate and added
4902 missing functions (including a catch-all ENGINE_cpy that duplicates
4903 all ENGINE values onto a new ENGINE except reference counts/state).
4904 - Removed NULL parameter checks in get/set functions. Setting a method
4905 or function to NULL is a way of cancelling out a previously set
4906 value. Passing a NULL ENGINE parameter is just plain stupid anyway
4907 and doesn't justify the extra error symbols and code.
4908 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
4909 flags from engine_int.h to engine.h.
4910 - Changed prototypes for ENGINE handler functions (init(), finish(),
4911 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
4914 *) Implement binary inversion algorithm for BN_mod_inverse in addition
4915 to the algorithm using long division. The binary algorithm can be
4916 used only if the modulus is odd. On 32-bit systems, it is faster
4917 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
4918 roughly 5-15% for 256-bit moduli), so we use it only for moduli
4919 up to 450 bits. In 64-bit environments, the binary algorithm
4920 appears to be advantageous for much longer moduli; here we use it
4921 for moduli up to 2048 bits.
4924 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
4925 could not support the combine flag in choice fields.
4928 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
4929 extensions from a certificate request to the certificate.
4932 *) Allow multiple 'certopt' and 'nameopt' options to be separated
4933 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
4934 file: this allows the display of the certificate about to be
4935 signed to be customised, to allow certain fields to be included
4936 or excluded and extension details. The old system didn't display
4937 multicharacter strings properly, omitted fields not in the policy
4938 and couldn't display additional details such as extensions.
4941 *) Function EC_POINTs_mul for multiple scalar multiplication
4942 of an arbitrary number of elliptic curve points
4943 \sum scalars[i]*points[i],
4944 optionally including the generator defined for the EC_GROUP:
4945 scalar*generator + \sum scalars[i]*points[i].
4947 EC_POINT_mul is a simple wrapper function for the typical case
4948 that the point list has just one item (besides the optional
4952 *) First EC_METHODs for curves over GF(p):
4954 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
4955 operations and provides various method functions that can also
4956 operate with faster implementations of modular arithmetic.
4958 EC_GFp_mont_method() reuses most functions that are part of
4959 EC_GFp_simple_method, but uses Montgomery arithmetic.
4961 [Bodo Moeller; point addition and point doubling
4962 implementation directly derived from source code provided by
4963 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
4965 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
4966 crypto/ec/ec_lib.c):
4968 Curves are EC_GROUP objects (with an optional group generator)
4969 based on EC_METHODs that are built into the library.
4971 Points are EC_POINT objects based on EC_GROUP objects.
4973 Most of the framework would be able to handle curves over arbitrary
4974 finite fields, but as there are no obvious types for fields other
4975 than GF(p), some functions are limited to that for now.
4978 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
4979 that the file contains a complete HTTP response.
4982 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
4983 change the def and num file printf format specifier from "%-40sXXX"
4984 to "%-39s XXX". The latter will always guarantee a space after the
4985 field while the former will cause them to run together if the field
4986 is 40 of more characters long.
4989 *) Constify the cipher and digest 'method' functions and structures
4990 and modify related functions to take constant EVP_MD and EVP_CIPHER
4994 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
4995 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
4998 *) Modify EVP_Digest*() routines so they now return values. Although the
4999 internal software routines can never fail additional hardware versions
5003 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
5005 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
5006 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
5009 ERR_R_NESTED_ASN1_ERROR
5011 ERR_R_MISSING_ASN1_EOS
5012 were 4 .. 9, conflicting with
5013 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5015 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5016 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5018 Add new error code 'ERR_R_INTERNAL_ERROR'.
5021 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5025 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5026 sets the subject name for a new request or supersedes the
5027 subject name in a given request. Formats that can be parsed are
5028 'CN=Some Name, OU=myOU, C=IT'
5030 'CN=Some Name/OU=myOU/C=IT'.
5032 Add options '-batch' and '-verbose' to 'openssl req'.
5033 [Massimiliano Pala <madwolf@hackmasters.net>]
5035 *) Introduce the possibility to access global variables through
5036 functions on platform were that's the best way to handle exporting
5037 global variables in shared libraries. To enable this functionality,
5038 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5039 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5040 is normally done by Configure or something similar).
5042 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5043 in the source file (foo.c) like this:
5045 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5046 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5048 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5049 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5051 OPENSSL_DECLARE_GLOBAL(int,foo);
5052 #define foo OPENSSL_GLOBAL_REF(foo)
5053 OPENSSL_DECLARE_GLOBAL(double,bar);
5054 #define bar OPENSSL_GLOBAL_REF(bar)
5056 The #defines are very important, and therefore so is including the
5057 header file everywhere where the defined globals are used.
5059 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5060 of ASN.1 items, but that structure is a bit different.
5062 The largest change is in util/mkdef.pl which has been enhanced with
5063 better and easier to understand logic to choose which symbols should
5064 go into the Windows .def files as well as a number of fixes and code
5065 cleanup (among others, algorithm keywords are now sorted
5066 lexicographically to avoid constant rewrites).
5069 *) In BN_div() keep a copy of the sign of 'num' before writing the
5070 result to 'rm' because if rm==num the value will be overwritten
5071 and produce the wrong result if 'num' is negative: this caused
5072 problems with BN_mod() and BN_nnmod().
5075 *) Function OCSP_request_verify(). This checks the signature on an
5076 OCSP request and verifies the signer certificate. The signer
5077 certificate is just checked for a generic purpose and OCSP request
5081 *) Add OCSP_check_validity() function to check the validity of OCSP
5082 responses. OCSP responses are prepared in real time and may only
5083 be a few seconds old. Simply checking that the current time lies
5084 between thisUpdate and nextUpdate max reject otherwise valid responses
5085 caused by either OCSP responder or client clock inaccuracy. Instead
5086 we allow thisUpdate and nextUpdate to fall within a certain period of
5087 the current time. The age of the response can also optionally be
5088 checked. Two new options -validity_period and -status_age added to
5092 *) If signature or public key algorithm is unrecognized print out its
5093 OID rather that just UNKNOWN.
5096 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5097 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5098 ID to be generated from the issuer certificate alone which can then be
5099 passed to OCSP_id_issuer_cmp().
5102 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5103 ASN1 modules to export functions returning ASN1_ITEM pointers
5104 instead of the ASN1_ITEM structures themselves. This adds several
5105 new macros which allow the underlying ASN1 function/structure to
5106 be accessed transparently. As a result code should not use ASN1_ITEM
5107 references directly (such as &X509_it) but instead use the relevant
5108 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5109 use of the new ASN1 code on platforms where exporting structures
5110 is problematical (for example in shared libraries) but exporting
5111 functions returning pointers to structures is not.
5114 *) Add support for overriding the generation of SSL/TLS session IDs.
5115 These callbacks can be registered either in an SSL_CTX or per SSL.
5116 The purpose of this is to allow applications to control, if they wish,
5117 the arbitrary values chosen for use as session IDs, particularly as it
5118 can be useful for session caching in multiple-server environments. A
5119 command-line switch for testing this (and any client code that wishes
5120 to use such a feature) has been added to "s_server".
5121 [Geoff Thorpe, Lutz Jaenicke]
5123 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5124 of the form '#if defined(...) || defined(...) || ...' and
5125 '#if !defined(...) && !defined(...) && ...'. This also avoids
5126 the growing number of special cases it was previously handling.
5129 *) Make all configuration macros available for application by making
5130 sure they are available in opensslconf.h, by giving them names starting
5131 with "OPENSSL_" to avoid conflicts with other packages and by making
5132 sure e_os2.h will cover all platform-specific cases together with
5134 Additionally, it is now possible to define configuration/platform-
5135 specific names (called "system identities"). In the C code, these
5136 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
5137 macro with the name beginning with "OPENSSL_SYS_", which is determined
5138 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
5142 *) New option -set_serial to 'req' and 'x509' this allows the serial
5143 number to use to be specified on the command line. Previously self
5144 signed certificates were hard coded with serial number 0 and the
5145 CA options of 'x509' had to use a serial number in a file which was
5149 *) New options to 'ca' utility to support V2 CRL entry extensions.
5150 Currently CRL reason, invalidity date and hold instruction are
5151 supported. Add new CRL extensions to V3 code and some new objects.
5154 *) New function EVP_CIPHER_CTX_set_padding() this is used to
5155 disable standard block padding (aka PKCS#5 padding) in the EVP
5156 API, which was previously mandatory. This means that the data is
5157 not padded in any way and so the total length much be a multiple
5158 of the block size, otherwise an error occurs.
5161 *) Initial (incomplete) OCSP SSL support.
5164 *) New function OCSP_parse_url(). This splits up a URL into its host,
5165 port and path components: primarily to parse OCSP URLs. New -url
5166 option to ocsp utility.
5169 *) New nonce behavior. The return value of OCSP_check_nonce() now
5170 reflects the various checks performed. Applications can decide
5171 whether to tolerate certain situations such as an absent nonce
5172 in a response when one was present in a request: the ocsp application
5173 just prints out a warning. New function OCSP_add1_basic_nonce()
5174 this is to allow responders to include a nonce in a response even if
5175 the request is nonce-less.
5178 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
5179 skipped when using openssl x509 multiple times on a single input file,
5180 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
5183 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
5184 set string type: to handle setting ASN1_TIME structures. Fix ca
5185 utility to correctly initialize revocation date of CRLs.
5188 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
5189 the clients preferred ciphersuites and rather use its own preferences.
5190 Should help to work around M$ SGC (Server Gated Cryptography) bug in
5191 Internet Explorer by ensuring unchanged hash method during stepup.
5192 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
5195 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
5196 to aes and add a new 'exist' option to print out symbols that don't
5200 *) Additional options to ocsp utility to allow flags to be set and
5201 additional certificates supplied.
5204 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
5205 OCSP client a number of certificate to only verify the response
5209 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
5210 handle the new API. Currently only ECB, CBC modes supported. Add new
5213 Add TLS AES ciphersuites as described in RFC3268, "Advanced
5214 Encryption Standard (AES) Ciphersuites for Transport Layer
5215 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
5216 not enabled by default and were not part of the "ALL" ciphersuite
5217 alias because they were not yet official; they could be
5218 explicitly requested by specifying the "AESdraft" ciphersuite
5219 group alias. In the final release of OpenSSL 0.9.7, the group
5220 alias is called "AES" and is part of "ALL".)
5221 [Ben Laurie, Steve Henson, Bodo Moeller]
5223 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
5224 request to response.
5227 *) Functions for OCSP responders. OCSP_request_onereq_count(),
5228 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
5229 extract information from a certificate request. OCSP_response_create()
5230 creates a response and optionally adds a basic response structure.
5231 OCSP_basic_add1_status() adds a complete single response to a basic
5232 response and returns the OCSP_SINGLERESP structure just added (to allow
5233 extensions to be included for example). OCSP_basic_add1_cert() adds a
5234 certificate to a basic response and OCSP_basic_sign() signs a basic
5235 response with various flags. New helper functions ASN1_TIME_check()
5236 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
5237 (converts ASN1_TIME to GeneralizedTime).
5240 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
5241 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
5242 structure from a certificate. X509_pubkey_digest() digests the public_key
5243 contents: this is used in various key identifiers.
5246 *) Make sk_sort() tolerate a NULL argument.
5247 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
5249 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
5250 passed by the function are trusted implicitly. If any of them signed the
5251 response then it is assumed to be valid and is not verified.
5254 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
5255 to data. This was previously part of the PKCS7 ASN1 code. This
5256 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
5257 [Steve Henson, reported by Kenneth R. Robinette
5258 <support@securenetterm.com>]
5260 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
5261 routines: without these tracing memory leaks is very painful.
5262 Fix leaks in PKCS12 and PKCS7 routines.
5265 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
5266 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
5267 effectively meant GeneralizedTime would never be used. Now it
5268 is initialised to -1 but X509_time_adj() now has to check the value
5269 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
5270 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
5271 [Steve Henson, reported by Kenneth R. Robinette
5272 <support@securenetterm.com>]
5274 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
5275 result in a zero length in the ASN1_INTEGER structure which was
5276 not consistent with the structure when d2i_ASN1_INTEGER() was used
5277 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
5278 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
5279 where it did not print out a minus for negative ASN1_INTEGER.
5282 *) Add summary printout to ocsp utility. The various functions which
5283 convert status values to strings have been renamed to:
5284 OCSP_response_status_str(), OCSP_cert_status_str() and
5285 OCSP_crl_reason_str() and are no longer static. New options
5286 to verify nonce values and to disable verification. OCSP response
5287 printout format cleaned up.
5290 *) Add additional OCSP certificate checks. These are those specified
5291 in RFC2560. This consists of two separate checks: the CA of the
5292 certificate being checked must either be the OCSP signer certificate
5293 or the issuer of the OCSP signer certificate. In the latter case the
5294 OCSP signer certificate must contain the OCSP signing extended key
5295 usage. This check is performed by attempting to match the OCSP
5296 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
5297 in the OCSP_CERTID structures of the response.
5300 *) Initial OCSP certificate verification added to OCSP_basic_verify()
5301 and related routines. This uses the standard OpenSSL certificate
5302 verify routines to perform initial checks (just CA validity) and
5303 to obtain the certificate chain. Then additional checks will be
5304 performed on the chain. Currently the root CA is checked to see
5305 if it is explicitly trusted for OCSP signing. This is used to set
5306 a root CA as a global signing root: that is any certificate that
5307 chains to that CA is an acceptable OCSP signing certificate.
5310 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5311 extensions from a separate configuration file.
5312 As when reading extensions from the main configuration file,
5313 the '-extensions ...' option may be used for specifying the
5315 [Massimiliano Pala <madwolf@comune.modena.it>]
5317 *) New OCSP utility. Allows OCSP requests to be generated or
5318 read. The request can be sent to a responder and the output
5319 parsed, outputed or printed in text form. Not complete yet:
5320 still needs to check the OCSP response validity.
5323 *) New subcommands for 'openssl ca':
5324 'openssl ca -status <serial>' prints the status of the cert with
5325 the given serial number (according to the index file).
5326 'openssl ca -updatedb' updates the expiry status of certificates
5328 [Massimiliano Pala <madwolf@comune.modena.it>]
5330 *) New '-newreq-nodes' command option to CA.pl. This is like
5331 '-newreq', but calls 'openssl req' with the '-nodes' option
5332 so that the resulting key is not encrypted.
5333 [Damien Miller <djm@mindrot.org>]
5335 *) New configuration for the GNU Hurd.
5336 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5338 *) Initial code to implement OCSP basic response verify. This
5339 is currently incomplete. Currently just finds the signer's
5340 certificate and verifies the signature on the response.
5343 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5344 value of OPENSSLDIR. This is available via the new '-d' option
5345 to 'openssl version', and is also included in 'openssl version -a'.
5348 *) Allowing defining memory allocation callbacks that will be given
5349 file name and line number information in additional arguments
5350 (a const char* and an int). The basic functionality remains, as
5351 well as the original possibility to just replace malloc(),
5352 realloc() and free() by functions that do not know about these
5353 additional arguments. To register and find out the current
5354 settings for extended allocation functions, the following
5355 functions are provided:
5357 CRYPTO_set_mem_ex_functions
5358 CRYPTO_set_locked_mem_ex_functions
5359 CRYPTO_get_mem_ex_functions
5360 CRYPTO_get_locked_mem_ex_functions
5362 These work the same way as CRYPTO_set_mem_functions and friends.
5363 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5364 extended allocation function is enabled.
5365 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5366 a conventional allocation function is enabled.
5367 [Richard Levitte, Bodo Moeller]
5369 *) Finish off removing the remaining LHASH function pointer casts.
5370 There should no longer be any prototype-casting required when using
5371 the LHASH abstraction, and any casts that remain are "bugs". See
5372 the callback types and macros at the head of lhash.h for details
5373 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5376 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5377 If /dev/[u]random devices are not available or do not return enough
5378 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5380 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5381 /etc/entropy will be queried once each in this sequence, quering stops
5382 when enough entropy was collected without querying more sockets.
5385 *) Change the Unix RAND_poll() variant to be able to poll several
5386 random devices, as specified by DEVRANDOM, until a sufficient amount
5387 of data has been collected. We spend at most 10 ms on each file
5388 (select timeout) and read in non-blocking mode. DEVRANDOM now
5389 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5390 (previously it was just the string "/dev/urandom"), so on typical
5391 platforms the 10 ms delay will never occur.
5392 Also separate out the Unix variant to its own file, rand_unix.c.
5393 For VMS, there's a currently-empty rand_vms.c.
5396 *) Move OCSP client related routines to ocsp_cl.c. These
5397 provide utility functions which an application needing
5398 to issue a request to an OCSP responder and analyse the
5399 response will typically need: as opposed to those which an
5400 OCSP responder itself would need which will be added later.
5402 OCSP_request_sign() signs an OCSP request with an API similar
5403 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5404 response. OCSP_response_get1_basic() extracts basic response
5405 from response. OCSP_resp_find_status(): finds and extracts status
5406 information from an OCSP_CERTID structure (which will be created
5407 when the request structure is built). These are built from lower
5408 level functions which work on OCSP_SINGLERESP structures but
5409 wont normally be used unless the application wishes to examine
5410 extensions in the OCSP response for example.
5412 Replace nonce routines with a pair of functions.
5413 OCSP_request_add1_nonce() adds a nonce value and optionally
5414 generates a random value. OCSP_check_nonce() checks the
5415 validity of the nonce in an OCSP response.
5418 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5419 This doesn't copy the supplied OCSP_CERTID and avoids the
5420 need to free up the newly created id. Change return type
5421 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5422 This can then be used to add extensions to the request.
5423 Deleted OCSP_request_new(), since most of its functionality
5424 is now in OCSP_REQUEST_new() (and the case insensitive name
5425 clash) apart from the ability to set the request name which
5426 will be added elsewhere.
5429 *) Update OCSP API. Remove obsolete extensions argument from
5430 various functions. Extensions are now handled using the new
5431 OCSP extension code. New simple OCSP HTTP function which
5432 can be used to send requests and parse the response.
5435 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5436 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5437 uses the special reorder version of SET OF to sort the attributes
5438 and reorder them to match the encoded order. This resolves a long
5439 standing problem: a verify on a PKCS7 structure just after signing
5440 it used to fail because the attribute order did not match the
5441 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5442 it uses the received order. This is necessary to tolerate some broken
5443 software that does not order SET OF. This is handled by encoding
5444 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5445 to produce the required SET OF.
5448 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5449 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5450 files to get correct declarations of the ASN.1 item variables.
5453 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5454 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5455 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5456 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5457 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5458 ASN1_ITEM and no wrapper functions.
5461 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5462 replace the old function pointer based I/O routines. Change most of
5463 the *_d2i_bio() and *_d2i_fp() functions to use these.
5466 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5467 lines, recognice more "algorithms" that can be deselected, and make
5468 it complain about algorithm deselection that isn't recognised.
5471 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5472 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5473 to use new functions. Add NO_ASN1_OLD which can be set to remove
5474 some old style ASN1 functions: this can be used to determine if old
5475 code will still work when these eventually go away.
5478 *) New extension functions for OCSP structures, these follow the
5479 same conventions as certificates and CRLs.
5482 *) New function X509V3_add1_i2d(). This automatically encodes and
5483 adds an extension. Its behaviour can be customised with various
5484 flags to append, replace or delete. Various wrappers added for
5485 certifcates and CRLs.
