5 Changes between 1.0.2 and 1.1.0 [xx XXX xxxx]
7 *) Harmonize version and its documentation. -f flag is used to display
9 [mancha <mancha1@zoho.com>]
11 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
13 [mancha <mancha1@zoho.com>]
15 *) Fix some double frees. These are not thought to be exploitable.
16 [mancha <mancha1@zoho.com>]
18 *) A missing bounds check in the handling of the TLS heartbeat extension
19 can be used to reveal up to 64k of memory to a connected client or
22 Thanks for Neel Mehta of Google Security for discovering this bug and to
23 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
24 preparing the fix (CVE-2014-0160)
25 [Adam Langley, Bodo Moeller]
27 *) Fix for the attack described in the paper "Recovering OpenSSL
28 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
29 by Yuval Yarom and Naomi Benger. Details can be obtained from:
30 http://eprint.iacr.org/2014/140
32 Thanks to Yuval Yarom and Naomi Benger for discovering this
33 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
34 [Yuval Yarom and Naomi Benger]
36 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
37 this fixes a limiation in previous versions of OpenSSL.
40 *) Experimental encrypt-then-mac support.
42 Experimental support for encrypt then mac from
43 draft-gutmann-tls-encrypt-then-mac-02.txt
45 To enable it set the appropriate extension number (0x42 for the test
46 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
48 For non-compliant peers (i.e. just about everything) this should have no
51 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
55 *) Add callbacks supporting generation and retrieval of supplemental
57 [Scott Deboy <sdeboy@apache.org>, Trevor Perrin and Ben Laurie]
59 *) Add EVP support for key wrapping algorithms, to avoid problems with
60 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
61 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
62 algorithms and include tests cases.
65 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
69 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
70 MGF1 digest and OAEP label.
73 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
74 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
75 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
78 *) Make openssl verify return errors.
79 [Chris Palmer <palmer@google.com> and Ben Laurie]
81 *) New function ASN1_TIME_diff to calculate the difference between two
82 ASN1_TIME structures or one structure and the current time.
85 *) Update fips_test_suite to support multiple command line options. New
86 test to induce all self test errors in sequence and check expected
90 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
91 sign or verify all in one operation.
94 *) Add fips_algvs: a multicall fips utility incorporaing all the algorithm
95 test programs and fips_test_suite. Includes functionality to parse
96 the minimal script output of fipsalgest.pl directly.
99 *) Add authorisation parameter to FIPS_module_mode_set().
102 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
105 *) Use separate DRBG fields for internal and external flags. New function
106 FIPS_drbg_health_check() to perform on demand health checking. Add
107 generation tests to fips_test_suite with reduced health check interval to
108 demonstrate periodic health checking. Add "nodh" option to
109 fips_test_suite to skip very slow DH test.
112 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
116 *) More extensive health check for DRBG checking many more failure modes.
117 New function FIPS_selftest_drbg_all() to handle every possible DRBG
118 combination: call this in fips_test_suite.
121 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
122 and POST to handle Dual EC cases.
125 *) Add support for canonical generation of DSA parameter 'g'. See
128 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
129 POST to handle HMAC cases.
132 *) Add functions FIPS_module_version() and FIPS_module_version_text()
133 to return numerical and string versions of the FIPS module number.
136 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
137 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
138 outside the validated module in the FIPS capable OpenSSL.
141 *) Minor change to DRBG entropy callback semantics. In some cases
142 there is no multiple of the block length between min_len and
143 max_len. Allow the callback to return more than max_len bytes
144 of entropy but discard any extra: it is the callback's responsibility
145 to ensure that the extra data discarded does not impact the
146 requested amount of entropy.
149 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
150 information in FIPS186-3, SP800-57 and SP800-131A.
153 *) CCM support via EVP. Interface is very similar to GCM case except we
154 must supply all data in one chunk (i.e. no update, final) and the
155 message length must be supplied if AAD is used. Add algorithm test
159 *) Initial version of POST overhaul. Add POST callback to allow the status
160 of POST to be monitored and/or failures induced. Modify fips_test_suite
161 to use callback. Always run all selftests even if one fails.
164 *) XTS support including algorithm test driver in the fips_gcmtest program.
165 Note: this does increase the maximum key length from 32 to 64 bytes but
166 there should be no binary compatibility issues as existing applications
167 will never use XTS mode.