5488 *) Fix to avoid calling the underlying ASN1 print routine when
5489 an extension cannot be parsed. Correct a typo in the
5490 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5493 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5494 entries for variables.
5497 *) Add functionality to apps/openssl.c for detecting locking
5498 problems: As the program is single-threaded, all we have
5499 to do is register a locking callback using an array for
5500 storing which locks are currently held by the program.
5503 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5504 SSL_get_ex_data_X509_STORE_idx(), which is used in
5505 ssl_verify_cert_chain() and thus can be called at any time
5506 during TLS/SSL handshakes so that thread-safety is essential.
5507 Unfortunately, the ex_data design is not at all suited
5508 for multi-threaded use, so it probably should be abolished.
5511 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5512 [Broadcom, tweaked and integrated by Geoff Thorpe]
5514 *) Move common extension printing code to new function
5515 X509V3_print_extensions(). Reorganise OCSP print routines and
5516 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5519 *) New function X509_signature_print() to remove duplication in some
5523 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5524 set (this was treated exactly the same as SET OF previously). This
5525 is used to reorder the STACK representing the structure to match the
5526 encoding. This will be used to get round a problem where a PKCS7
5527 structure which was signed could not be verified because the STACK
5528 order did not reflect the encoded order.
5531 *) Reimplement the OCSP ASN1 module using the new code.
5534 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5535 for its ASN1 operations. The old style function pointers still exist
5536 for now but they will eventually go away.
5539 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5540 completely replaces the old ASN1 functionality with a table driven
5541 encoder and decoder which interprets an ASN1_ITEM structure describing
5542 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5543 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5544 has also been converted to the new form.
5547 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5548 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5549 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5550 for negative moduli.
5553 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5554 of not touching the result's sign bit.
5557 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5561 *) Changed the LHASH code to use prototypes for callbacks, and created
5562 macros to declare and implement thin (optionally static) functions
5563 that provide type-safety and avoid function pointer casting for the
5564 type-specific callbacks.
5567 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5569 [Veers Staats <staatsvr@asc.hpc.mil>,
5570 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5572 *) Reformat the FAQ so the different questions and answers can be divided
5573 in sections depending on the subject.
5576 *) Have the zlib compression code load ZLIB.DLL dynamically under
5580 *) New function BN_mod_sqrt for computing square roots modulo a prime
5581 (using the probabilistic Tonelli-Shanks algorithm unless
5582 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5583 be handled deterministically).
5584 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5586 *) Make BN_mod_inverse faster by explicitly handling small quotients
5587 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5588 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5591 *) New function BN_kronecker.
5594 *) Fix BN_gcd so that it works on negative inputs; the result is
5595 positive unless both parameters are zero.
5596 Previously something reasonably close to an infinite loop was
5597 possible because numbers could be growing instead of shrinking
5598 in the implementation of Euclid's algorithm.
5601 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5602 sign of the number in question.
5604 Fix BN_is_word(a,w) to work correctly for w == 0.
5606 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5607 because its test if the absolute value of 'a' equals 'w'.
5608 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5609 it exists mostly for use in the implementations of BN_is_zero(),
5610 BN_is_one(), and BN_is_word().
5613 *) New function BN_swap.
5616 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5617 the exponentiation functions are more likely to produce reasonable
5618 results on negative inputs.
5621 *) Change BN_mod_mul so that the result is always non-negative.
5622 Previously, it could be negative if one of the factors was negative;
5623 I don't think anyone really wanted that behaviour.
5626 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5627 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5628 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5629 and add new functions:
5638 BN_mod_lshift1_quick
5642 These functions always generate non-negative results.
5644 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5645 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5647 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5648 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5649 be reduced modulo m.
5650 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5653 The following entry accidentily appeared in the CHANGES file
5654 distributed with OpenSSL 0.9.7. The modifications described in
5655 it do *not* apply to OpenSSL 0.9.7.
5657 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5658 was actually never needed) and in BN_mul(). The removal in BN_mul()
5659 required a small change in bn_mul_part_recursive() and the addition
5660 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5661 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5662 bn_sub_words() and bn_add_words() except they take arrays with
5667 *) In 'openssl passwd', verify passwords read from the terminal
5668 unless the '-salt' option is used (which usually means that
5669 verification would just waste user's time since the resulting
5670 hash is going to be compared with some given password hash)
5671 or the new '-noverify' option is used.
5673 This is an incompatible change, but it does not affect
5674 non-interactive use of 'openssl passwd' (passwords on the command
5675 line, '-stdin' option, '-in ...' option) and thus should not
5679 *) Remove all references to RSAref, since there's no more need for it.
5682 *) Make DSO load along a path given through an environment variable
5683 (SHLIB_PATH) with shl_load().
5686 *) Constify the ENGINE code as a result of BIGNUM constification.
5687 Also constify the RSA code and most things related to it. In a
5688 few places, most notable in the depth of the ASN.1 code, ugly
5689 casts back to non-const were required (to be solved at a later
5693 *) Make it so the openssl application has all engines loaded by default.
5696 *) Constify the BIGNUM routines a little more.
5699 *) Add the following functions:
5701 ENGINE_load_cswift()
5703 ENGINE_load_atalla()
5705 ENGINE_load_builtin_engines()
5707 That way, an application can itself choose if external engines that
5708 are built-in in OpenSSL shall ever be used or not. The benefit is
5709 that applications won't have to be linked with libdl or other dso
5710 libraries unless it's really needed.
5712 Changed 'openssl engine' to load all engines on demand.
5713 Changed the engine header files to avoid the duplication of some
5714 declarations (they differed!).
5717 *) 'openssl engine' can now list capabilities.
5720 *) Better error reporting in 'openssl engine'.
5723 *) Never call load_dh_param(NULL) in s_server.
5726 *) Add engine application. It can currently list engines by name and
5727 identity, and test if they are actually available.
5730 *) Improve RPM specification file by forcing symbolic linking and making
5731 sure the installed documentation is also owned by root.root.
5732 [Damien Miller <djm@mindrot.org>]
5734 *) Give the OpenSSL applications more possibilities to make use of
5735 keys (public as well as private) handled by engines.
5738 *) Add OCSP code that comes from CertCo.
5741 *) Add VMS support for the Rijndael code.
5744 *) Added untested support for Nuron crypto accelerator.
5747 *) Add support for external cryptographic devices. This code was
5748 previously distributed separately as the "engine" branch.
5749 [Geoff Thorpe, Richard Levitte]
5751 *) Rework the filename-translation in the DSO code. It is now possible to
5752 have far greater control over how a "name" is turned into a filename
5753 depending on the operating environment and any oddities about the
5754 different shared library filenames on each system.
5757 *) Support threads on FreeBSD-elf in Configure.
5760 *) Fix for SHA1 assembly problem with MASM: it produces
5761 warnings about corrupt line number information when assembling
5762 with debugging information. This is caused by the overlapping
5764 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5767 NCONF_get_number() has no error checking at all. As a replacement,
5768 NCONF_get_number_e() is defined (_e for "error checking") and is
5769 promoted strongly. The old NCONF_get_number is kept around for
5770 binary backward compatibility.
5771 Make it possible for methods to load from something other than a BIO,
5772 by providing a function pointer that is given a name instead of a BIO.
5773 For example, this could be used to load configuration data from an
5777 *) Fix for non blocking accept BIOs. Added new I/O special reason
5778 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5779 with non blocking I/O was not possible because no retry code was
5780 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5784 *) Added the beginnings of Rijndael support.
5787 *) Fix for bug in DirectoryString mask setting. Add support for
5788 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5789 to allow certificate printing to more controllable, additional
5790 'certopt' option to 'x509' to allow new printing options to be
5794 *) Clean old EAY MD5 hack from e_os.h.
5797 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5799 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5800 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5801 [Joe Orton, Steve Henson]
5803 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5805 *) Fix additional bug revealed by the NISCC test suite:
5807 Stop bug triggering large recursion when presented with
5808 certain ASN.1 tags (CVE-2003-0851)
5811 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5813 *) Fix various bugs revealed by running the NISCC test suite:
5815 Stop out of bounds reads in the ASN1 code when presented with
5816 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5818 If verify callback ignores invalid public key errors don't try to check
5819 certificate signature with the NULL public key.
5823 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5824 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5828 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5829 extra data after the compression methods not only for TLS 1.0
5830 but also for SSL 3.0 (as required by the specification).
5831 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5833 *) Change X509_certificate_type() to mark the key as exported/exportable
5834 when it's 512 *bits* long, not 512 bytes.
5837 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5839 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5840 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5841 a protocol version number mismatch like a decryption error
5842 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5845 *) Turn on RSA blinding by default in the default implementation
5846 to avoid a timing attack. Applications that don't want it can call
5847 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5848 They would be ill-advised to do so in most cases.
5849 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5851 *) Change RSA blinding code so that it works when the PRNG is not
5852 seeded (in this case, the secret RSA exponent is abused as
5853 an unpredictable seed -- if it is not unpredictable, there
5854 is no point in blinding anyway). Make RSA blinding thread-safe
5855 by remembering the creator's thread ID in rsa->blinding and
5856 having all other threads use local one-time blinding factors
5857 (this requires more computation than sharing rsa->blinding, but
5858 avoids excessive locking; and if an RSA object is not shared
5859 between threads, blinding will still be very fast).
5862 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5864 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5865 via timing by performing a MAC computation even if incorrrect
5866 block cipher padding has been found. This is a countermeasure
5867 against active attacks where the attacker has to distinguish
5868 between bad padding and a MAC verification error. (CVE-2003-0078)
5870 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5871 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5872 Martin Vuagnoux (EPFL, Ilion)]
5874 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5876 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5877 memory from it's contents. This is done with a counter that will
5878 place alternating values in each byte. This can be used to solve
5879 two issues: 1) the removal of calls to memset() by highly optimizing
5880 compilers, and 2) cleansing with other values than 0, since those can
5881 be read through on certain media, for example a swap space on disk.
5884 *) Bugfix: client side session caching did not work with external caching,
5885 because the session->cipher setting was not restored when reloading
5886 from the external cache. This problem was masked, when
5887 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
5888 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
5891 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
5892 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
5893 [Zeev Lieber <zeev-l@yahoo.com>]
5895 *) Undo an undocumented change introduced in 0.9.6e which caused
5896 repeated calls to OpenSSL_add_all_ciphers() and
5897 OpenSSL_add_all_digests() to be ignored, even after calling
5901 *) Change the default configuration reader to deal with last line not
5902 being properly terminated.
5905 *) Change X509_NAME_cmp() so it applies the special rules on handling
5906 DN values that are of type PrintableString, as well as RDNs of type
5907 emailAddress where the value has the type ia5String.
5908 [stefank@valicert.com via Richard Levitte]
5910 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
5911 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
5912 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
5913 the bitwise-OR of the two for use by the majority of applications
5914 wanting this behaviour, and update the docs. The documented
5915 behaviour and actual behaviour were inconsistent and had been
5916 changing anyway, so this is more a bug-fix than a behavioural
5918 [Geoff Thorpe, diagnosed by Nadav Har'El]
5920 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
5921 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
5924 *) Fix initialization code race conditions in
5925 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
5926 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
5927 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
5928 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
5929 ssl2_get_cipher_by_char(),
5930 ssl3_get_cipher_by_char().
5931 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
5933 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
5934 the cached sessions are flushed, as the remove_cb() might use ex_data
5935 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
5936 (see [openssl.org #212]).
5937 [Geoff Thorpe, Lutz Jaenicke]
5939 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
5940 length, instead of the encoding length to d2i_ASN1_OBJECT.
5943 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
5945 *) [In 0.9.6g-engine release:]
5946 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
5947 [Lynn Gazis <lgazis@rainbow.com>]
5949 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
5951 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
5952 and get fix the header length calculation.
5953 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
5954 Alon Kantor <alonk@checkpoint.com> (and others),
5957 *) Use proper error handling instead of 'assertions' in buffer
5958 overflow checks added in 0.9.6e. This prevents DoS (the
5959 assertions could call abort()).
5960 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
5962 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
5964 *) Add various sanity checks to asn1_get_length() to reject
5965 the ASN1 length bytes if they exceed sizeof(long), will appear
5966 negative or the content length exceeds the length of the
5968 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5970 *) Fix cipher selection routines: ciphers without encryption had no flags
5971 for the cipher strength set and where therefore not handled correctly
5972 by the selection routines (PR #130).
5975 *) Fix EVP_dsa_sha macro.
5979 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
5980 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
5981 that was added in OpenSSL 0.9.6d.
5983 As the countermeasure turned out to be incompatible with some
5984 broken SSL implementations, the new option is part of SSL_OP_ALL.
5985 SSL_OP_ALL is usually employed when compatibility with weird SSL
5986 implementations is desired (e.g. '-bugs' option to 's_client' and
5987 's_server'), so the new option is automatically set in many
5991 *) Changes in security patch:
5993 Changes marked "(CHATS)" were sponsored by the Defense Advanced
5994 Research Projects Agency (DARPA) and Air Force Research Laboratory,
5995 Air Force Materiel Command, USAF, under agreement number
5998 *) Add various sanity checks to asn1_get_length() to reject
5999 the ASN1 length bytes if they exceed sizeof(long), will appear
6000 negative or the content length exceeds the length of the
6001 supplied buffer. (CVE-2002-0659)
6002 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6004 *) Assertions for various potential buffer overflows, not known to
6006 [Ben Laurie (CHATS)]
6008 *) Various temporary buffers to hold ASCII versions of integers were
6009 too small for 64 bit platforms. (CVE-2002-0655)
6010 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6012 *) Remote buffer overflow in SSL3 protocol - an attacker could
6013 supply an oversized session ID to a client. (CVE-2002-0656)
6014 [Ben Laurie (CHATS)]
6016 *) Remote buffer overflow in SSL2 protocol - an attacker could
6017 supply an oversized client master key. (CVE-2002-0656)
6018 [Ben Laurie (CHATS)]
6020 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6022 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6023 encoded as NULL) with id-dsa-with-sha1.
6024 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6026 *) Check various X509_...() return values in apps/req.c.
6027 [Nils Larsch <nla@trustcenter.de>]
6029 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6030 an end-of-file condition would erronously be flagged, when the CRLF
6031 was just at the end of a processed block. The bug was discovered when
6032 processing data through a buffering memory BIO handing the data to a
6033 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6034 <ptsekov@syntrex.com> and Nedelcho Stanev.
6037 *) Implement a countermeasure against a vulnerability recently found
6038 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6039 before application data chunks to avoid the use of known IVs
6040 with data potentially chosen by the attacker.
6043 *) Fix length checks in ssl3_get_client_hello().
6046 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6047 to prevent ssl3_read_internal() from incorrectly assuming that
6048 ssl3_read_bytes() found application data while handshake
6049 processing was enabled when in fact s->s3->in_read_app_data was
6050 merely automatically cleared during the initial handshake.
6051 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6053 *) Fix object definitions for Private and Enterprise: they were not
6054 recognized in their shortname (=lowercase) representation. Extend
6055 obj_dat.pl to issue an error when using undefined keywords instead
6056 of silently ignoring the problem (Svenning Sorensen
6057 <sss@sss.dnsalias.net>).
6060 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6061 generators, i.e. generators other than 2 and 5. (Previously, the
6062 code did not properly initialise the 'add' and 'rem' values to
6063 BN_generate_prime().)
6065 In the new general case, we do not insist that 'generator' is
6066 actually a primitive root: This requirement is rather pointless;
6067 a generator of the order-q subgroup is just as good, if not
6071 *) Map new X509 verification errors to alerts. Discovered and submitted by
6072 Tom Wu <tom@arcot.com>.
6075 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6076 returning non-zero before the data has been completely received
6077 when using non-blocking I/O.
6078 [Bodo Moeller; problem pointed out by John Hughes]
6080 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6081 [Ben Laurie, Lutz Jaenicke]
6083 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6084 Yoram Zahavi <YoramZ@gilian.com>).
6087 *) Add information about CygWin 1.3 and on, and preserve proper
6088 configuration for the versions before that.
6089 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6091 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6092 check whether we deal with a copy of a session and do not delete from
6093 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6094 <izhar@checkpoint.com>.
6097 *) Do not store session data into the internal session cache, if it
6098 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6099 flag is set). Proposed by Aslam <aslam@funk.com>.
6102 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6106 *) [In 0.9.6d-engine release:]
6107 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6108 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6110 *) Add the configuration target linux-s390x.
6111 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6113 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6114 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6115 variable as an indication that a ClientHello message has been
6116 received. As the flag value will be lost between multiple
6117 invocations of ssl3_accept when using non-blocking I/O, the
6118 function may not be aware that a handshake has actually taken
6119 place, thus preventing a new session from being added to the
6122 To avoid this problem, we now set s->new_session to 2 instead of
6123 using a local variable.
6124 [Lutz Jaenicke, Bodo Moeller]
6126 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6127 if the SSL_R_LENGTH_MISMATCH error is detected.
6128 [Geoff Thorpe, Bodo Moeller]
6130 *) New 'shared_ldflag' column in Configure platform table.
6133 *) Fix EVP_CIPHER_mode macro.
6134 ["Dan S. Camper" <dan@bti.net>]
6136 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
6137 type, we must throw them away by setting rr->length to 0.
6138 [D P Chang <dpc@qualys.com>]
6140 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
6142 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
6143 <Dominikus.Scherkl@biodata.com>. (The previous implementation
6144 worked incorrectly for those cases where range = 10..._2 and
6145 3*range is two bits longer than range.)
6148 *) Only add signing time to PKCS7 structures if it is not already
6152 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
6153 OBJ_ld_ce should be OBJ_id_ce.
6154 Also some ip-pda OIDs in crypto/objects/objects.txt were
6155 incorrect (cf. RFC 3039).
6156 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
6158 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
6159 returns early because it has nothing to do.
6160 [Andy Schneider <andy.schneider@bjss.co.uk>]
6162 *) [In 0.9.6c-engine release:]
6163 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
6164 [Andy Schneider <andy.schneider@bjss.co.uk>]
6166 *) [In 0.9.6c-engine release:]
6167 Add support for Cryptographic Appliance's keyserver technology.
6168 (Use engine 'keyclient')
6169 [Cryptographic Appliances and Geoff Thorpe]
6171 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
6172 is called via tools/c89.sh because arguments have to be
6173 rearranged (all '-L' options must appear before the first object
6175 [Richard Shapiro <rshapiro@abinitio.com>]
6177 *) [In 0.9.6c-engine release:]
6178 Add support for Broadcom crypto accelerator cards, backported
6180 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
6182 *) [In 0.9.6c-engine release:]
6183 Add support for SureWare crypto accelerator cards from
6184 Baltimore Technologies. (Use engine 'sureware')
6185 [Baltimore Technologies and Mark Cox]
6187 *) [In 0.9.6c-engine release:]
6188 Add support for crypto accelerator cards from Accelerated
6189 Encryption Processing, www.aep.ie. (Use engine 'aep')
6190 [AEP Inc. and Mark Cox]
6192 *) Add a configuration entry for gcc on UnixWare.
6193 [Gary Benson <gbenson@redhat.com>]
6195 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
6196 messages are stored in a single piece (fixed-length part and
6197 variable-length part combined) and fix various bugs found on the way.
6200 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
6201 instead. BIO_gethostbyname() does not know what timeouts are
6202 appropriate, so entries would stay in cache even when they have
6204 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
6206 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
6207 faced with a pathologically small ClientHello fragment that does
6208 not contain client_version: Instead of aborting with an error,
6209 simply choose the highest available protocol version (i.e.,
6210 TLS 1.0 unless it is disabled). In practice, ClientHello
6211 messages are never sent like this, but this change gives us
6212 strictly correct behaviour at least for TLS.
6215 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
6216 never resets s->method to s->ctx->method when called from within
6217 one of the SSL handshake functions.
6218 [Bodo Moeller; problem pointed out by Niko Baric]
6220 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
6221 (sent using the client's version number) if client_version is
6222 smaller than the protocol version in use. Also change
6223 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
6224 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
6225 the client will at least see that alert.
6228 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
6232 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
6233 client receives HelloRequest while in a handshake.
6234 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
6236 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
6237 should end in 'break', not 'goto end' which circuments various
6238 cleanups done in state SSL_ST_OK. But session related stuff
6239 must be disabled for SSL_ST_OK in the case that we just sent a
6242 Also avoid some overhead by not calling ssl_init_wbio_buffer()
6243 before just sending a HelloRequest.
6244 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
6246 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
6247 reveal whether illegal block cipher padding was found or a MAC
6248 verification error occured. (Neither SSLerr() codes nor alerts
6249 are directly visible to potential attackers, but the information
6250 may leak via logfiles.)
6252 Similar changes are not required for the SSL 2.0 implementation
6253 because the number of padding bytes is sent in clear for SSL 2.0,
6254 and the extra bytes are just ignored. However ssl/s2_pkt.c
6255 failed to verify that the purported number of padding bytes is in
6259 *) Add OpenUNIX-8 support including shared libraries
6260 (Boyd Lynn Gerber <gerberb@zenez.com>).
6263 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
6264 'wristwatch attack' using huge encoding parameters (cf.
6265 James H. Manger's CRYPTO 2001 paper). Note that the
6266 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
6267 encoding parameters and hence was not vulnerable.
6270 *) BN_sqr() bug fix.
6271 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
6273 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
6274 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
6275 followed by modular reduction.
6276 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
6278 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
6279 equivalent based on BN_pseudo_rand() instead of BN_rand().
6282 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
6283 This function was broken, as the check for a new client hello message
6284 to handle SGC did not allow these large messages.
6285 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
6288 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
6291 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
6292 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
6295 *) Rework the configuration and shared library support for Tru64 Unix.
6296 The configuration part makes use of modern compiler features and
6297 still retains old compiler behavior for those that run older versions
6298 of the OS. The shared library support part includes a variant that
6299 uses the RPATH feature, and is available through the special
6300 configuration target "alpha-cc-rpath", which will never be selected
6302 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
6304 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
6305 with the same message size as in ssl3_get_certificate_request().
6306 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
6307 messages might inadvertently be reject as too long.
6308 [Petr Lampa <lampa@fee.vutbr.cz>]
6310 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6313 *) Modified SSL library such that the verify_callback that has been set
6314 specificly for an SSL object with SSL_set_verify() is actually being
6315 used. Before the change, a verify_callback set with this function was
6316 ignored and the verify_callback() set in the SSL_CTX at the time of
6317 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6318 to allow the necessary settings.
6321 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6322 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6323 done automatically (in contradiction to the requirements of the C
6324 standard). This made problems when used from OpenSSH.
6327 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6328 dh->length and always used
6330 BN_rand_range(priv_key, dh->p).
6332 BN_rand_range() is not necessary for Diffie-Hellman, and this
6333 specific range makes Diffie-Hellman unnecessarily inefficient if
6334 dh->length (recommended exponent length) is much smaller than the
6335 length of dh->p. We could use BN_rand_range() if the order of
6336 the subgroup was stored in the DH structure, but we only have
6341 BN_rand(priv_key, l, ...)
6343 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6349 RSA_eay_public_encrypt
6350 RSA_eay_private_decrypt
6351 RSA_eay_private_encrypt (signing)
6352 RSA_eay_public_decrypt (signature verification)
6354 (default implementations for RSA_public_encrypt,
6355 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6356 always reject numbers >= n.
6359 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6360 to synchronize access to 'locking_thread'. This is necessary on
6361 systems where access to 'locking_thread' (an 'unsigned long'
6362 variable) is not atomic.
6365 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6366 *before* setting the 'crypto_lock_rand' flag. The previous code had
6367 a race condition if 0 is a valid thread ID.
6368 [Travis Vitek <vitek@roguewave.com>]
6370 *) Add support for shared libraries under Irix.
6371 [Albert Chin-A-Young <china@thewrittenword.com>]
6373 *) Add configuration option to build on Linux on both big-endian and
6375 [Ralf Baechle <ralf@uni-koblenz.de>]
6377 *) Add the possibility to create shared libraries on HP-UX.
6380 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6382 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6383 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6384 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6385 PRNG state recovery was possible based on the output of
6386 one PRNG request appropriately sized to gain knowledge on
6387 'md' followed by enough consecutive 1-byte PRNG requests
6388 to traverse all of 'state'.
6390 1. When updating 'md_local' (the current thread's copy of 'md')
6391 during PRNG output generation, hash all of the previous
6392 'md_local' value, not just the half used for PRNG output.
6394 2. Make the number of bytes from 'state' included into the hash
6395 independent from the number of PRNG bytes requested.
6397 The first measure alone would be sufficient to avoid
6398 Markku-Juhani's attack. (Actually it had never occurred
6399 to me that the half of 'md_local' used for chaining was the
6400 half from which PRNG output bytes were taken -- I had always
6401 assumed that the secret half would be used.) The second
6402 measure makes sure that additional data from 'state' is never
6403 mixed into 'md_local' in small portions; this heuristically
6404 further strengthens the PRNG.
6407 *) Fix crypto/bn/asm/mips3.s.
6410 *) When only the key is given to "enc", the IV is undefined. Print out
6411 an error message in this case.
6414 *) Handle special case when X509_NAME is empty in X509 printing routines.
6417 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6418 positive and less than q.
6421 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6422 used: it isn't thread safe and the add_lock_callback should handle
6424 [Paul Rose <Paul.Rose@bridge.com>]
6426 *) Verify that incoming data obeys the block size in
6427 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6431 [Ulf Möller, Bodo Möller]
6433 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6434 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6435 when fixing the server behaviour for backwards-compatible 'client
6436 hello' messages. (Note that the attack is impractical against
6437 SSL 3.0 and TLS 1.0 anyway because length and version checking
6438 means that the probability of guessing a valid ciphertext is
6439 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6442 Before 0.9.5, the countermeasure (hide the error by generating a
6443 random 'decryption result') did not work properly because
6444 ERR_clear_error() was missing, meaning that SSL_get_error() would
6445 detect the supposedly ignored error.
6447 Both problems are now fixed.
6450 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6451 (previously it was 1024).
6454 *) Fix for compatibility mode trust settings: ignore trust settings
6455 unless some valid trust or reject settings are present.
6458 *) Fix for blowfish EVP: its a variable length cipher.
6461 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6462 parameters in DSA public key structures and return an error in the
6463 DSA routines if parameters are absent.
6466 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6467 in the current directory if neither $RANDFILE nor $HOME was set.
6468 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6469 caused some confusion to Windows users who haven't defined $HOME.
6470 Thus RAND_file_name() is changed again: e_os.h can define a
6471 DEFAULT_HOME, which will be used if $HOME is not set.
6472 For Windows, we use "C:"; on other platforms, we still require
6473 environment variables.
6475 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6476 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6477 having multiple threads call RAND_poll() concurrently.
6480 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6481 combination of a flag and a thread ID variable.
6482 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6483 flag), *other* threads can enter ssleay_add_bytes without obeying
6484 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6485 that they do not hold after the first thread unsets add_do_not_lock).
6488 *) Change bctest again: '-x' expressions are not available in all
6492 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6494 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6495 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6497 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6498 the default extension for executables, if any. Also, make the perl
6499 scripts that use symlink() to test if it really exists and use "cp"
6500 if it doesn't. All this made OpenSSL compilable and installable in
6504 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6505 If SEQUENCE is length is indefinite just set c->slen to the total
6506 amount of data available.
6507 [Steve Henson, reported by shige@FreeBSD.org]
6508 [This change does not apply to 0.9.7.]
6510 *) Change bctest to avoid here-documents inside command substitution
6511 (workaround for FreeBSD /bin/sh bug).
6512 For compatibility with Ultrix, avoid shell functions (introduced
6513 in the bctest version that searches along $PATH).
6516 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6517 with des_encrypt() defined on some operating systems, like Solaris
6521 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6522 On the Importance of Eliminating Errors in Cryptographic
6523 Computations, J. Cryptology 14 (2001) 2, 101-119,
6524 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6527 *) MIPS assembler BIGNUM division bug fix.
6530 *) Disabled incorrect Alpha assembler code.
6533 *) Fix PKCS#7 decode routines so they correctly update the length
6534 after reading an EOC for the EXPLICIT tag.
6536 [This change does not apply to 0.9.7.]
6538 *) Fix bug in PKCS#12 key generation routines. This was triggered
6539 if a 3DES key was generated with a 0 initial byte. Include
6540 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6541 (but broken) behaviour.
6544 *) Enhance bctest to search for a working bc along $PATH and print
6546 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6548 *) Fix memory leaks in err.c: free err_data string if necessary;
6549 don't write to the wrong index in ERR_set_error_data.
6552 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6556 *) Replace rdtsc with _emit statements for VC++ version 5.
6557 [Jeremy Cooper <jeremy@baymoo.org>]
6559 *) Make it possible to reuse SSLv2 sessions.
6562 *) In copy_email() check for >= 0 as a return value for
6563 X509_NAME_get_index_by_NID() since 0 is a valid index.
6564 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6566 *) Avoid coredump with unsupported or invalid public keys by checking if
6567 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6568 PKCS7_verify() fails with non detached data.
6571 *) Don't use getenv in library functions when run as setuid/setgid.
6572 New function OPENSSL_issetugid().
6575 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6576 due to incorrect handling of multi-threading:
6578 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6580 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6582 3. Count how many times MemCheck_off() has been called so that
6583 nested use can be treated correctly. This also avoids
6584 inband-signalling in the previous code (which relied on the
6585 assumption that thread ID 0 is impossible).
6588 *) Add "-rand" option also to s_client and s_server.
6591 *) Fix CPU detection on Irix 6.x.
6592 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6593 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6595 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6598 [This change does not apply to 0.9.7.]
6600 *) Use the cached encoding of an X509_NAME structure rather than
6601 copying it. This is apparently the reason for the libsafe "errors"
6602 but the code is actually correct.
6605 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6606 Bleichenbacher's DSA attack.
6607 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6608 to be set and top=0 forces the highest bit to be set; top=-1 is new
6609 and leaves the highest bit random.
6610 [Ulf Moeller, Bodo Moeller]
6612 *) In the NCONF_...-based implementations for CONF_... queries
6613 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6614 a temporary CONF structure with the data component set to NULL
6615 (which gives segmentation faults in lh_retrieve).
6616 Instead, use NULL for the CONF pointer in CONF_get_string and
6617 CONF_get_number (which may use environment variables) and directly
6618 return NULL from CONF_get_section.
6621 *) Fix potential buffer overrun for EBCDIC.
6624 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6625 keyUsage if basicConstraints absent for a CA.
6628 *) Make SMIME_write_PKCS7() write mail header values with a format that
6629 is more generally accepted (no spaces before the semicolon), since
6630 some programs can't parse those values properly otherwise. Also make
6631 sure BIO's that break lines after each write do not create invalid
6635 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6636 macros previously used would not encode an empty SEQUENCE OF
6637 and break the signature.
6639 [This change does not apply to 0.9.7.]
6641 *) Zero the premaster secret after deriving the master secret in
6645 *) Add some EVP_add_digest_alias registrations (as found in
6646 OpenSSL_add_all_digests()) to SSL_library_init()
6647 aka OpenSSL_add_ssl_algorithms(). This provides improved
6648 compatibility with peers using X.509 certificates
6649 with unconventional AlgorithmIdentifier OIDs.
6652 *) Fix for Irix with NO_ASM.
6653 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6655 *) ./config script fixes.
6656 [Ulf Moeller, Richard Levitte]
6658 *) Fix 'openssl passwd -1'.
6661 *) Change PKCS12_key_gen_asc() so it can cope with non null
6662 terminated strings whose length is passed in the passlen
6663 parameter, for example from PEM callbacks. This was done
6664 by adding an extra length parameter to asc2uni().
6665 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6667 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6668 call failed, free the DSA structure.
6671 *) Fix to uni2asc() to cope with zero length Unicode strings.
6672 These are present in some PKCS#12 files.
6675 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6676 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6677 when writing a 32767 byte record.
6678 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6680 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6681 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6683 (RSA objects have a reference count access to which is protected
6684 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6685 so they are meant to be shared between threads.)
6686 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6687 "Reddie, Steven" <Steven.Reddie@ca.com>]
6689 *) Fix a deadlock in CRYPTO_mem_leaks().
6692 *) Use better test patterns in bntest.
6695 *) rand_win.c fix for Borland C.
6698 *) BN_rshift bugfix for n == 0.
6701 *) Add a 'bctest' script that checks for some known 'bc' bugs
6702 so that 'make test' does not abort just because 'bc' is broken.
6705 *) Store verify_result within SSL_SESSION also for client side to
6706 avoid potential security hole. (Re-used sessions on the client side
6707 always resulted in verify_result==X509_V_OK, not using the original
6708 result of the server certificate verification.)
6711 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6712 SSL3_RT_APPLICATION_DATA, return 0.
6713 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6717 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6718 releases, have been re-implemented by renaming the previous
6719 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6720 and ssl3_read_internal, respectively, and adding 'peek' parameters
6721 to them. The new ssl[23]_{read,peek} functions are calls to
6722 ssl[23]_read_internal with the 'peek' flag set appropriately.
6723 A 'peek' parameter has also been added to ssl3_read_bytes, which
6724 does the actual work for ssl3_read_internal.
6727 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6728 the method-specific "init()" handler. Also clean up ex_data after
6729 calling the method-specific "finish()" handler. Previously, this was
6730 happening the other way round.
6733 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6734 The previous value, 12, was not always sufficient for BN_mod_exp().
6737 *) Make sure that shared libraries get the internal name engine with
6738 the full version number and not just 0. This should mark the
6739 shared libraries as not backward compatible. Of course, this should
6740 be changed again when we can guarantee backward binary compatibility.
6743 *) Fix typo in get_cert_by_subject() in by_dir.c
6744 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6746 *) Rework the system to generate shared libraries:
6748 - Make note of the expected extension for the shared libraries and
6749 if there is a need for symbolic links from for example libcrypto.so.0
6750 to libcrypto.so.0.9.7. There is extended info in Configure for
6753 - Make as few rebuilds of the shared libraries as possible.
6755 - Still avoid linking the OpenSSL programs with the shared libraries.
6757 - When installing, install the shared libraries separately from the
6761 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6763 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6764 and not in SSL_clear because the latter is also used by the
6765 accept/connect functions; previously, the settings made by
6766 SSL_set_read_ahead would be lost during the handshake.
6767 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6769 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6770 Previously, it would create entries for disableed algorithms no
6774 *) Added several new manual pages for SSL_* function.
6777 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6779 *) In ssl23_get_client_hello, generate an error message when faced
6780 with an initial SSL 3.0/TLS record that is too small to contain the
6781 first two bytes of the ClientHello message, i.e. client_version.
6782 (Note that this is a pathologic case that probably has never happened
6783 in real life.) The previous approach was to use the version number
6784 from the record header as a substitute; but our protocol choice
6785 should not depend on that one because it is not authenticated
6786 by the Finished messages.
6789 *) More robust randomness gathering functions for Windows.
6790 [Jeffrey Altman <jaltman@columbia.edu>]
6792 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6793 not set then we don't setup the error code for issuer check errors
6794 to avoid possibly overwriting other errors which the callback does
6795 handle. If an application does set the flag then we assume it knows
6796 what it is doing and can handle the new informational codes
6800 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6801 a general "ANY" type, as such it should be able to decode anything
6802 including tagged types. However it didn't check the class so it would
6803 wrongly interpret tagged types in the same way as their universal
6804 counterpart and unknown types were just rejected. Changed so that the
6805 tagged and unknown types are handled in the same way as a SEQUENCE:
6806 that is the encoding is stored intact. There is also a new type
6807 "V_ASN1_OTHER" which is used when the class is not universal, in this
6808 case we have no idea what the actual type is so we just lump them all
6812 *) On VMS, stdout may very well lead to a file that is written to
6813 in a record-oriented fashion. That means that every write() will
6814 write a separate record, which will be read separately by the
6815 programs trying to read from it. This can be very confusing.
6817 The solution is to put a BIO filter in the way that will buffer
6818 text until a linefeed is reached, and then write everything a
6819 line at a time, so every record written will be an actual line,
6820 not chunks of lines and not (usually doesn't happen, but I've
6821 seen it once) several lines in one record. BIO_f_linebuffer() is
6824 Currently, it's a VMS-only method, because that's where it has
6825 been tested well enough.
6828 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6829 it can return incorrect results.
6830 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6831 but it was in 0.9.6-beta[12].)
6834 *) Disable the check for content being present when verifying detached
6835 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6836 include zero length content when signing messages.
6839 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6840 BIO_ctrl (for BIO pairs).
6843 *) Add DSO method for VMS.
6846 *) Bug fix: Montgomery multiplication could produce results with the
6850 *) Add RPM specification openssl.spec and modify it to build three
6851 packages. The default package contains applications, application
6852 documentation and run-time libraries. The devel package contains
6853 include files, static libraries and function documentation. The
6854 doc package contains the contents of the doc directory. The original
6855 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6858 *) Add a large number of documentation files for many SSL routines.
6859 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6861 *) Add a configuration entry for Sony News 4.
6862 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6864 *) Don't set the two most significant bits to one when generating a
6865 random number < q in the DSA library.
6868 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6869 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6870 the underlying transport is blocking) if a handshake took place.
6871 (The default behaviour is needed by applications such as s_client
6872 and s_server that use select() to determine when to use SSL_read;
6873 but for applications that know in advance when to expect data, it
6874 just makes things more complicated.)
6877 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
6881 *) Add a few more EBCDIC conditionals that make `req' and `x509'
6882 work better on such systems.
6883 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6885 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
6886 Update PKCS12_parse() so it copies the friendlyName and the
6887 keyid to the certificates aux info.
6890 *) Fix bug in PKCS7_verify() which caused an infinite loop
6891 if there was more than one signature.
6892 [Sven Uszpelkat <su@celocom.de>]
6894 *) Major change in util/mkdef.pl to include extra information
6895 about each symbol, as well as presentig variables as well
6896 as functions. This change means that there's n more need
6897 to rebuild the .num files when some algorithms are excluded.
6900 *) Allow the verify time to be set by an application,
6901 rather than always using the current time.
6904 *) Phase 2 verify code reorganisation. The certificate
6905 verify code now looks up an issuer certificate by a
6906 number of criteria: subject name, authority key id
6907 and key usage. It also verifies self signed certificates
6908 by the same criteria. The main comparison function is
6909 X509_check_issued() which performs these checks.
6911 Lot of changes were necessary in order to support this
6912 without completely rewriting the lookup code.
6914 Authority and subject key identifier are now cached.
6916 The LHASH 'certs' is X509_STORE has now been replaced
6917 by a STACK_OF(X509_OBJECT). This is mainly because an
6918 LHASH can't store or retrieve multiple objects with
6919 the same hash value.