170 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
171 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
172 performs algorithm blocking for unapproved PRNG types. Also do not
173 set PRNG type in FIPS_mode_set(): leave this to the application.
174 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
175 the standard OpenSSL PRNG: set additional data to a date time vector.
178 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
179 This shouldn't present any incompatibility problems because applications
180 shouldn't be using these directly and any that are will need to rethink
181 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
184 *) Extensive self tests and health checking required by SP800-90 DRBG.
185 Remove strength parameter from FIPS_drbg_instantiate and always
186 instantiate at maximum supported strength.
189 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
192 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
195 *) New function DH_compute_key_padded() to compute a DH key and pad with
196 leading zeroes if needed: this complies with SP800-56A et al.
199 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
200 anything, incomplete, subject to change and largely untested at present.
203 *) Modify fipscanisteronly build option to only build the necessary object
204 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
207 *) Add experimental option FIPSSYMS to give all symbols in
208 fipscanister.o and FIPS or fips prefix. This will avoid
209 conflicts with future versions of OpenSSL. Add perl script
210 util/fipsas.pl to preprocess assembly language source files
211 and rename any affected symbols.
214 *) Add selftest checks and algorithm block of non-fips algorithms in
215 FIPS mode. Remove DES2 from selftests.
218 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
219 return internal method without any ENGINE dependencies. Add new
220 tiny fips sign and verify functions.
223 *) New build option no-ec2m to disable characteristic 2 code.
226 *) New build option "fipscanisteronly". This only builds fipscanister.o
227 and (currently) associated fips utilities. Uses the file Makefile.fips
228 instead of Makefile.org as the prototype.
231 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
232 Update fips_gcmtest to use IV generator.
235 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
236 setting output buffer to NULL. The *Final function must be
237 called although it will not retrieve any additional data. The tag
238 can be set or retrieved with a ctrl. The IV length is by default 12
239 bytes (96 bits) but can be set to an alternative value. If the IV
240 length exceeds the maximum IV length (currently 16 bytes) it cannot be
244 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
245 underlying do_cipher function handles all cipher semantics itself
246 including padding and finalisation. This is useful if (for example)
247 an ENGINE cipher handles block padding itself. The behaviour of
248 do_cipher is subtly changed if this flag is set: the return value
249 is the number of characters written to the output buffer (zero is
250 no longer an error code) or a negative error code. Also if the
251 input buffer is NULL and length 0 finalisation should be performed.
254 *) If a candidate issuer certificate is already part of the constructed
255 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
258 *) Improve forward-security support: add functions
260 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
261 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
263 for use by SSL/TLS servers; the callback function will be called whenever a
264 new session is created, and gets to decide whether the session may be
265 cached to make it resumable (return 0) or not (return 1). (As by the
266 SSL/TLS protocol specifications, the session_id sent by the server will be
267 empty to indicate that the session is not resumable; also, the server will
268 not generate RFC 4507 (RFC 5077) session tickets.)
270 A simple reasonable callback implementation is to return is_forward_secure.
271 This parameter will be set to 1 or 0 depending on the ciphersuite selected
272 by the SSL/TLS server library, indicating whether it can provide forward
274 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
276 *) New -verify_name option in command line utilities to set verification
280 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
281 Add CMAC pkey methods.
284 *) Experimental regnegotiation in s_server -www mode. If the client
285 browses /reneg connection is renegotiated. If /renegcert it is
286 renegotiated requesting a certificate.
289 *) Add an "external" session cache for debugging purposes to s_server. This
290 should help trace issues which normally are only apparent in deployed
291 multi-process servers.
294 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
295 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
296 BIO_set_cipher() and some obscure PEM functions were changed so they
297 can now return an error. The RAND changes required a change to the
298 RAND_METHOD structure.
301 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
302 a gcc attribute to warn if the result of a function is ignored. This
303 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
304 whose return value is often ignored.
307 Changes between 1.0.1e and 1.0.2 [xx XXX xxxx]
309 *) Keep original DTLS digest and encryption contexts in retransmission
310 structures so we can use the previous session parameters if they need
311 to be resent. (CVE-2013-6450)
314 *) TLS pad extension: draft-agl-tls-padding-03
316 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
317 TLS client Hello record length value would otherwise be > 255 and
318 less that 512 pad with a dummy extension containing zeroes so it
319 is at least 512 bytes long.