6921 As a result various functions (which were all internal
6922 use only) have changed to handle the new X509_STORE
6923 structure. This will break anything that messed round
6924 with X509_STORE internally.
6926 The functions X509_STORE_add_cert() now checks for an
6927 exact match, rather than just subject name.
6929 The X509_STORE API doesn't directly support the retrieval
6930 of multiple certificates matching a given criteria, however
6931 this can be worked round by performing a lookup first
6932 (which will fill the cache with candidate certificates)
6933 and then examining the cache for matches. This is probably
6934 the best we can do without throwing out X509_LOOKUP
6935 entirely (maybe later...).
6937 The X509_VERIFY_CTX structure has been enhanced considerably.
6939 All certificate lookup operations now go via a get_issuer()
6940 callback. Although this currently uses an X509_STORE it
6941 can be replaced by custom lookups. This is a simple way
6942 to bypass the X509_STORE hackery necessary to make this
6943 work and makes it possible to use more efficient techniques
6944 in future. A very simple version which uses a simple
6945 STACK for its trusted certificate store is also provided
6946 using X509_STORE_CTX_trusted_stack().
6948 The verify_cb() and verify() callbacks now have equivalents
6949 in the X509_STORE_CTX structure.
6951 X509_STORE_CTX also has a 'flags' field which can be used
6952 to customise the verify behaviour.
6955 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
6956 excludes S/MIME capabilities.
6959 *) When a certificate request is read in keep a copy of the
6960 original encoding of the signed data and use it when outputing
6961 again. Signatures then use the original encoding rather than
6962 a decoded, encoded version which may cause problems if the
6963 request is improperly encoded.
6966 *) For consistency with other BIO_puts implementations, call
6967 buffer_write(b, ...) directly in buffer_puts instead of calling
6970 In BIO_puts, increment b->num_write as in BIO_write.
6971 [Peter.Sylvester@EdelWeb.fr]
6973 *) Fix BN_mul_word for the case where the word is 0. (We have to use
6974 BN_zero, we may not return a BIGNUM with an array consisting of
6978 *) Avoid calling abort() from within the library when problems are
6979 detected, except if preprocessor symbols have been defined
6980 (such as REF_CHECK, BN_DEBUG etc.).
6983 *) New openssl application 'rsautl'. This utility can be
6984 used for low level RSA operations. DER public key
6985 BIO/fp routines also added.
6988 *) New Configure entry and patches for compiling on QNX 4.
6989 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
6991 *) A demo state-machine implementation was sponsored by
6992 Nuron (http://www.nuron.com/) and is now available in
6993 demos/state_machine.
6996 *) New options added to the 'dgst' utility for signature
6997 generation and verification.
7000 *) Unrecognized PKCS#7 content types are now handled via a
7001 catch all ASN1_TYPE structure. This allows unsupported
7002 types to be stored as a "blob" and an application can
7003 encode and decode it manually.
7006 *) Fix various signed/unsigned issues to make a_strex.c
7008 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7010 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7011 length if passed a buffer. ASN1_INTEGER_to_BN failed
7012 if passed a NULL BN and its argument was negative.
7013 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7015 *) Modification to PKCS#7 encoding routines to output definite
7016 length encoding. Since currently the whole structures are in
7017 memory there's not real point in using indefinite length
7018 constructed encoding. However if OpenSSL is compiled with
7019 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7022 *) Added BIO_vprintf() and BIO_vsnprintf().
7025 *) Added more prefixes to parse for in the the strings written
7026 through a logging bio, to cover all the levels that are available
7027 through syslog. The prefixes are now:
7029 PANIC, EMERG, EMR => LOG_EMERG
7030 ALERT, ALR => LOG_ALERT
7031 CRIT, CRI => LOG_CRIT
7032 ERROR, ERR => LOG_ERR
7033 WARNING, WARN, WAR => LOG_WARNING
7034 NOTICE, NOTE, NOT => LOG_NOTICE
7035 INFO, INF => LOG_INFO
7036 DEBUG, DBG => LOG_DEBUG
7038 and as before, if none of those prefixes are present at the
7039 beginning of the string, LOG_ERR is chosen.
7041 On Win32, the LOG_* levels are mapped according to this:
7043 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7044 LOG_WARNING => EVENTLOG_WARNING_TYPE
7045 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7049 *) Made it possible to reconfigure with just the configuration
7050 argument "reconf" or "reconfigure". The command line arguments
7051 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7052 and are retrieved from there when reconfiguring.
7056 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7058 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7061 *) The obj_dat.pl script was messing up the sorting of object
7062 names. The reason was that it compared the quoted version
7063 of strings as a result "OCSP" > "OCSP Signing" because
7064 " > SPACE. Changed script to store unquoted versions of
7065 names and add quotes on output. It was also omitting some
7066 names from the lookup table if they were given a default
7067 value (that is if SN is missing it is given the same
7068 value as LN and vice versa), these are now added on the
7069 grounds that if an object has a name we should be able to
7070 look it up. Finally added warning output when duplicate
7071 short or long names are found.
7074 *) Changes needed for Tandem NSK.
7075 [Scott Uroff <scott@xypro.com>]
7077 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7078 RSA_padding_check_SSLv23(), special padding was never detected
7079 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7080 version rollback attacks was not effective.
7082 In s23_clnt.c, don't use special rollback-attack detection padding
7083 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7084 client; similarly, in s23_srvr.c, don't do the rollback check if
7085 SSL 2.0 is the only protocol enabled in the server.
7088 *) Make it possible to get hexdumps of unprintable data with 'openssl
7089 asn1parse'. By implication, the functions ASN1_parse_dump() and
7090 BIO_dump_indent() are added.
7093 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7094 these print out strings and name structures based on various
7095 flags including RFC2253 support and proper handling of
7096 multibyte characters. Added options to the 'x509' utility
7097 to allow the various flags to be set.
7100 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7101 Also change the functions X509_cmp_current_time() and
7102 X509_gmtime_adj() work with an ASN1_TIME structure,
7103 this will enable certificates using GeneralizedTime in validity
7104 dates to be checked.
7107 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7108 negative public key encodings) on by default,
7109 NO_NEG_PUBKEY_BUG can be set to disable it.
7112 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7113 content octets. An i2c_ASN1_OBJECT is unnecessary because
7114 the encoding can be trivially obtained from the structure.
7117 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7118 not read locks (CRYPTO_r_[un]lock).
7121 *) A first attempt at creating official support for shared
7122 libraries through configuration. I've kept it so the
7123 default is static libraries only, and the OpenSSL programs
7124 are always statically linked for now, but there are
7125 preparations for dynamic linking in place.
7126 This has been tested on Linux and Tru64.
7129 *) Randomness polling function for Win9x, as described in:
7130 Peter Gutmann, Software Generation of Practically Strong
7134 *) Fix so PRNG is seeded in req if using an already existing
7138 *) New options to smime application. -inform and -outform
7139 allow alternative formats for the S/MIME message including
7140 PEM and DER. The -content option allows the content to be
7141 specified separately. This should allow things like Netscape
7142 form signing output easier to verify.
7145 *) Fix the ASN1 encoding of tags using the 'long form'.
7148 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
7149 STRING types. These convert content octets to and from the
7150 underlying type. The actual tag and length octets are
7151 already assumed to have been read in and checked. These
7152 are needed because all other string types have virtually
7153 identical handling apart from the tag. By having versions
7154 of the ASN1 functions that just operate on content octets
7155 IMPLICIT tagging can be handled properly. It also allows
7156 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
7157 and ASN1_INTEGER are identical apart from the tag.
7160 *) Change the handling of OID objects as follows:
7162 - New object identifiers are inserted in objects.txt, following
7163 the syntax given in objects.README.
7164 - objects.pl is used to process obj_mac.num and create a new
7166 - obj_dat.pl is used to create a new obj_dat.h, using the data in
7169 This is currently kind of a hack, and the perl code in objects.pl
7170 isn't very elegant, but it works as I intended. The simplest way
7171 to check that it worked correctly is to look in obj_dat.h and
7172 check the array nid_objs and make sure the objects haven't moved
7173 around (this is important!). Additions are OK, as well as
7174 consistent name changes.
7177 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
7180 *) Addition of the command line parameter '-rand file' to 'openssl req'.
7181 The given file adds to whatever has already been seeded into the
7182 random pool through the RANDFILE configuration file option or
7183 environment variable, or the default random state file.
7186 *) mkstack.pl now sorts each macro group into lexical order.
7187 Previously the output order depended on the order the files
7188 appeared in the directory, resulting in needless rewriting
7192 *) Patches to make OpenSSL compile under Win32 again. Mostly
7193 work arounds for the VC++ problem that it treats func() as
7194 func(void). Also stripped out the parts of mkdef.pl that
7195 added extra typesafe functions: these no longer exist.
7198 *) Reorganisation of the stack code. The macros are now all
7199 collected in safestack.h . Each macro is defined in terms of
7200 a "stack macro" of the form SKM_<name>(type, a, b). The
7201 DEBUG_SAFESTACK is now handled in terms of function casts,
7202 this has the advantage of retaining type safety without the
7203 use of additional functions. If DEBUG_SAFESTACK is not defined
7204 then the non typesafe macros are used instead. Also modified the
7205 mkstack.pl script to handle the new form. Needs testing to see
7206 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
7207 the default if no major problems. Similar behaviour for ASN1_SET_OF
7208 and PKCS12_STACK_OF.
7211 *) When some versions of IIS use the 'NET' form of private key the
7212 key derivation algorithm is different. Normally MD5(password) is
7213 used as a 128 bit RC4 key. In the modified case
7214 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
7215 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
7216 as the old Netscape_RSA functions except they have an additional
7217 'sgckey' parameter which uses the modified algorithm. Also added
7218 an -sgckey command line option to the rsa utility. Thanks to
7219 Adrian Peck <bertie@ncipher.com> for posting details of the modified
7220 algorithm to openssl-dev.
7223 *) The evp_local.h macros were using 'c.##kname' which resulted in
7224 invalid expansion on some systems (SCO 5.0.5 for example).
7225 Corrected to 'c.kname'.
7226 [Phillip Porch <root@theporch.com>]
7228 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
7229 a STACK of email addresses from a certificate or request, these look
7230 in the subject name and the subject alternative name extensions and
7231 omit any duplicate addresses.
7234 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
7235 This makes DSA verification about 2 % faster.
7238 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
7239 (meaning that now 2^5 values will be precomputed, which is only 4 KB
7240 plus overhead for 1024 bit moduli).
7241 This makes exponentiations about 0.5 % faster for 1024 bit
7242 exponents (as measured by "openssl speed rsa2048").
7245 *) Rename memory handling macros to avoid conflicts with other
7247 Malloc => OPENSSL_malloc
7248 Malloc_locked => OPENSSL_malloc_locked
7249 Realloc => OPENSSL_realloc
7250 Free => OPENSSL_free
7253 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
7254 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
7257 *) CygWin32 support.
7258 [John Jarvie <jjarvie@newsguy.com>]
7260 *) The type-safe stack code has been rejigged. It is now only compiled
7261 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
7262 by default all type-specific stack functions are "#define"d back to
7263 standard stack functions. This results in more streamlined output
7264 but retains the type-safety checking possibilities of the original
7268 *) The STACK code has been cleaned up, and certain type declarations
7269 that didn't make a lot of sense have been brought in line. This has
7270 also involved a cleanup of sorts in safestack.h to more correctly
7271 map type-safe stack functions onto their plain stack counterparts.
7272 This work has also resulted in a variety of "const"ifications of
7273 lots of the code, especially "_cmp" operations which should normally
7274 be prototyped with "const" parameters anyway.
7277 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
7278 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
7279 (The PRNG state consists of two parts, the large pool 'state' and 'md',
7280 where all of 'md' is used each time the PRNG is used, but 'state'
7281 is used only indexed by a cyclic counter. As entropy may not be
7282 well distributed from the beginning, 'md' is important as a
7283 chaining variable. However, the output function chains only half
7284 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
7285 all of 'md', and seeding with STATE_SIZE dummy bytes will result
7286 in all of 'state' being rewritten, with the new values depending
7287 on virtually all of 'md'. This overcomes the 80 bit limitation.)
7290 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
7291 the handshake is continued after ssl_verify_cert_chain();
7292 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
7293 can lead to 'unexplainable' connection aborts later.
7294 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
7296 *) Major EVP API cipher revision.
7297 Add hooks for extra EVP features. This allows various cipher
7298 parameters to be set in the EVP interface. Support added for variable
7299 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
7300 setting of RC2 and RC5 parameters.
7302 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
7305 Remove lots of duplicated code from the EVP library. For example *every*
7306 cipher init() function handles the 'iv' in the same way according to the
7307 cipher mode. They also all do nothing if the 'key' parameter is NULL and
7308 for CFB and OFB modes they zero ctx->num.
7310 New functionality allows removal of S/MIME code RC2 hack.
7312 Most of the routines have the same form and so can be declared in terms
7315 By shifting this to the top level EVP_CipherInit() it can be removed from
7316 all individual ciphers. If the cipher wants to handle IVs or keys
7317 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7320 Change lots of functions like EVP_EncryptUpdate() to now return a
7321 value: although software versions of the algorithms cannot fail
7322 any installed hardware versions can.
7325 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7326 this option is set, tolerate broken clients that send the negotiated
7327 protocol version number instead of the requested protocol version
7331 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7332 i.e. non-zero for export ciphersuites, zero otherwise.
7333 Previous versions had this flag inverted, inconsistent with
7334 rsa_tmp_cb (..._TMP_RSA_CB).
7335 [Bodo Moeller; problem reported by Amit Chopra]
7337 *) Add missing DSA library text string. Work around for some IIS
7338 key files with invalid SEQUENCE encoding.
7341 *) Add a document (doc/standards.txt) that list all kinds of standards
7342 and so on that are implemented in OpenSSL.
7345 *) Enhance c_rehash script. Old version would mishandle certificates
7346 with the same subject name hash and wouldn't handle CRLs at all.
7347 Added -fingerprint option to crl utility, to support new c_rehash
7351 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7354 *) Fix for SSL server purpose checking. Server checking was
7355 rejecting certificates which had extended key usage present
7356 but no ssl client purpose.
7357 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7359 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7360 is a little unclear about how a blank password is handled.
7361 Since the password in encoded as a BMPString with terminating
7362 double NULL a zero length password would end up as just the
7363 double NULL. However no password at all is different and is
7364 handled differently in the PKCS#12 key generation code. NS
7365 treats a blank password as zero length. MSIE treats it as no
7366 password on export: but it will try both on import. We now do
7367 the same: PKCS12_parse() tries zero length and no password if
7368 the password is set to "" or NULL (NULL is now a valid password:
7369 it wasn't before) as does the pkcs12 application.
7372 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7373 perror when PEM_read_bio_X509_REQ fails, the error message must
7374 be obtained from the error queue.
7377 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7378 it in ERR_remove_state if appropriate, and change ERR_get_state
7379 accordingly to avoid race conditions (this is necessary because
7380 thread_hash is no longer constant once set).
7383 *) Bugfix for linux-elf makefile.one.
7386 *) RSA_get_default_method() will now cause a default
7387 RSA_METHOD to be chosen if one doesn't exist already.
7388 Previously this was only set during a call to RSA_new()
7389 or RSA_new_method(NULL) meaning it was possible for
7390 RSA_get_default_method() to return NULL.
7393 *) Added native name translation to the existing DSO code
7394 that will convert (if the flag to do so is set) filenames
7395 that are sufficiently small and have no path information
7396 into a canonical native form. Eg. "blah" converted to
7397 "libblah.so" or "blah.dll" etc.
7400 *) New function ERR_error_string_n(e, buf, len) which is like
7401 ERR_error_string(e, buf), but writes at most 'len' bytes
7402 including the 0 terminator. For ERR_error_string_n, 'buf'
7404 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7406 *) CONF library reworked to become more general. A new CONF
7407 configuration file reader "class" is implemented as well as a
7408 new functions (NCONF_*, for "New CONF") to handle it. The now
7409 old CONF_* functions are still there, but are reimplemented to
7410 work in terms of the new functions. Also, a set of functions
7411 to handle the internal storage of the configuration data is
7412 provided to make it easier to write new configuration file
7413 reader "classes" (I can definitely see something reading a
7414 configuration file in XML format, for example), called _CONF_*,
7415 or "the configuration storage API"...
7417 The new configuration file reading functions are:
7419 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7420 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7422 NCONF_default, NCONF_WIN32
7424 NCONF_dump_fp, NCONF_dump_bio
7426 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7427 NCONF_new creates a new CONF object. This works in the same way
7428 as other interfaces in OpenSSL, like the BIO interface.
7429 NCONF_dump_* dump the internal storage of the configuration file,
7430 which is useful for debugging. All other functions take the same
7431 arguments as the old CONF_* functions wth the exception of the
7432 first that must be a `CONF *' instead of a `LHASH *'.
7434 To make it easer to use the new classes with the old CONF_* functions,
7435 the function CONF_set_default_method is provided.
7438 *) Add '-tls1' option to 'openssl ciphers', which was already
7439 mentioned in the documentation but had not been implemented.
7440 (This option is not yet really useful because even the additional
7441 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7444 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7445 OpenSSL-based applications) load shared libraries and bind to
7446 them in a portable way.
7447 [Geoff Thorpe, with contributions from Richard Levitte]
7449 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7451 *) Make sure _lrotl and _lrotr are only used with MSVC.
7453 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7454 (the default implementation of RAND_status).
7456 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7457 to '-clrext' (= clear extensions), as intended and documented.
7458 [Bodo Moeller; inconsistency pointed out by Michael Attili
7461 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7462 was larger than the MD block size.
7463 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7465 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7466 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7467 using the passed key: if the passed key was a private key the result
7468 of X509_print(), for example, would be to print out all the private key
7472 *) des_quad_cksum() byte order bug fix.
7473 [Ulf Möller, using the problem description in krb4-0.9.7, where
7474 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7476 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7478 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7480 *) For easily testing in shell scripts whether some command
7481 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7482 returns with exit code 0 iff no command of the given name is available.
7483 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7484 the output goes to stdout and nothing is printed to stderr.
7485 Additional arguments are always ignored.
7487 Since for each cipher there is a command of the same name,
7488 the 'no-cipher' compilation switches can be tested this way.
7490 ('openssl no-XXX' is not able to detect pseudo-commands such
7491 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7494 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7497 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7498 is set; it will be thrown away anyway because each handshake creates
7500 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7501 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7502 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7503 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7506 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7507 'Q' and 'R' lose their special meanings (quit/renegotiate).
7508 This is part of what -quiet does; unlike -quiet, -ign_eof
7509 does not suppress any output.
7512 *) Add compatibility options to the purpose and trust code. The
7513 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7514 accepts a certificate or CA, this was the previous behaviour,
7515 with all the associated security issues.
7517 X509_TRUST_COMPAT is the old trust behaviour: only and
7518 automatically trust self signed roots in certificate store. A
7519 new trust setting X509_TRUST_DEFAULT is used to specify that
7520 a purpose has no associated trust setting and it should instead
7521 use the value in the default purpose.
7524 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7525 and fix a memory leak.
7528 *) In util/mkerr.pl (which implements 'make errors'), preserve
7529 reason strings from the previous version of the .c file, as
7530 the default to have only downcase letters (and digits) in
7531 automatically generated reasons codes is not always appropriate.
7534 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7535 using strerror. Previously, ERR_reason_error_string() returned
7536 library names as reason strings for SYSerr; but SYSerr is a special
7537 case where small numbers are errno values, not library numbers.
7540 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7541 converts DSA parameters into DH parameters. (When creating parameters,
7542 DSA_generate_parameters is used.)
7545 *) Include 'length' (recommended exponent length) in C code generated
7546 by 'openssl dhparam -C'.
7549 *) The second argument to set_label in perlasm was already being used
7550 so couldn't be used as a "file scope" flag. Moved to third argument
7554 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7555 instead of RAND_bytes for encryption IVs and salts.
7558 *) Include RAND_status() into RAND_METHOD instead of implementing
7559 it only for md_rand.c Otherwise replacing the PRNG by calling
7560 RAND_set_rand_method would be impossible.
7563 *) Don't let DSA_generate_key() enter an infinite loop if the random
7564 number generation fails.
7567 *) New 'rand' application for creating pseudo-random output.
7570 *) Added configuration support for Linux/IA64
7571 [Rolf Haberrecker <rolf@suse.de>]
7573 *) Assembler module support for Mingw32.
7576 *) Shared library support for HPUX (in shlib/).
7577 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7579 *) Shared library support for Solaris gcc.
7580 [Lutz Behnke <behnke@trustcenter.de>]
7582 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7584 *) PKCS7_encrypt() was adding text MIME headers twice because they
7585 were added manually and by SMIME_crlf_copy().
7588 *) In bntest.c don't call BN_rand with zero bits argument.
7589 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7591 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7592 case was implemented. This caused BN_div_recp() to fail occasionally.
7595 *) Add an optional second argument to the set_label() in the perl
7596 assembly language builder. If this argument exists and is set
7597 to 1 it signals that the assembler should use a symbol whose
7598 scope is the entire file, not just the current function. This
7599 is needed with MASM which uses the format label:: for this scope.
7600 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7602 *) Change the ASN1 types so they are typedefs by default. Before
7603 almost all types were #define'd to ASN1_STRING which was causing
7604 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7608 *) Change names of new functions to the new get1/get0 naming
7609 convention: After 'get1', the caller owns a reference count
7610 and has to call ..._free; 'get0' returns a pointer to some
7611 data structure without incrementing reference counters.
7612 (Some of the existing 'get' functions increment a reference
7613 counter, some don't.)
7614 Similarly, 'set1' and 'add1' functions increase reference
7615 counters or duplicate objects.
7618 *) Allow for the possibility of temp RSA key generation failure:
7619 the code used to assume it always worked and crashed on failure.
7622 *) Fix potential buffer overrun problem in BIO_printf().
7623 [Ulf Möller, using public domain code by Patrick Powell; problem
7624 pointed out by David Sacerdote <das33@cornell.edu>]
7626 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7627 RAND_egd() and RAND_status(). In the command line application,
7628 the EGD socket can be specified like a seed file using RANDFILE
7632 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7633 Some CAs (e.g. Verisign) distribute certificates in this form.
7636 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7637 list to exclude them. This means that no special compilation option
7638 is needed to use anonymous DH: it just needs to be included in the
7642 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7643 EVP_MD_type. The old functionality is available in a new macro called
7644 EVP_MD_md(). Change code that uses it and update docs.
7647 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7648 where the 'void *' argument is replaced by a function pointer argument.
7649 Previously 'void *' was abused to point to functions, which works on
7650 many platforms, but is not correct. As these functions are usually
7651 called by macros defined in OpenSSL header files, most source code
7652 should work without changes.
7655 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7656 sections with information on -D... compiler switches used for
7657 compiling the library so that applications can see them. To enable
7658 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7659 must be defined. E.g.,
7660 #define OPENSSL_ALGORITHM_DEFINES
7661 #include <openssl/opensslconf.h>
7662 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7663 [Richard Levitte, Ulf and Bodo Möller]
7665 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7669 *) Change the 'other' type in certificate aux info to a STACK_OF
7670 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7671 the required ASN1 format: arbitrary types determined by an OID.
7674 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7675 argument to 'req'. This is not because the function is newer or
7676 better than others it just uses the work 'NEW' in the certificate
7677 request header lines. Some software needs this.
7680 *) Reorganise password command line arguments: now passwords can be
7681 obtained from various sources. Delete the PEM_cb function and make
7682 it the default behaviour: i.e. if the callback is NULL and the
7683 usrdata argument is not NULL interpret it as a null terminated pass
7684 phrase. If usrdata and the callback are NULL then the pass phrase
7685 is prompted for as usual.
7688 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7689 the support is automatically enabled. The resulting binaries will
7690 autodetect the card and use it if present.
7691 [Ben Laurie and Compaq Inc.]
7693 *) Work around for Netscape hang bug. This sends certificate request
7694 and server done in one record. Since this is perfectly legal in the
7695 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7696 the bugs/SSLv3 entry for more info.
7699 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7702 *) Add -rand argument to smime and pkcs12 applications and read/write
7706 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7709 *) Add command line password options to the remaining applications.
7712 *) Bug fix for BN_div_recp() for numerators with an even number of
7716 *) More tests in bntest.c, and changed test_bn output.
7719 *) ./config recognizes MacOS X now.
7722 *) Bug fix for BN_div() when the first words of num and divsor are
7723 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7726 *) Add support for various broken PKCS#8 formats, and command line
7727 options to produce them.
7730 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7731 get temporary BIGNUMs from a BN_CTX.
7734 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7738 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7739 include a #define from the old name to the new. The original intent
7740 was that statically linked binaries could for example just call
7741 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7742 link with digests. This never worked becayse SSLeay_add_all_digests()
7743 and SSLeay_add_all_ciphers() were in the same source file so calling
7744 one would link with the other. They are now in separate source files.
7747 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7750 *) Use a less unusual form of the Miller-Rabin primality test (it used
7751 a binary algorithm for exponentiation integrated into the Miller-Rabin
7752 loop, our standard modexp algorithms are faster).
7755 *) Support for the EBCDIC character set completed.
7756 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7758 *) Source code cleanups: use const where appropriate, eliminate casts,
7759 use void * instead of char * in lhash.
7762 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7763 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7764 this the server could overwrite ephemeral keys that the client
7768 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7769 using 50 iterations of the Rabin-Miller test.
7771 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7772 iterations of the Rabin-Miller test as required by the appendix
7773 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7774 As BN_is_prime_fasttest includes trial division, DSA parameter
7775 generation becomes much faster.
7777 This implies a change for the callback functions in DSA_is_prime
7778 and DSA_generate_parameters: The callback function is called once
7779 for each positive witness in the Rabin-Miller test, not just
7780 occasionally in the inner loop; and the parameters to the
7781 callback function now provide an iteration count for the outer
7782 loop rather than for the current invocation of the inner loop.
7783 DSA_generate_parameters additionally can call the callback
7784 function with an 'iteration count' of -1, meaning that a
7785 candidate has passed the trial division test (when q is generated
7786 from an application-provided seed, trial division is skipped).
7789 *) New function BN_is_prime_fasttest that optionally does trial
7790 division before starting the Rabin-Miller test and has
7791 an additional BN_CTX * argument (whereas BN_is_prime always
7792 has to allocate at least one BN_CTX).
7793 'callback(1, -1, cb_arg)' is called when a number has passed the
7794 trial division stage.
7797 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7801 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7804 *) New function BN_pseudo_rand().
7807 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7808 bignum version of BN_from_montgomery() with the working code from
7809 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7813 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7814 made it impossible to use the same SSL_SESSION data structure in
7815 SSL2 clients in multiple threads.
7818 *) The return value of RAND_load_file() no longer counts bytes obtained
7819 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7820 to seed the PRNG (previously an explicit byte count was required).
7821 [Ulf Möller, Bodo Möller]
7823 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7824 used (char *) instead of (void *) and had casts all over the place.
7827 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7830 *) Retain source code compatibility for BN_prime_checks macro:
7831 BN_is_prime(..., BN_prime_checks, ...) now uses
7832 BN_prime_checks_for_size to determine the appropriate number of
7833 Rabin-Miller iterations.
7836 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7837 DH_CHECK_P_NOT_SAFE_PRIME.
7838 (Check if this is true? OpenPGP calls them "strong".)
7841 *) Merge the functionality of "dh" and "gendh" programs into a new program
7842 "dhparam". The old programs are retained for now but will handle DH keys
7843 (instead of parameters) in future.
7846 *) Make the ciphers, s_server and s_client programs check the return values
7847 when a new cipher list is set.
7850 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7851 ciphers. Before when the 56bit ciphers were enabled the sorting was
7854 The syntax for the cipher sorting has been extended to support sorting by
7855 cipher-strength (using the strength_bits hard coded in the tables).
7856 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7858 Fix a bug in the cipher-command parser: when supplying a cipher command
7859 string with an "undefined" symbol (neither command nor alphanumeric
7860 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7861 an error is flagged.
7863 Due to the strength-sorting extension, the code of the
7864 ssl_create_cipher_list() function was completely rearranged. I hope that
7865 the readability was also increased :-)
7866 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7868 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7869 for the first serial number and places 2 in the serial number file. This
7870 avoids problems when the root CA is created with serial number zero and
7871 the first user certificate has the same issuer name and serial number
7875 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7876 the new code. Add documentation for this stuff.
7879 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
7880 X509_*() to X509at_*() on the grounds that they don't handle X509
7881 structures and behave in an analagous way to the X509v3 functions:
7882 they shouldn't be called directly but wrapper functions should be used
7885 So we also now have some wrapper functions that call the X509at functions
7886 when passed certificate requests. (TO DO: similar things can be done with
7887 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
7888 things. Some of these need some d2i or i2d and print functionality
7889 because they handle more complex structures.)
7892 *) Add missing #ifndefs that caused missing symbols when building libssl
7893 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
7894 NO_RSA in ssl/s2*.c.
7895 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
7897 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
7898 has a return value which indicates the quality of the random data
7899 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
7900 error queue. New function RAND_pseudo_bytes() generates output that is
7901 guaranteed to be unique but not unpredictable. RAND_add is like
7902 RAND_seed, but takes an extra argument for an entropy estimate
7903 (RAND_seed always assumes full entropy).
7906 *) Do more iterations of Rabin-Miller probable prime test (specifically,
7907 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
7908 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
7909 in crypto/bn/bn_prime.c for the complete table). This guarantees a
7910 false-positive rate of at most 2^-80 for random input.
7913 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
7916 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
7917 in the 0.9.5 release), this returns the chain
7918 from an X509_CTX structure with a dup of the stack and all
7919 the X509 reference counts upped: so the stack will exist
7920 after X509_CTX_cleanup() has been called. Modify pkcs12.c
7923 Also make SSL_SESSION_print() print out the verify return
7927 *) Add manpage for the pkcs12 command. Also change the default
7928 behaviour so MAC iteration counts are used unless the new
7929 -nomaciter option is used. This improves file security and
7930 only older versions of MSIE (4.0 for example) need it.
7933 *) Honor the no-xxx Configure options when creating .DEF files.
7936 *) Add PKCS#10 attributes to field table: challengePassword,
7937 unstructuredName and unstructuredAddress. These are taken from
7938 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
7939 international characters are used.
7941 More changes to X509_ATTRIBUTE code: allow the setting of types
7942 based on strings. Remove the 'loc' parameter when adding
7943 attributes because these will be a SET OF encoding which is sorted
7947 *) Initial changes to the 'req' utility to allow request generation
7948 automation. This will allow an application to just generate a template
7949 file containing all the field values and have req construct the
7952 Initial support for X509_ATTRIBUTE handling. Stacks of these are
7953 used all over the place including certificate requests and PKCS#7
7954 structures. They are currently handled manually where necessary with
7955 some primitive wrappers for PKCS#7. The new functions behave in a
7956 manner analogous to the X509 extension functions: they allow
7957 attributes to be looked up by NID and added.
7959 Later something similar to the X509V3 code would be desirable to
7960 automatically handle the encoding, decoding and printing of the
7961 more complex types. The string types like challengePassword can
7962 be handled by the string table functions.
7964 Also modified the multi byte string table handling. Now there is
7965 a 'global mask' which masks out certain types. The table itself
7966 can use the flag STABLE_NO_MASK to ignore the mask setting: this
7967 is useful when for example there is only one permissible type
7968 (as in countryName) and using the mask might result in no valid
7972 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
7973 SSL_get_peer_finished to allow applications to obtain the latest
7974 Finished messages sent to the peer or expected from the peer,
7975 respectively. (SSL_get_peer_finished is usually the Finished message
7976 actually received from the peer, otherwise the protocol will be aborted.)
7978 As the Finished message are message digests of the complete handshake
7979 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
7980 be used for external authentication procedures when the authentication
7981 provided by SSL/TLS is not desired or is not enough.
7984 *) Enhanced support for Alpha Linux is added. Now ./config checks if
7985 the host supports BWX extension and if Compaq C is present on the
7986 $PATH. Just exploiting of the BWX extension results in 20-30%
7987 performance kick for some algorithms, e.g. DES and RC4 to mention
7988 a couple. Compaq C in turn generates ~20% faster code for MD5 and
7992 *) Add support for MS "fast SGC". This is arguably a violation of the
7993 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
7994 weak crypto and after checking the certificate is SGC a second one
7995 with strong crypto. MS SGC stops the first handshake after receiving
7996 the server certificate message and sends a second client hello. Since
7997 a server will typically do all the time consuming operations before
7998 expecting any further messages from the client (server key exchange
7999 is the most expensive) there is little difference between the two.
8001 To get OpenSSL to support MS SGC we have to permit a second client
8002 hello message after we have sent server done. In addition we have to
8003 reset the MAC if we do get this second client hello.
8006 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
8007 if a DER encoded private key is RSA or DSA traditional format. Changed
8008 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
8009 format DER encoded private key. Newer code should use PKCS#8 format which
8010 has the key type encoded in the ASN1 structure. Added DER private key
8011 support to pkcs8 application.
8014 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8015 ciphersuites has been selected (as required by the SSL 3/TLS 1
8016 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8017 is set, we interpret this as a request to violate the specification
8018 (the worst that can happen is a handshake failure, and 'correct'
8019 behaviour would result in a handshake failure anyway).
8022 *) In SSL_CTX_add_session, take into account that there might be multiple
8023 SSL_SESSION structures with the same session ID (e.g. when two threads
8024 concurrently obtain them from an external cache).
8025 The internal cache can handle only one SSL_SESSION with a given ID,
8026 so if there's a conflict, we now throw out the old one to achieve
8030 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8031 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8032 some routines that use cipher OIDs: some ciphers do not have OIDs
8033 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8037 *) Simplify the trust setting structure and code. Now we just have
8038 two sequences of OIDs for trusted and rejected settings. These will
8039 typically have values the same as the extended key usage extension
8040 and any application specific purposes.
8042 The trust checking code now has a default behaviour: it will just
8043 check for an object with the same NID as the passed id. Functions can
8044 be provided to override either the default behaviour or the behaviour
8045 for a given id. SSL client, server and email already have functions
8046 in place for compatibility: they check the NID and also return "trusted"
8047 if the certificate is self signed.
8050 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8051 traditional format into an EVP_PKEY structure.
8054 *) Add a password callback function PEM_cb() which either prompts for
8055 a password if usr_data is NULL or otherwise assumes it is a null
8056 terminated password. Allow passwords to be passed on command line
8057 environment or config files in a few more utilities.
8060 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8061 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8062 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8063 Update documentation.
8066 *) Support for ASN1 "NULL" type. This could be handled before by using
8067 ASN1_TYPE but there wasn't any function that would try to read a NULL
8068 and produce an error if it couldn't. For compatibility we also have
8069 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8070 don't allocate anything because they don't need to.
8073 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8075 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8077 *) Rebuild of the memory allocation routines used by OpenSSL code and
8078 possibly others as well. The purpose is to make an interface that
8079 provide hooks so anyone can build a separate set of allocation and
8080 deallocation routines to be used by OpenSSL, for example memory
8081 pool implementations, or something else, which was previously hard
8082 since Malloc(), Realloc() and Free() were defined as macros having
8083 the values malloc, realloc and free, respectively (except for Win32
8084 compilations). The same is provided for memory debugging code.
8085 OpenSSL already comes with functionality to find memory leaks, but
8086 this gives people a chance to debug other memory problems.
8088 With these changes, a new set of functions and macros have appeared:
8090 CRYPTO_set_mem_debug_functions() [F]
8091 CRYPTO_get_mem_debug_functions() [F]
8092 CRYPTO_dbg_set_options() [F]
8093 CRYPTO_dbg_get_options() [F]
8094 CRYPTO_malloc_debug_init() [M]
8096 The memory debug functions are NULL by default, unless the library
8097 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8098 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8099 gives the standard debugging functions that come with OpenSSL) or
8100 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8101 provided by the library user) must be used. When the standard
8102 debugging functions are used, CRYPTO_dbg_set_options can be used to
8103 request additional information:
8104 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8105 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8107 Also, things like CRYPTO_set_mem_functions will always give the
8108 expected result (the new set of functions is used for allocation
8109 and deallocation) at all times, regardless of platform and compiler
8112 To finish it up, some functions that were never use in any other
8113 way than through macros have a new API and new semantic:
8116 CRYPTO_dbg_realloc()
8119 All macros of value have retained their old syntax.
8120 [Richard Levitte and Bodo Moeller]
8122 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8123 ordering of SMIMECapabilities wasn't in "strength order" and there
8124 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8128 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8129 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8130 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
8132 *) Merge in my S/MIME library for OpenSSL. This provides a simple
8133 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
8134 functionality to handle multipart/signed properly) and a utility
8135 called 'smime' to call all this stuff. This is based on code I
8136 originally wrote for Celo who have kindly allowed it to be
8137 included in OpenSSL.
8140 *) Add variants des_set_key_checked and des_set_key_unchecked of
8141 des_set_key (aka des_key_sched). Global variable des_check_key
8142 decides which of these is called by des_set_key; this way
8143 des_check_key behaves as it always did, but applications and
8144 the library itself, which was buggy for des_check_key == 1,
8145 have a cleaner way to pick the version they need.
8148 *) New function PKCS12_newpass() which changes the password of a
8152 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
8153 dynamic mix. In both cases the ids can be used as an index into the
8154 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
8155 functions so they accept a list of the field values and the
8156 application doesn't need to directly manipulate the X509_TRUST
8160 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
8164 *) Modify the way the V3 extension code looks up extensions. This now
8165 works in a similar way to the object code: we have some "standard"
8166 extensions in a static table which is searched with OBJ_bsearch()
8167 and the application can add dynamic ones if needed. The file
8168 crypto/x509v3/ext_dat.h now has the info: this file needs to be
8169 updated whenever a new extension is added to the core code and kept
8170 in ext_nid order. There is a simple program 'tabtest.c' which checks
8171 this. New extensions are not added too often so this file can readily
8172 be maintained manually.
8174 There are two big advantages in doing things this way. The extensions
8175 can be looked up immediately and no longer need to be "added" using
8176 X509V3_add_standard_extensions(): this function now does nothing.
8177 [Side note: I get *lots* of email saying the extension code doesn't
8178 work because people forget to call this function]
8179 Also no dynamic allocation is done unless new extensions are added:
8180 so if we don't add custom extensions there is no need to call
8181 X509V3_EXT_cleanup().
8184 *) Modify enc utility's salting as follows: make salting the default. Add a
8185 magic header, so unsalted files fail gracefully instead of just decrypting
8186 to garbage. This is because not salting is a big security hole, so people
8187 should be discouraged from doing it.