321 [Adam Langley, Steve Henson]
323 *) Add functions to allocate and set the fields of an ECDSA_METHOD
325 [Douglas E. Engert, Steve Henson]
327 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
328 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
329 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
330 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
331 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
332 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
333 [Rob Stradling, Adam Langley]
335 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
336 difference in days and seconds between two tm or ASN1_TIME structures.
339 *) Add -rev test option to s_server to just reverse order of characters
340 received by client and send back to server. Also prints an abbreviated
341 summary of the connection parameters.
344 *) New option -brief for s_client and s_server to print out a brief summary
345 of connection parameters.
348 *) Add callbacks for arbitrary TLS extensions.
349 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
351 *) New option -crl_download in several openssl utilities to download CRLs
352 from CRLDP extension in certificates.
355 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
358 *) New function X509_CRL_diff to generate a delta CRL from the difference
359 of two full CRLs. Add support to "crl" utility.
362 *) New functions to set lookup_crls function and to retrieve
363 X509_STORE from X509_STORE_CTX.
366 *) Print out deprecated issuer and subject unique ID fields in
370 *) Extend OCSP I/O functions so they can be used for simple general purpose
371 HTTP as well as OCSP. New wrapper function which can be used to download
372 CRLs using the OCSP API.
375 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
378 *) SSL_CONF* functions. These provide a common framework for application
379 configuration using configuration files or command lines.
382 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
383 message callback and prints the results. Needs compile time option
384 "enable-ssl-trace". New options to s_client and s_server to enable
388 *) New ctrl and macro to retrieve supported points extensions.
389 Print out extension in s_server and s_client.
392 *) New functions to retrieve certificate signature and signature
396 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
400 *) New Suite B modes for TLS code. These use and enforce the requirements
401 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
402 only use Suite B curves. The Suite B modes can be set by using the
403 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
406 *) New chain verification flags for Suite B levels of security. Check
407 algorithms are acceptable when flags are set in X509_verify_cert.
410 *) Make tls1_check_chain return a set of flags indicating checks passed
411 by a certificate chain. Add additional tests to handle client
412 certificates: checks for matching certificate type and issuer name
416 *) If an attempt is made to use a signature algorithm not in the peer
417 preference list abort the handshake. If client has no suitable
418 signature algorithms in response to a certificate request do not
422 *) If server EC tmp key is not in client preference list abort handshake.
425 *) Add support for certificate stores in CERT structure. This makes it
426 possible to have different stores per SSL structure or one store in
427 the parent SSL_CTX. Include distint stores for certificate chain
428 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
429 to build and store a certificate chain in CERT structure: returing
430 an error if the chain cannot be built: this will allow applications
431 to test if a chain is correctly configured.
433 Note: if the CERT based stores are not set then the parent SSL_CTX
434 store is used to retain compatibility with existing behaviour.
438 *) New function ssl_set_client_disabled to set a ciphersuite disabled
439 mask based on the current session, check mask when sending client
440 hello and checking the requested ciphersuite.
443 *) New ctrls to retrieve and set certificate types in a certificate
444 request message. Print out received values in s_client. If certificate
445 types is not set with custom values set sensible values based on
446 supported signature algorithms.
449 *) Support for distinct client and server supported signature algorithms.
452 *) Add certificate callback. If set this is called whenever a certificate
453 is required by client or server. An application can decide which
454 certificate chain to present based on arbitrary criteria: for example
455 supported signature algorithms. Add very simple example to s_server.
456 This fixes many of the problems and restrictions of the existing client
457 certificate callback: for example you can now clear an existing
458 certificate and specify the whole chain.
461 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
462 the certificate can be used for (if anything). Set valid_flags field
463 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
464 to have similar checks in it.
466 Add new "cert_flags" field to CERT structure and include a "strict mode".
467 This enforces some TLS certificate requirements (such as only permitting
468 certificate signature algorithms contained in the supported algorithms
469 extension) which some implementations ignore: this option should be used
470 with caution as it could cause interoperability issues.
473 *) Update and tidy signature algorithm extension processing. Work out
474 shared signature algorithms based on preferences and peer algorithms
475 and print them out in s_client and s_server. Abort handshake if no
476 shared signature algorithms.