8190 *) Fixes and enhancements to the 'x509' utility. It allowed a message
8191 digest to be passed on the command line but it only used this
8192 parameter when signing a certificate. Modified so all relevant
8193 operations are affected by the digest parameter including the
8194 -fingerprint and -x509toreq options. Also -x509toreq choked if a
8195 DSA key was used because it didn't fix the digest.
8198 *) Initial certificate chain verify code. Currently tests the untrusted
8199 certificates for consistency with the verify purpose (which is set
8200 when the X509_STORE_CTX structure is set up) and checks the pathlength.
8202 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
8203 this is because it will reject chains with invalid extensions whereas
8204 every previous version of OpenSSL and SSLeay made no checks at all.
8206 Trust code: checks the root CA for the relevant trust settings. Trust
8207 settings have an initial value consistent with the verify purpose: e.g.
8208 if the verify purpose is for SSL client use it expects the CA to be
8209 trusted for SSL client use. However the default value can be changed to
8210 permit custom trust settings: one example of this would be to only trust
8211 certificates from a specific "secure" set of CAs.
8213 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
8214 which should be used for version portability: especially since the
8215 verify structure is likely to change more often now.
8217 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
8218 to set them. If not set then assume SSL clients will verify SSL servers
8221 Two new options to the verify program: -untrusted allows a set of
8222 untrusted certificates to be passed in and -purpose which sets the
8223 intended purpose of the certificate. If a purpose is set then the
8224 new chain verify code is used to check extension consistency.
8227 *) Support for the authority information access extension.
8230 *) Modify RSA and DSA PEM read routines to transparently handle
8231 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
8232 public keys in a format compatible with certificate
8233 SubjectPublicKeyInfo structures. Unfortunately there were already
8234 functions called *_PublicKey_* which used various odd formats so
8235 these are retained for compatibility: however the DSA variants were
8236 never in a public release so they have been deleted. Changed dsa/rsa
8237 utilities to handle the new format: note no releases ever handled public
8238 keys so we should be OK.
8240 The primary motivation for this change is to avoid the same fiasco
8241 that dogs private keys: there are several incompatible private key
8242 formats some of which are standard and some OpenSSL specific and
8243 require various evil hacks to allow partial transparent handling and
8244 even then it doesn't work with DER formats. Given the option anything
8245 other than PKCS#8 should be dumped: but the other formats have to
8246 stay in the name of compatibility.
8248 With public keys and the benefit of hindsight one standard format
8249 is used which works with EVP_PKEY, RSA or DSA structures: though
8250 it clearly returns an error if you try to read the wrong kind of key.
8252 Added a -pubkey option to the 'x509' utility to output the public key.
8253 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
8254 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
8255 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
8256 that do the same as the EVP_PKEY_assign_*() except they up the
8257 reference count of the added key (they don't "swallow" the
8261 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
8262 CRLs would fail if the file contained no certificates or no CRLs:
8263 added a new function to read in both types and return the number
8264 read: this means that if none are read it will be an error. The
8265 DER versions of the certificate and CRL reader would always fail
8266 because it isn't possible to mix certificates and CRLs in DER format
8267 without choking one or the other routine. Changed this to just read
8268 a certificate: this is the best we can do. Also modified the code
8269 in apps/verify.c to take notice of return codes: it was previously
8270 attempting to read in certificates from NULL pointers and ignoring
8271 any errors: this is one reason why the cert and CRL reader seemed
8272 to work. It doesn't check return codes from the default certificate
8273 routines: these may well fail if the certificates aren't installed.
8276 *) Code to support otherName option in GeneralName.
8279 *) First update to verify code. Change the verify utility
8280 so it warns if it is passed a self signed certificate:
8281 for consistency with the normal behaviour. X509_verify
8282 has been modified to it will now verify a self signed
8283 certificate if *exactly* the same certificate appears
8284 in the store: it was previously impossible to trust a
8285 single self signed certificate. This means that:
8286 openssl verify ss.pem
8287 now gives a warning about a self signed certificate but
8288 openssl verify -CAfile ss.pem ss.pem
8292 *) For servers, store verify_result in SSL_SESSION data structure
8293 (and add it to external session representation).
8294 This is needed when client certificate verifications fails,
8295 but an application-provided verification callback (set by
8296 SSL_CTX_set_cert_verify_callback) allows accepting the session
8297 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
8298 but returns 1): When the session is reused, we have to set
8299 ssl->verify_result to the appropriate error code to avoid
8301 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
8303 *) Fix a bug in the new PKCS#7 code: it didn't consider the
8304 case in PKCS7_dataInit() where the signed PKCS7 structure
8305 didn't contain any existing data because it was being created.
8306 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
8308 *) Add a salt to the key derivation routines in enc.c. This
8309 forms the first 8 bytes of the encrypted file. Also add a
8310 -S option to allow a salt to be input on the command line.
8313 *) New function X509_cmp(). Oddly enough there wasn't a function
8314 to compare two certificates. We do this by working out the SHA1
8315 hash and comparing that. X509_cmp() will be needed by the trust
8319 *) SSL_get1_session() is like SSL_get_session(), but increments
8320 the reference count in the SSL_SESSION returned.
8321 [Geoff Thorpe <geoff@eu.c2.net>]
8323 *) Fix for 'req': it was adding a null to request attributes.
8324 Also change the X509_LOOKUP and X509_INFO code to handle
8325 certificate auxiliary information.
8328 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8332 *) Add the possibility to add extra information to the memory leak
8333 detecting output, to form tracebacks, showing from where each
8334 allocation was originated: CRYPTO_push_info("constant string") adds
8335 the string plus current file name and line number to a per-thread
8336 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8337 is like calling CYRPTO_pop_info() until the stack is empty.
8338 Also updated memory leak detection code to be multi-thread-safe.
8341 *) Add options -text and -noout to pkcs7 utility and delete the
8342 encryption options which never did anything. Update docs.
8345 *) Add options to some of the utilities to allow the pass phrase
8346 to be included on either the command line (not recommended on
8347 OSes like Unix) or read from the environment. Update the
8348 manpages and fix a few bugs.
8351 *) Add a few manpages for some of the openssl commands.
8354 *) Fix the -revoke option in ca. It was freeing up memory twice,
8355 leaking and not finding already revoked certificates.
8358 *) Extensive changes to support certificate auxiliary information.
8359 This involves the use of X509_CERT_AUX structure and X509_AUX
8360 functions. An X509_AUX function such as PEM_read_X509_AUX()
8361 can still read in a certificate file in the usual way but it
8362 will also read in any additional "auxiliary information". By
8363 doing things this way a fair degree of compatibility can be
8364 retained: existing certificates can have this information added
8365 using the new 'x509' options.
8367 Current auxiliary information includes an "alias" and some trust
8368 settings. The trust settings will ultimately be used in enhanced
8369 certificate chain verification routines: currently a certificate
8370 can only be trusted if it is self signed and then it is trusted
8374 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8375 The problem was that one of the replacement routines had not been working
8376 since SSLeay releases. For now the offending routine has been replaced
8377 with non-optimised assembler. Even so, this now gives around 95%
8378 performance improvement for 1024 bit RSA signs.
8381 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8382 handling. Most clients have the effective key size in bits equal to
8383 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8384 A few however don't do this and instead use the size of the decrypted key
8385 to determine the RC2 key length and the AlgorithmIdentifier to determine
8386 the effective key length. In this case the effective key length can still
8387 be 40 bits but the key length can be 168 bits for example. This is fixed
8388 by manually forcing an RC2 key into the EVP_PKEY structure because the
8389 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8390 the key length and effective key length are equal.
8393 *) Add a bunch of functions that should simplify the creation of
8394 X509_NAME structures. Now you should be able to do:
8395 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8396 and have it automatically work out the correct field type and fill in
8397 the structures. The more adventurous can try:
8398 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8399 and it will (hopefully) work out the correct multibyte encoding.
8402 *) Change the 'req' utility to use the new field handling and multibyte
8403 copy routines. Before the DN field creation was handled in an ad hoc
8404 way in req, ca, and x509 which was rather broken and didn't support
8405 BMPStrings or UTF8Strings. Since some software doesn't implement
8406 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8407 using the dirstring_type option. See the new comment in the default
8408 openssl.cnf for more info.
8411 *) Make crypto/rand/md_rand.c more robust:
8412 - Assure unique random numbers after fork().
8413 - Make sure that concurrent threads access the global counter and
8414 md serializably so that we never lose entropy in them
8415 or use exactly the same state in multiple threads.
8416 Access to the large state is not always serializable because
8417 the additional locking could be a performance killer, and
8418 md should be large enough anyway.
8421 *) New file apps/app_rand.c with commonly needed functionality
8422 for handling the random seed file.
8424 Use the random seed file in some applications that previously did not:
8426 dsaparam -genkey (which also ignored its '-rand' option),
8429 x509 (when signing).
8430 Except on systems with /dev/urandom, it is crucial to have a random
8431 seed file at least for key creation, DSA signing, and for DH exchanges;
8432 for RSA signatures we could do without one.
8434 gendh and gendsa (unlike genrsa) used to read only the first byte
8435 of each file listed in the '-rand' option. The function as previously
8436 found in genrsa is now in app_rand.c and is used by all programs
8437 that support '-rand'.
8440 *) In RAND_write_file, use mode 0600 for creating files;
8441 don't just chmod when it may be too late.
8444 *) Report an error from X509_STORE_load_locations
8445 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8448 *) New function ASN1_mbstring_copy() this copies a string in either
8449 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8450 into an ASN1_STRING type. A mask of permissible types is passed
8451 and it chooses the "minimal" type to use or an error if not type
8455 *) Add function equivalents to the various macros in asn1.h. The old
8456 macros are retained with an M_ prefix. Code inside the library can
8457 use the M_ macros. External code (including the openssl utility)
8458 should *NOT* in order to be "shared library friendly".
8461 *) Add various functions that can check a certificate's extensions
8462 to see if it usable for various purposes such as SSL client,
8463 server or S/MIME and CAs of these types. This is currently
8464 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8465 verification. Also added a -purpose flag to x509 utility to
8466 print out all the purposes.
8469 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8473 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8474 for, obtain and decode and extension and obtain its critical flag.
8475 This allows all the necessary extension code to be handled in a
8476 single function call.
8479 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8480 platforms. See crypto/rc4/rc4_enc.c for further details.
8483 *) New -noout option to asn1parse. This causes no output to be produced
8484 its main use is when combined with -strparse and -out to extract data
8485 from a file (which may not be in ASN.1 format).
8488 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8489 when producing the local key id.
8490 [Richard Levitte <levitte@stacken.kth.se>]
8492 *) New option -dhparam in s_server. This allows a DH parameter file to be
8493 stated explicitly. If it is not stated then it tries the first server
8494 certificate file. The previous behaviour hard coded the filename
8498 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8499 a public key to be input or output. For example:
8500 openssl rsa -in key.pem -pubout -out pubkey.pem
8501 Also added necessary DSA public key functions to handle this.
8504 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8505 in the message. This was handled by allowing
8506 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8507 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8509 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8510 to the end of the strings whereas this didn't. This would cause problems
8511 if strings read with d2i_ASN1_bytes() were later modified.
8512 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8514 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8515 data and it contains EOF it will end up returning an error. This is
8516 caused by input 46 bytes long. The cause is due to the way base64
8517 BIOs find the start of base64 encoded data. They do this by trying a
8518 trial decode on each line until they find one that works. When they
8519 do a flag is set and it starts again knowing it can pass all the
8520 data directly through the decoder. Unfortunately it doesn't reset
8521 the context it uses. This means that if EOF is reached an attempt
8522 is made to pass two EOFs through the context and this causes the
8523 resulting error. This can also cause other problems as well. As is
8524 usual with these problems it takes *ages* to find and the fix is
8525 trivial: move one line.
8526 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8528 *) Ugly workaround to get s_client and s_server working under Windows. The
8529 old code wouldn't work because it needed to select() on sockets and the
8530 tty (for keypresses and to see if data could be written). Win32 only
8531 supports select() on sockets so we select() with a 1s timeout on the
8532 sockets and then see if any characters are waiting to be read, if none
8533 are present then we retry, we also assume we can always write data to
8534 the tty. This isn't nice because the code then blocks until we've
8535 received a complete line of data and it is effectively polling the
8536 keyboard at 1s intervals: however it's quite a bit better than not
8537 working at all :-) A dedicated Windows application might handle this
8538 with an event loop for example.
8541 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8542 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8543 will be called when RSA_sign() and RSA_verify() are used. This is useful
8544 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8545 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8546 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8547 This necessitated the support of an extra signature type NID_md5_sha1
8548 for SSL signatures and modifications to the SSL library to use it instead
8549 of calling RSA_public_decrypt() and RSA_private_encrypt().
8552 *) Add new -verify -CAfile and -CApath options to the crl program, these
8553 will lookup a CRL issuers certificate and verify the signature in a
8554 similar way to the verify program. Tidy up the crl program so it
8555 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8556 less strict. It will now permit CRL extensions even if it is not
8557 a V2 CRL: this will allow it to tolerate some broken CRLs.
8560 *) Initialize all non-automatic variables each time one of the openssl
8561 sub-programs is started (this is necessary as they may be started
8562 multiple times from the "OpenSSL>" prompt).
8563 [Lennart Bang, Bodo Moeller]
8565 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8566 removing all other RSA functionality (this is what NO_RSA does). This
8567 is so (for example) those in the US can disable those operations covered
8568 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8572 *) Non-copying interface to BIO pairs.
8573 (still largely untested)
8576 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8577 ASCII string. This was handled independently in various places before.
8580 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8581 UTF8 strings a character at a time.
8584 *) Use client_version from client hello to select the protocol
8585 (s23_srvr.c) and for RSA client key exchange verification
8586 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8589 *) Add various utility functions to handle SPKACs, these were previously
8590 handled by poking round in the structure internals. Added new function
8591 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8592 print, verify and generate SPKACs. Based on an original idea from
8593 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8596 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8599 *) Allow the config file extension section to be overwritten on the
8600 command line. Based on an original idea from Massimiliano Pala
8601 <madwolf@comune.modena.it>. The new option is called -extensions
8602 and can be applied to ca, req and x509. Also -reqexts to override
8603 the request extensions in req and -crlexts to override the crl extensions
8607 *) Add new feature to the SPKAC handling in ca. Now you can include
8608 the same field multiple times by preceding it by "XXXX." for example:
8611 this is the same syntax as used in the req config file.
8614 *) Allow certificate extensions to be added to certificate requests. These
8615 are specified in a 'req_extensions' option of the req section of the
8616 config file. They can be printed out with the -text option to req but
8617 are otherwise ignored at present.
8620 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8621 data read consists of only the final block it would not decrypted because
8622 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8623 A misplaced 'break' also meant the decrypted final block might not be
8624 copied until the next read.
8627 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8628 a few extra parameters to the DH structure: these will be useful if
8629 for example we want the value of 'q' or implement X9.42 DH.
8632 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8633 provides hooks that allow the default DSA functions or functions on a
8634 "per key" basis to be replaced. This allows hardware acceleration and
8635 hardware key storage to be handled without major modification to the
8636 library. Also added low level modexp hooks and CRYPTO_EX structure and
8637 associated functions.
8640 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8641 as "read only": it can't be written to and the buffer it points to will
8642 not be freed. Reading from a read only BIO is much more efficient than
8643 a normal memory BIO. This was added because there are several times when
8644 an area of memory needs to be read from a BIO. The previous method was
8645 to create a memory BIO and write the data to it, this results in two
8646 copies of the data and an O(n^2) reading algorithm. There is a new
8647 function BIO_new_mem_buf() which creates a read only memory BIO from
8648 an area of memory. Also modified the PKCS#7 routines to use read only
8652 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8653 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8654 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8655 but a retry condition occured while trying to read the rest.
8658 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8659 NID_pkcs7_encrypted by default: this was wrong since this should almost
8660 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8661 the encrypted data type: this is a more sensible place to put it and it
8662 allows the PKCS#12 code to be tidied up that duplicated this
8666 *) Changed obj_dat.pl script so it takes its input and output files on
8667 the command line. This should avoid shell escape redirection problems
8671 *) Initial support for certificate extension requests, these are included
8672 in things like Xenroll certificate requests. Included functions to allow
8673 extensions to be obtained and added.
8676 *) -crlf option to s_client and s_server for sending newlines as
8677 CRLF (as required by many protocols).
8680 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8682 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8683 [Ralf S. Engelschall]
8685 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8686 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8688 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8692 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8693 DH parameters/keys (q is lost during that conversion, but the resulting
8694 DH parameters contain its length).
8696 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8697 much faster than DH_generate_parameters (which creates parameters
8698 where p = 2*q + 1), and also the smaller q makes DH computations
8699 much more efficient (160-bit exponentiation instead of 1024-bit
8700 exponentiation); so this provides a convenient way to support DHE
8701 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8702 utter importance to use
8703 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8705 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8706 when such DH parameters are used, because otherwise small subgroup
8707 attacks may become possible!
8710 *) Avoid memory leak in i2d_DHparams.
8713 *) Allow the -k option to be used more than once in the enc program:
8714 this allows the same encrypted message to be read by multiple recipients.
8717 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8718 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8719 it will always use the numerical form of the OID, even if it has a short
8723 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8724 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8725 otherwise bn_mod_exp was called. In the case of hardware keys for example
8726 no private key components need be present and it might store extra data
8727 in the RSA structure, which cannot be accessed from bn_mod_exp.
8728 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8729 private key operations.
8732 *) Added support for SPARC Linux.
8735 *) pem_password_cb function type incompatibly changed from
8736 typedef int pem_password_cb(char *buf, int size, int rwflag);
8738 ....(char *buf, int size, int rwflag, void *userdata);
8739 so that applications can pass data to their callbacks:
8740 The PEM[_ASN1]_{read,write}... functions and macros now take an
8741 additional void * argument, which is just handed through whenever
8742 the password callback is called.
8743 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8745 New function SSL_CTX_set_default_passwd_cb_userdata.
8747 Compatibility note: As many C implementations push function arguments
8748 onto the stack in reverse order, the new library version is likely to
8749 interoperate with programs that have been compiled with the old
8750 pem_password_cb definition (PEM_whatever takes some data that
8751 happens to be on the stack as its last argument, and the callback
8752 just ignores this garbage); but there is no guarantee whatsoever that
8755 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8756 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8757 problems not only on Windows, but also on some Unix platforms.
8758 To avoid problematic command lines, these definitions are now in an
8759 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8760 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8763 *) MIPS III/IV assembler module is reimplemented.
8766 *) More DES library cleanups: remove references to srand/rand and
8767 delete an unused file.
8770 *) Add support for the the free Netwide assembler (NASM) under Win32,
8771 since not many people have MASM (ml) and it can be hard to obtain.
8772 This is currently experimental but it seems to work OK and pass all
8773 the tests. Check out INSTALL.W32 for info.
8776 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8777 without temporary keys kept an extra copy of the server key,
8778 and connections with temporary keys did not free everything in case
8782 *) New function RSA_check_key and new openssl rsa option -check
8783 for verifying the consistency of RSA keys.
8784 [Ulf Moeller, Bodo Moeller]
8786 *) Various changes to make Win32 compile work:
8787 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8788 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8789 comparison" warnings.
8790 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8793 *) Add a debugging option to PKCS#5 v2 key generation function: when
8794 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8795 derived keys are printed to stderr.
8798 *) Copy the flags in ASN1_STRING_dup().
8799 [Roman E. Pavlov <pre@mo.msk.ru>]
8801 *) The x509 application mishandled signing requests containing DSA
8802 keys when the signing key was also DSA and the parameters didn't match.
8804 It was supposed to omit the parameters when they matched the signing key:
8805 the verifying software was then supposed to automatically use the CA's
8806 parameters if they were absent from the end user certificate.
8808 Omitting parameters is no longer recommended. The test was also
8809 the wrong way round! This was probably due to unusual behaviour in
8810 EVP_cmp_parameters() which returns 1 if the parameters match.
8811 This meant that parameters were omitted when they *didn't* match and
8812 the certificate was useless. Certificates signed with 'ca' didn't have
8814 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8816 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8817 The interface is as follows:
8818 Applications can use
8819 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8820 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8821 "off" is now the default.
8822 The library internally uses
8823 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8824 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8825 to disable memory-checking temporarily.
8827 Some inconsistent states that previously were possible (and were
8828 even the default) are now avoided.
8830 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8831 with each memory chunk allocated; this is occasionally more helpful
8832 than just having a counter.
8834 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8836 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8840 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8841 which largely parallels "options", but is for changing API behaviour,
8842 whereas "options" are about protocol behaviour.
8843 Initial "mode" flags are:
8845 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8846 a single record has been written.
8847 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8848 retries use the same buffer location.
8849 (But all of the contents must be
8853 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8856 *) Fix problems with no-hmac etc.
8857 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8859 *) New functions RSA_get_default_method(), RSA_set_method() and
8860 RSA_get_method(). These allows replacement of RSA_METHODs without having
8861 to mess around with the internals of an RSA structure.
8864 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8865 Also really enable memory leak checks in openssl.c and in some
8867 [Chad C. Mulligan, Bodo Moeller]
8869 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8870 up the length of negative integers. This has now been simplified to just
8871 store the length when it is first determined and use it later, rather
8872 than trying to keep track of where data is copied and updating it to
8874 [Steve Henson, reported by Brien Wheeler
8875 <bwheeler@authentica-security.com>]
8877 *) Add a new function PKCS7_signatureVerify. This allows the verification
8878 of a PKCS#7 signature but with the signing certificate passed to the
8879 function itself. This contrasts with PKCS7_dataVerify which assumes the
8880 certificate is present in the PKCS#7 structure. This isn't always the
8881 case: certificates can be omitted from a PKCS#7 structure and be
8882 distributed by "out of band" means (such as a certificate database).
8885 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
8886 function prototypes in pem.h, also change util/mkdef.pl to add the
8887 necessary function names.
8890 *) mk1mf.pl (used by Windows builds) did not properly read the
8891 options set by Configure in the top level Makefile, and Configure
8892 was not even able to write more than one option correctly.
8893 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
8896 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
8897 file to be loaded from a BIO or FILE pointer. The BIO version will
8898 for example allow memory BIOs to contain config info.
8901 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
8902 Whoever hopes to achieve shared-library compatibility across versions
8903 must use this, not the compile-time macro.
8904 (Exercise 0.9.4: Which is the minimum library version required by
8906 Note: All this applies only to multi-threaded programs, others don't
8910 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
8911 through a BIO pair triggered the default case, i.e.
8912 SSLerr(...,SSL_R_UNKNOWN_STATE).
8915 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
8916 can use the SSL library even if none of the specific BIOs is
8920 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
8921 for the encoded length.
8922 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
8924 *) Add initial documentation of the X509V3 functions.
8927 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
8928 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
8929 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
8930 secure PKCS#8 private key format with a high iteration count.
8933 *) Fix determination of Perl interpreter: A perl or perl5
8934 _directory_ in $PATH was also accepted as the interpreter.
8935 [Ralf S. Engelschall]
8937 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
8938 wrong with it but it was very old and did things like calling
8939 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
8943 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
8944 to use the new extension code.
8947 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
8948 with macros. This should make it easier to change their form, add extra
8949 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
8953 *) Add to configuration table a new entry that can specify an alternative
8954 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
8955 according to Mark Crispin <MRC@Panda.COM>.
8959 *) DES CBC did not update the IV. Weird.
8962 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
8963 Changing the behaviour of the former might break existing programs --
8964 where IV updating is needed, des_ncbc_encrypt can be used.
8967 *) When bntest is run from "make test" it drives bc to check its
8968 calculations, as well as internally checking them. If an internal check
8969 fails, it needs to cause bc to give a non-zero result or make test carries
8970 on without noticing the failure. Fixed.
8973 *) DES library cleanups.
8976 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
8977 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
8978 ciphers. NOTE: although the key derivation function has been verified
8979 against some published test vectors it has not been extensively tested
8980 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
8984 *) Instead of "mkdir -p", which is not fully portable, use new
8985 Perl script "util/mkdir-p.pl".
8988 *) Rewrite the way password based encryption (PBE) is handled. It used to
8989 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
8990 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
8991 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
8992 the 'parameter' field of the AlgorithmIdentifier is passed to the
8993 underlying key generation function so it must do its own ASN1 parsing.
8994 This has also changed the EVP_PBE_CipherInit() function which now has a
8995 'parameter' argument instead of literal salt and iteration count values
8996 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
8999 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
9000 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
9001 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
9002 KEY" because this clashed with PKCS#8 unencrypted string. Since this
9003 value was just used as a "magic string" and not used directly its
9004 value doesn't matter.
9007 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9011 *) "linux-sparc64" configuration (ultrapenguin).
9012 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9013 "linux-sparc" configuration.
9014 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9016 *) config now generates no-xxx options for missing ciphers.
9019 *) Support the EBCDIC character set (work in progress).
9020 File ebcdic.c not yet included because it has a different license.
9021 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9023 *) Support BS2000/OSD-POSIX.
9024 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9026 *) Make callbacks for key generation use void * instead of char *.
9029 *) Make S/MIME samples compile (not yet tested).
9032 *) Additional typesafe stacks.
9035 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9039 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9041 *) New configuration variant "sco5-gcc".
9043 *) Updated some demos.
9044 [Sean O Riordain, Wade Scholine]
9046 *) Add missing BIO_free at exit of pkcs12 application.
9049 *) Fix memory leak in conf.c.
9052 *) Updates for Win32 to assembler version of MD5.
9055 *) Set #! path to perl in apps/der_chop to where we found it
9056 instead of using a fixed path.
9059 *) SHA library changes for irix64-mips4-cc.
9062 *) Improvements for VMS support.
9066 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9068 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9069 This also avoids the problems with SC4.2 and unpatched SC5.
9070 [Andy Polyakov <appro@fy.chalmers.se>]
9072 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9073 These are required because of the typesafe stack would otherwise break
9074 existing code. If old code used a structure member which used to be STACK
9075 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9076 sk_num or sk_value it would produce an error because the num, data members
9077 are not present in STACK_OF. Now it just produces a warning. sk_set
9078 replaces the old method of assigning a value to sk_value
9079 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9080 that does this will no longer work (and should use sk_set instead) but
9081 this could be regarded as a "questionable" behaviour anyway.
9084 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9085 correctly handle encrypted S/MIME data.
9088 *) Change type of various DES function arguments from des_cblock
9089 (which means, in function argument declarations, pointer to char)
9090 to des_cblock * (meaning pointer to array with 8 char elements),
9091 which allows the compiler to do more typechecking; it was like
9092 that back in SSLeay, but with lots of ugly casts.
9094 Introduce new type const_des_cblock.
9097 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9098 problems: find RecipientInfo structure that matches recipient certificate
9099 and initialise the ASN1 structures properly based on passed cipher.
9102 *) Belatedly make the BN tests actually check the results.
9105 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9106 to and from BNs: it was completely broken. New compilation option
9107 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9108 key elements as negative integers.
9111 *) Reorganize and speed up MD5.
9112 [Andy Polyakov <appro@fy.chalmers.se>]
9115 [Richard Levitte <richard@levitte.org>]
9117 *) New option -out to asn1parse to allow the parsed structure to be
9118 output to a file. This is most useful when combined with the -strparse
9119 option to examine the output of things like OCTET STRINGS.
9122 *) Make SSL library a little more fool-proof by not requiring any longer
9123 that SSL_set_{accept,connect}_state be called before
9124 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9125 in many applications because usually everything *appeared* to work as
9126 intended anyway -- now it really works as intended).
9129 *) Move openssl.cnf out of lib/.
9132 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
9133 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
9134 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
9135 [Ralf S. Engelschall]
9137 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
9138 handle PKCS#7 enveloped data properly.
9139 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
9141 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
9142 copying pointers. The cert_st handling is changed by this in
9143 various ways (and thus what used to be known as ctx->default_cert
9144 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
9145 any longer when s->cert does not give us what we need).
9146 ssl_cert_instantiate becomes obsolete by this change.
9147 As soon as we've got the new code right (possibly it already is?),
9148 we have solved a couple of bugs of the earlier code where s->cert
9149 was used as if it could not have been shared with other SSL structures.
9151 Note that using the SSL API in certain dirty ways now will result
9152 in different behaviour than observed with earlier library versions:
9153 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
9154 does not influence s as it used to.
9156 In order to clean up things more thoroughly, inside SSL_SESSION
9157 we don't use CERT any longer, but a new structure SESS_CERT
9158 that holds per-session data (if available); currently, this is
9159 the peer's certificate chain and, for clients, the server's certificate
9160 and temporary key. CERT holds only those values that can have
9161 meaningful defaults in an SSL_CTX.
9164 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
9165 from the internal representation. Various PKCS#7 fixes: remove some
9166 evil casts and set the enc_dig_alg field properly based on the signing
9170 *) Allow PKCS#12 password to be set from the command line or the
9171 environment. Let 'ca' get its config file name from the environment
9172 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
9176 *) Allow certificate policies extension to use an IA5STRING for the
9177 organization field. This is contrary to the PKIX definition but
9178 VeriSign uses it and IE5 only recognises this form. Document 'x509'
9182 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
9183 without disallowing inline assembler and the like for non-pedantic builds.
9186 *) Support Borland C++ builder.
9187 [Janez Jere <jj@void.si>, modified by Ulf Möller]
9192 *) SHA-1 cleanups and performance enhancements.
9193 [Andy Polyakov <appro@fy.chalmers.se>]
9195 *) Sparc v8plus assembler for the bignum library.
9196 [Andy Polyakov <appro@fy.chalmers.se>]
9198 *) Accept any -xxx and +xxx compiler options in Configure.
9201 *) Update HPUX configuration.
9204 *) Add missing sk_<type>_unshift() function to safestack.h
9205 [Ralf S. Engelschall]
9207 *) New function SSL_CTX_use_certificate_chain_file that sets the
9208 "extra_cert"s in addition to the certificate. (This makes sense
9209 only for "PEM" format files, as chains as a whole are not
9213 *) Support verify_depth from the SSL API.
9214 x509_vfy.c had what can be considered an off-by-one-error:
9215 Its depth (which was not part of the external interface)
9216 was actually counting the number of certificates in a chain;
9217 now it really counts the depth.
9220 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
9221 instead of X509err, which often resulted in confusing error
9222 messages since the error codes are not globally unique
9223 (e.g. an alleged error in ssl3_accept when a certificate
9224 didn't match the private key).
9226 *) New function SSL_CTX_set_session_id_context that allows to set a default
9227 value (so that you don't need SSL_set_session_id_context for each
9228 connection using the SSL_CTX).
9231 *) OAEP decoding bug fix.
9234 *) Support INSTALL_PREFIX for package builders, as proposed by
9238 *) New Configure options "threads" and "no-threads". For systems
9239 where the proper compiler options are known (currently Solaris
9240 and Linux), "threads" is the default.
9243 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
9246 *) Install various scripts to $(OPENSSLDIR)/misc, not to
9247 $(INSTALLTOP)/bin -- they shouldn't clutter directories
9248 such as /usr/local/bin.
9251 *) "make linux-shared" to build shared libraries.
9252 [Niels Poppe <niels@netbox.org>]
9254 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
9257 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
9258 extension adding in x509 utility.
9261 *) Remove NOPROTO sections and error code comments.
9264 *) Partial rewrite of the DEF file generator to now parse the ANSI
9268 *) New Configure options --prefix=DIR and --openssldir=DIR.
9271 *) Complete rewrite of the error code script(s). It is all now handled
9272 by one script at the top level which handles error code gathering,
9273 header rewriting and C source file generation. It should be much better
9274 than the old method: it now uses a modified version of Ulf's parser to
9275 read the ANSI prototypes in all header files (thus the old K&R definitions
9276 aren't needed for error creation any more) and do a better job of
9277 translating function codes into names. The old 'ASN1 error code imbedded
9278 in a comment' is no longer necessary and it doesn't use .err files which
9279 have now been deleted. Also the error code call doesn't have to appear all
9280 on one line (which resulted in some large lines...).
9283 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
9286 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
9287 0 (which usually indicates a closed connection), but continue reading.
9290 *) Fix some race conditions.
9293 *) Add support for CRL distribution points extension. Add Certificate
9294 Policies and CRL distribution points documentation.
9297 *) Move the autogenerated header file parts to crypto/opensslconf.h.
9300 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
9301 8 of keying material. Merlin has also confirmed interop with this fix
9302 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
9303 [Merlin Hughes <merlin@baltimore.ie>]
9305 *) Fix lots of warnings.
9306 [Richard Levitte <levitte@stacken.kth.se>]
9308 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
9309 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9310 [Richard Levitte <levitte@stacken.kth.se>]
9312 *) Fix problems with sizeof(long) == 8.
9313 [Andy Polyakov <appro@fy.chalmers.se>]
9315 *) Change functions to ANSI C.
9318 *) Fix typos in error codes.
9319 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9321 *) Remove defunct assembler files from Configure.
9324 *) SPARC v8 assembler BIGNUM implementation.
9325 [Andy Polyakov <appro@fy.chalmers.se>]
9327 *) Support for Certificate Policies extension: both print and set.
9328 Various additions to support the r2i method this uses.
9331 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9332 return a const string when you are expecting an allocated buffer.
9335 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9336 types DirectoryString and DisplayText.
9339 *) Add code to allow r2i extensions to access the configuration database,
9340 add an LHASH database driver and add several ctx helper functions.
9343 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9344 fail when they extended the size of a BIGNUM.
9347 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9348 support typesafe stack.
9351 *) Fix typo in SSL_[gs]et_options().
9352 [Nils Frostberg <nils@medcom.se>]
9354 *) Delete various functions and files that belonged to the (now obsolete)
9355 old X509V3 handling code.
9358 *) New Configure option "rsaref".
9361 *) Don't auto-generate pem.h.
9364 *) Introduce type-safe ASN.1 SETs.
9367 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9368 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9370 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9371 that links with OpenSSL (well at least cause lots of warnings), but fear
9372 not: the conversion is trivial, and it eliminates loads of evil casts. A
9373 few STACKed things have been converted already. Feel free to convert more.
9374 In the fullness of time, I'll do away with the STACK type altogether.
9377 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9378 specified in <certfile> by updating the entry in the index.txt file.
9379 This way one no longer has to edit the index.txt file manually for
9380 revoking a certificate. The -revoke option does the gory details now.
9381 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9383 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9384 `-text' option at all and this way the `-noout -text' combination was
9385 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9386 [Ralf S. Engelschall]
9388 *) Make sure a corresponding plain text error message exists for the
9389 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9390 verify callback function determined that a certificate was revoked.
9391 [Ralf S. Engelschall]
9393 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9394 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9395 all available cipers including rc5, which was forgotten until now.
9396 In order to let the testing shell script know which algorithms
9397 are available, a new (up to now undocumented) command
9398 "openssl list-cipher-commands" is used.
9401 *) Bugfix: s_client occasionally would sleep in select() when
9402 it should have checked SSL_pending() first.
9405 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9406 the raw DSA values prior to ASN.1 encoding.
9409 *) Tweaks to Configure
9410 [Niels Poppe <niels@netbox.org>]
9412 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9416 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9419 *) New config option to avoid instructions that are illegal on the 80386.
9420 The default code is faster, but requires at least a 486.
9423 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9424 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9425 same as SSL2_VERSION anyway.
9428 *) New "-showcerts" option for s_client.
9431 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9432 application. Various cleanups and fixes.
9435 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9436 modify error routines to work internally. Add error codes and PBE init
9437 to library startup routines.
9440 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9441 packing functions to asn1 and evp. Changed function names and error
9442 codes along the way.
9445 *) PKCS12 integration: and so it begins... First of several patches to
9446 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9447 objects to objects.h
9450 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9451 and display support for Thawte strong extranet extension.
9454 *) Add LinuxPPC support.
9455 [Jeff Dubrule <igor@pobox.org>]
9457 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9458 bn_div_words in alpha.s.
9459 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9461 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9462 OAEP isn't supported when OpenSSL is built with RSAref.
9463 [Ulf Moeller <ulf@fitug.de>]
9465 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9466 so they no longer are missing under -DNOPROTO.
9467 [Soren S. Jorvang <soren@t.dk>]
9470 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9472 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9473 doesn't work when the session is reused. Coming soon!
9476 *) Fix a security hole, that allows sessions to be reused in the wrong
9477 context thus bypassing client cert protection! All software that uses
9478 client certs and session caches in multiple contexts NEEDS PATCHING to
9479 allow session reuse! A fuller solution is in the works.
9480 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9482 *) Some more source tree cleanups (removed obsolete files
9483 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9484 permission on "config" script to be executable) and a fix for the INSTALL
9486 [Ulf Moeller <ulf@fitug.de>]
9488 *) Remove some legacy and erroneous uses of malloc, free instead of
9490 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9492 *) Make rsa_oaep_test return non-zero on error.
9493 [Ulf Moeller <ulf@fitug.de>]
9495 *) Add support for native Solaris shared libraries. Configure
9496 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9497 if someone would make that last step automatic.
9498 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9500 *) ctx_size was not built with the right compiler during "make links". Fixed.
9503 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9504 except NULL ciphers". This means the default cipher list will no longer
9505 enable NULL ciphers. They need to be specifically enabled e.g. with
9506 the string "DEFAULT:eNULL".
9509 *) Fix to RSA private encryption routines: if p < q then it would
9510 occasionally produce an invalid result. This will only happen with
9511 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9514 *) Be less restrictive and allow also `perl util/perlpath.pl
9515 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9516 because this way one can also use an interpreter named `perl5' (which is
9517 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9518 installed as `perl').
9519 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9521 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9522 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9524 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9525 advapi32.lib to Win32 build and change the pem test comparision
9526 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9527 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9528 and crypto/des/ede_cbcm_enc.c.
9531 *) DES quad checksum was broken on big-endian architectures. Fixed.
9534 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9535 Win32 test batch file so it (might) work again. The Win32 test batch file
9536 is horrible: I feel ill....
9539 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9540 in e_os.h. Audit of header files to check ANSI and non ANSI
9541 sections: 10 functions were absent from non ANSI section and not exported
9542 from Windows DLLs. Fixed up libeay.num for new functions.
9545 *) Make `openssl version' output lines consistent.
9546 [Ralf S. Engelschall]
9548 *) Fix Win32 symbol export lists for BIO functions: Added
9549 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9550 to ms/libeay{16,32}.def.
9551 [Ralf S. Engelschall]
9553 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9554 fine under Unix and passes some trivial tests I've now added. But the
9555 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9556 added to make sure no one expects that this stuff really works in the
9557 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9558 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9560 [Ralf S. Engelschall]
9562 *) Fix the generation of two part addresses in perl.
9563 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9565 *) Add config entry for Linux on MIPS.
9566 [John Tobey <jtobey@channel1.com>]
9568 *) Make links whenever Configure is run, unless we are on Windoze.
9571 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9572 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9576 *) Add a useful kludge to allow package maintainers to specify compiler and
9577 other platforms details on the command line without having to patch the
9578 Configure script everytime: One now can use ``perl Configure
9579 <id>:<details>'', i.e. platform ids are allowed to have details appended
9580 to them (seperated by colons). This is treated as there would be a static
9581 pre-configured entry in Configure's %table under key <id> with value
9582 <details> and ``perl Configure <id>'' is called. So, when you want to
9583 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9584 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9585 now, which overrides the FreeBSD-elf entry on-the-fly.
9586 [Ralf S. Engelschall]
9588 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9591 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9592 on the `perl Configure ...' command line. This way one can compile
9593 OpenSSL libraries with Position Independent Code (PIC) which is needed
9594 for linking it into DSOs.
9595 [Ralf S. Engelschall]
9597 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9601 *) Cleaned up the LICENSE document: The official contact for any license
9602 questions now is the OpenSSL core team under openssl-core@openssl.org.
9603 And add a paragraph about the dual-license situation to make sure people
9604 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9605 to the OpenSSL toolkit.
9606 [Ralf S. Engelschall]
9608 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9609 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9610 Additonally cleaned up the `make links' target: Remove unnecessary
9611 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9612 to speed processing and no longer clutter the display with confusing
9613 stuff. Instead only the actually done links are displayed.
9614 [Ralf S. Engelschall]
9616 *) Permit null encryption ciphersuites, used for authentication only. It used
9617 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9618 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9622 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9623 signed attributes when verifying signatures (this would break them),
9624 the detached data encoding was wrong and public keys obtained using
9625 X509_get_pubkey() weren't freed.
9628 *) Add text documentation for the BUFFER functions. Also added a work around
9629 to a Win95 console bug. This was triggered by the password read stuff: the
9630 last character typed gets carried over to the next fread(). If you were
9631 generating a new cert request using 'req' for example then the last
9632 character of the passphrase would be CR which would then enter the first
9636 *) Added the new `Includes OpenSSL Cryptography Software' button as
9637 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9638 button and can be used by applications based on OpenSSL to show the
9639 relationship to the OpenSSL project.
9640 [Ralf S. Engelschall]
9642 *) Remove confusing variables in function signatures in files
9643 ssl/ssl_lib.c and ssl/ssl.h.
9644 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9646 *) Don't install bss_file.c under PREFIX/include/
9647 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9649 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9650 functions that return function pointers and has support for NT specific
9651 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9652 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9653 unsigned to signed types: this was killing the Win32 compile.
9656 *) Add new certificate file to stack functions,
9657 SSL_add_dir_cert_subjects_to_stack() and
9658 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9659 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9660 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9661 This means that Apache-SSL and similar packages don't have to mess around
9662 to add as many CAs as they want to the preferred list.
9665 *) Experiment with doxygen documentation. Currently only partially applied to
9667 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9668 openssl.doxy as the configuration file.
9671 *) Get rid of remaining C++-style comments which strict C compilers hate.
9672 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9674 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9675 compiled in by default: it has problems with large keys.
9678 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9679 DH private keys and/or callback functions which directly correspond to
9680 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9681 is needed for applications which have to configure certificates on a
9682 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9684 For the RSA certificate situation is makes no difference, but
9685 for the DSA certificate situation this fixes the "no shared cipher"
9686 problem where the OpenSSL cipher selection procedure failed because the
9687 temporary keys were not overtaken from the context and the API provided
9688 no way to reconfigure them.
9689 The new functions now let applications reconfigure the stuff and they
9690 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9691 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9692 non-public-API function ssl_cert_instantiate() is used as a helper
9693 function and also to reduce code redundancy inside ssl_rsa.c.
9694 [Ralf S. Engelschall]
9696 *) Move s_server -dcert and -dkey options out of the undocumented feature
9697 area because they are useful for the DSA situation and should be
9698 recognized by the users.
9699 [Ralf S. Engelschall]
9701 *) Fix the cipher decision scheme for export ciphers: the export bits are
9702 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9703 SSL_EXP_MASK. So, the original variable has to be used instead of the
9704 already masked variable.
9705 [Richard Levitte <levitte@stacken.kth.se>]
9707 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9708 [Richard Levitte <levitte@stacken.kth.se>]
9710 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9711 from `int' to `unsigned int' because it's a length and initialized by
9712 EVP_DigestFinal() which expects an `unsigned int *'.
9713 [Richard Levitte <levitte@stacken.kth.se>]
9715 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9716 script. Instead use the usual Shell->Perl transition trick.
9717 [Ralf S. Engelschall]
9719 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9720 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9721 -noout -modulus' as it's already the case for `openssl rsa -noout
9722 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9723 currently the public key is printed (a decision which was already done by
9724 `openssl dsa -modulus' in the past) which serves a similar purpose.
9725 Additionally the NO_RSA no longer completely removes the whole -modulus
9726 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9728 [Ralf S. Engelschall]
9730 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9731 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9732 [Arne Ansper <arne@ats.cyber.ee>]
9734 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9735 to be added. Now both 'req' and 'ca' can use new objects defined in the
9739 *) Add cool BIO that does syslog (or event log on NT).
9740 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9742 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9743 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9744 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9745 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9748 *) Add preliminary config info for new extension code.
9751 *) Make RSA_NO_PADDING really use no padding.
9752 [Ulf Moeller <ulf@fitug.de>]
9754 *) Generate errors when private/public key check is done.
9757 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9758 for some CRL extensions and new objects added.
9761 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9762 key usage extension and fuller support for authority key id.
9765 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9766 padding method for RSA, which is recommended for new applications in PKCS
9767 #1 v2.0 (RFC 2437, October 1998).
9768 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9769 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9770 against Bleichbacher's attack on RSA.
9771 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9774 *) Updates to the new SSL compression code
9775 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9777 *) Fix so that the version number in the master secret, when passed
9778 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9779 (because the server will not accept higher), that the version number
9780 is 0x03,0x01, not 0x03,0x00
9781 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9783 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9784 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9785 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9788 *) Support for RAW extensions where an arbitrary extension can be
9789 created by including its DER encoding. See apps/openssl.cnf for
9793 *) Make sure latest Perl versions don't interpret some generated C array
9794 code as Perl array code in the crypto/err/err_genc.pl script.
9795 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9797 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9798 not many people have the assembler. Various Win32 compilation fixes and
9799 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9803 *) Modify configure script 'Configure' to automatically create crypto/date.h
9804 file under Win32 and also build pem.h from pem.org. New script
9805 util/mkfiles.pl to create the MINFO file on environments that can't do a
9806 'make files': perl util/mkfiles.pl >MINFO should work.
9809 *) Major rework of DES function declarations, in the pursuit of correctness
9810 and purity. As a result, many evil casts evaporated, and some weirdness,
9811 too. You may find this causes warnings in your code. Zapping your evil
9812 casts will probably fix them. Mostly.
9815 *) Fix for a typo in asn1.h. Bug fix to object creation script
9816 obj_dat.pl. It considered a zero in an object definition to mean
9817 "end of object": none of the objects in objects.h have any zeros
9818 so it wasn't spotted.
9819 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9821 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9822 Masking (CBCM). In the absence of test vectors, the best I have been able
9823 to do is check that the decrypt undoes the encrypt, so far. Send me test
9824 vectors if you have them.
9827 *) Correct calculation of key length for export ciphers (too much space was
9828 allocated for null ciphers). This has not been tested!
9831 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9832 message is now correct (it understands "crypto" and "ssl" on its
9833 command line). There is also now an "update" option. This will update
9834 the util/ssleay.num and util/libeay.num files with any new functions.
9836 perl util/mkdef.pl crypto ssl update
9837 it will update them.
9840 *) Overhauled the Perl interface (perl/*):
9841 - ported BN stuff to OpenSSL's different BN library
9842 - made the perl/ source tree CVS-aware
9843 - renamed the package from SSLeay to OpenSSL (the files still contain
9844 their history because I've copied them in the repository)
9845 - removed obsolete files (the test scripts will be replaced
9846 by better Test::Harness variants in the future)
9847 [Ralf S. Engelschall]
9849 *) First cut for a very conservative source tree cleanup:
9850 1. merge various obsolete readme texts into doc/ssleay.txt
9851 where we collect the old documents and readme texts.
9852 2. remove the first part of files where I'm already sure that we no
9853 longer need them because of three reasons: either they are just temporary
9854 files which were left by Eric or they are preserved original files where
9855 I've verified that the diff is also available in the CVS via "cvs diff
9856 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9857 the crypto/md/ stuff).
9858 [Ralf S. Engelschall]
9860 *) More extension code. Incomplete support for subject and issuer alt
9861 name, issuer and authority key id. Change the i2v function parameters
9862 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9863 what that's for :-) Fix to ASN1 macro which messed up
9864 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9867 *) Preliminary support for ENUMERATED type. This is largely copied from the
9871 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9872 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9874 *) Make sure `make rehash' target really finds the `openssl' program.
9875 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9877 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9878 like to hear about it if this slows down other processors.
9881 *) Add CygWin32 platform information to Configure script.
9882 [Alan Batie <batie@aahz.jf.intel.com>]
9884 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
9885 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
9887 *) New program nseq to manipulate netscape certificate sequences
9890 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
9894 *) Fixes to BN code. Previously the default was to define BN_RECURSION
9895 but the BN code had some problems that would cause failures when
9896 doing certificate verification and some other functions.
9897 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9899 *) Add ASN1 and PEM code to support netscape certificate sequences.
9902 *) Add ASN1 and PEM code to support netscape certificate sequences.
9905 *) Add several PKIX and private extended key usage OIDs.
9908 *) Modify the 'ca' program to handle the new extension code. Modify
9909 openssl.cnf for new extension format, add comments.
9912 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
9913 and add a sample to openssl.cnf so req -x509 now adds appropriate
9917 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
9918 error code, add initial support to X509_print() and x509 application.
9921 *) Takes a deep breath and start addding X509 V3 extension support code. Add
9922 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
9923 stuff is currently isolated and isn't even compiled yet.
9926 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
9927 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
9928 Removed the versions check from X509 routines when loading extensions:
9929 this allows certain broken certificates that don't set the version
9930 properly to be processed.
9933 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
9934 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
9935 can still be regenerated with "make depend".
9938 *) Spelling mistake in C version of CAST-128.
9939 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
9941 *) Changes to the error generation code. The perl script err-code.pl
9942 now reads in the old error codes and retains the old numbers, only
9943 adding new ones if necessary. It also only changes the .err files if new
9944 codes are added. The makefiles have been modified to only insert errors
9945 when needed (to avoid needlessly modifying header files). This is done
9946 by only inserting errors if the .err file is newer than the auto generated
9947 C file. To rebuild all the error codes from scratch (the old behaviour)
9948 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
9949 or delete all the .err files.
9952 *) CAST-128 was incorrectly implemented for short keys. The C version has
9953 been fixed, but is untested. The assembler versions are also fixed, but
9954 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
9955 to regenerate it if needed.
9956 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
9957 Hagino <itojun@kame.net>]
9959 *) File was opened incorrectly in randfile.c.
9960 [Ulf Möller <ulf@fitug.de>]
9962 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
9963 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
9964 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
9965 al: it's just almost always a UTCTime. Note this patch adds new error
9966 codes so do a "make errors" if there are problems.
9969 *) Correct Linux 1 recognition in config.
9970 [Ulf Möller <ulf@fitug.de>]
9972 *) Remove pointless MD5 hash when using DSA keys in ca.
9973 [Anonymous <nobody@replay.com>]
9975 *) Generate an error if given an empty string as a cert directory. Also
9976 generate an error if handed NULL (previously returned 0 to indicate an
9977 error, but didn't set one).
9978 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
9980 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
9983 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
9984 parameters. This was causing a warning which killed off the Win32 compile.
9987 *) Remove C++ style comments from crypto/bn/bn_local.h.
9988 [Neil Costigan <neil.costigan@celocom.com>]
9990 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
9991 based on a text string, looking up short and long names and finally
9992 "dot" format. The "dot" format stuff didn't work. Added new function
9993 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
9994 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
9995 OID is not part of the table.
9998 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
9999 X509_LOOKUP_by_alias().
10002 *) Sort openssl functions by name.
10005 *) Get the gendsa program working (hopefully) and add it to app list. Remove
10006 encryption from sample DSA keys (in case anyone is interested the password
10010 *) Make _all_ *_free functions accept a NULL pointer.
10011 [Frans Heymans <fheymans@isaserver.be>]
10013 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10015 [Anonymous <nobody@replay.com>]
10017 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10018 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10020 *) Don't blow it for numeric -newkey arguments to apps/req.
10021 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10023 *) Temp key "for export" tests were wrong in s3_srvr.c.
10024 [Anonymous <nobody@replay.com>]
10026 *) Add prototype for temp key callback functions
10027 SSL_CTX_set_tmp_{rsa,dh}_callback().
10030 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10031 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10034 *) X509_name_add_entry() freed the wrong thing after an error.
10035 [Arne Ansper <arne@ats.cyber.ee>]
10037 *) rsa_eay.c would attempt to free a NULL context.
10038 [Arne Ansper <arne@ats.cyber.ee>]
10040 *) BIO_s_socket() had a broken should_retry() on Windoze.
10041 [Arne Ansper <arne@ats.cyber.ee>]
10043 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10044 [Arne Ansper <arne@ats.cyber.ee>]
10046 *) Make sure the already existing X509_STORE->depth variable is initialized
10047 in X509_STORE_new(), but document the fact that this variable is still
10048 unused in the certificate verification process.
10049 [Ralf S. Engelschall]
10051 *) Fix the various library and apps files to free up pkeys obtained from
10052 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10055 *) Fix reference counting in X509_PUBKEY_get(). This makes
10056 demos/maurice/example2.c work, amongst others, probably.
10057 [Steve Henson and Ben Laurie]
10059 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10060 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10061 are no longer created. This way we have a single and consistent command
10062 line interface `openssl <command>', similar to `cvs <command>'.
10063 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10065 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10066 BIT STRING wrapper always have zero unused bits.
10069 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10072 *) Make the top-level INSTALL documentation easier to understand.
10075 *) Makefiles updated to exit if an error occurs in a sub-directory
10076 make (including if user presses ^C) [Paul Sutton]
10078 *) Make Montgomery context stuff explicit in RSA data structure.
10081 *) Fix build order of pem and err to allow for generated pem.h.
10084 *) Fix renumbering bug in X509_NAME_delete_entry().
10087 *) Enhanced the err-ins.pl script so it makes the error library number
10088 global and can add a library name. This is needed for external ASN1 and
10089 other error libraries.
10092 *) Fixed sk_insert which never worked properly.
10095 *) Fix ASN1 macros so they can handle indefinite length construted
10096 EXPLICIT tags. Some non standard certificates use these: they can now
10100 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10101 into a single doc/ssleay.txt bundle. This way the information is still
10102 preserved but no longer messes up this directory. Now it's new room for
10103 the new set of documenation files.
10104 [Ralf S. Engelschall]
10106 *) SETs were incorrectly DER encoded. This was a major pain, because they
10107 shared code with SEQUENCEs, which aren't coded the same. This means that
10108 almost everything to do with SETs or SEQUENCEs has either changed name or
10109 number of arguments.
10110 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10112 *) Fix test data to work with the above.
10115 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10116 was already fixed by Eric for 0.9.1 it seems.
10117 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10119 *) Autodetect FreeBSD3.
10122 *) Fix various bugs in Configure. This affects the following platforms:
10126 unixware-2.0-pentium
10130 *) Eliminate generated files from CVS. Reorder tests to regenerate files
10131 before they are needed.
10134 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
10138 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
10140 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
10141 changed SSLeay to OpenSSL in version strings.
10142 [Ralf S. Engelschall]
10144 *) Some fixups to the top-level documents.
10147 *) Fixed the nasty bug where rsaref.h was not found under compile-time
10148 because the symlink to include/ was missing.
10149 [Ralf S. Engelschall]
10151 *) Incorporated the popular no-RSA/DSA-only patches
10152 which allow to compile a RSA-free SSLeay.
10153 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
10155 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
10156 when "ssleay" is still not found.
10157 [Ralf S. Engelschall]
10159 *) Added more platforms to Configure: Cray T3E, HPUX 11,
10160 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
10162 *) Updated the README file.
10163 [Ralf S. Engelschall]
10165 *) Added various .cvsignore files in the CVS repository subdirs
10166 to make a "cvs update" really silent.
10167 [Ralf S. Engelschall]
10169 *) Recompiled the error-definition header files and added
10170 missing symbols to the Win32 linker tables.
10171 [Ralf S. Engelschall]
10173 *) Cleaned up the top-level documents;
10174 o new files: CHANGES and LICENSE
10175 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
10176 o merged COPYRIGHT into LICENSE
10177 o removed obsolete TODO file
10178 o renamed MICROSOFT to INSTALL.W32
10179 [Ralf S. Engelschall]
10181 *) Removed dummy files from the 0.9.1b source tree:
10182 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
10183 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
10184 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
10185 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
10186 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
10187 [Ralf S. Engelschall]
10189 *) Added various platform portability fixes.
10192 *) The Genesis of the OpenSSL rpject:
10193 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
10194 Young and Tim J. Hudson created while they were working for C2Net until
10196 [The OpenSSL Project]
10199 Changes between 0.9.0b and 0.9.1b [not released]
10201 *) Updated a few CA certificates under certs/
10204 *) Changed some BIGNUM api stuff.
10207 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
10208 DGUX x86, Linux Alpha, etc.
10211 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
10212 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
10216 *) Add -strparse option to asn1pars program which parses nested
10218 [Dr Stephen Henson <shenson@bigfoot.com>]
10220 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
10223 *) DSA fix for "ca" program.
10226 *) Added "-genkey" option to "dsaparam" program.
10229 *) Added RIPE MD160 (rmd160) message digest.
10232 *) Added -a (all) option to "ssleay version" command.
10235 *) Added PLATFORM define which is the id given to Configure.
10238 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
10241 *) Extended the ASN.1 parser routines.
10244 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
10247 *) Added a BN_CTX to the BN library.
10250 *) Fixed the weak key values in DES library
10253 *) Changed API in EVP library for cipher aliases.
10256 *) Added support for RC2/64bit cipher.
10259 *) Converted the lhash library to the crypto/mem.c functions.
10262 *) Added more recognized ASN.1 object ids.
10265 *) Added more RSA padding checks for SSL/TLS.
10268 *) Added BIO proxy/filter functionality.
10271 *) Added extra_certs to SSL_CTX which can be used
10272 send extra CA certificates to the client in the CA cert chain sending
10273 process. It can be configured with SSL_CTX_add_extra_chain_cert().
10276 *) Now Fortezza is denied in the authentication phase because
10277 this is key exchange mechanism is not supported by SSLeay at all.
10280 *) Additional PKCS1 checks.
10283 *) Support the string "TLSv1" for all TLS v1 ciphers.
10286 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
10287 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
10290 *) Fixed a few memory leaks.
10293 *) Fixed various code and comment typos.
10296 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
10297 bytes sent in the client random.
10298 [Edward Bishop <ebishop@spyglass.com>]