479 *) Add new functions to allow customised supported signature algorithms
480 for SSL and SSL_CTX structures. Add options to s_client and s_server
484 *) New function SSL_certs_clear() to delete all references to certificates
485 from an SSL structure. Before this once a certificate had been added
486 it couldn't be removed.
489 *) Integrate hostname, email address and IP address checking with certificate
490 verification. New verify options supporting checking in opensl utility.
493 *) Fixes and wildcard matching support to hostname and email checking
494 functions. Add manual page.
495 [Florian Weimer (Red Hat Product Security Team)]
497 *) New functions to check a hostname email or IP address against a
498 certificate. Add options x509 utility to print results of checks against
502 *) Fix OCSP checking.
503 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
505 *) Initial experimental support for explicitly trusted non-root CAs.
506 OpenSSL still tries to build a complete chain to a root but if an
507 intermediate CA has a trust setting included that is used. The first
508 setting is used: whether to trust (e.g., -addtrust option to the x509
512 *) Add -trusted_first option which attempts to find certificates in the
513 trusted store even if an untrusted chain is also supplied.
516 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
517 platform support for Linux and Android.
520 *) Support for linux-x32, ILP32 environment in x86_64 framework.
523 *) RFC 5878 (TLS Authorization Extensions) support.
524 [Emilia Kasper, Adam Langley, Ben Laurie (Google)]
526 *) Experimental multi-implementation support for FIPS capable OpenSSL.
527 When in FIPS mode the approved implementations are used as normal,
528 when not in FIPS mode the internal unapproved versions are used instead.
529 This means that the FIPS capable OpenSSL isn't forced to use the
530 (often lower perfomance) FIPS implementations outside FIPS mode.
533 *) Transparently support X9.42 DH parameters when calling
534 PEM_read_bio_DHparameters. This means existing applications can handle
535 the new parameter format automatically.
538 *) Initial experimental support for X9.42 DH parameter format: mainly
539 to support use of 'q' parameter for RFC5114 parameters.
542 *) Add DH parameters from RFC5114 including test data to dhtest.
545 *) Support for automatic EC temporary key parameter selection. If enabled
546 the most preferred EC parameters are automatically used instead of
547 hardcoded fixed parameters. Now a server just has to call:
548 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
549 support ECDH and use the most appropriate parameters.
552 *) Enhance and tidy EC curve and point format TLS extension code. Use
553 static structures instead of allocation if default values are used.
554 New ctrls to set curves we wish to support and to retrieve shared curves.
555 Print out shared curves in s_server. New options to s_server and s_client
556 to set list of supported curves.
559 *) New ctrls to retrieve supported signature algorithms and
560 supported curve values as an array of NIDs. Extend openssl utility
561 to print out received values.
564 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
565 between NIDs and the more common NIST names such as "P-256". Enhance
566 ecparam utility and ECC method to recognise the NIST names for curves.
569 *) Enhance SSL/TLS certificate chain handling to support different
570 chains for each certificate instead of one chain in the parent SSL_CTX.
573 *) Support for fixed DH ciphersuite client authentication: where both
574 server and client use DH certificates with common parameters.
577 *) Support for fixed DH ciphersuites: those requiring DH server
581 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
583 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
584 supporting platforms or when small records were transferred.
585 [Andy Polyakov, Steve Henson]
587 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
589 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
591 This addresses the flaw in CBC record processing discovered by
592 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
593 at: http://www.isg.rhul.ac.uk/tls/
595 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
596 Security Group at Royal Holloway, University of London
597 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
598 Emilia Käsper for the initial patch.
600 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
602 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
603 ciphersuites which can be exploited in a denial of service attack.
604 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
605 and detecting this bug and to Wolfgang Ettlinger
606 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
610 *) Return an error when checking OCSP signatures when key is NULL.
611 This fixes a DoS attack. (CVE-2013-0166)
614 *) Make openssl verify return errors.
615 [Chris Palmer <palmer@google.com> and Ben Laurie]
617 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
618 the right response is stapled. Also change SSL_get_certificate()
619 so it returns the certificate actually sent.
620 See http://rt.openssl.org/Ticket/Display.html?id=2836.
621 [Rob Stradling <rob.stradling@comodo.com>]
623 *) Fix possible deadlock when decoding public keys.
626 *) Don't use TLS 1.0 record version number in initial client hello
630 Changes between 1.0.1b and 1.0.1c [10 May 2012]
632 *) Sanity check record length before skipping explicit IV in TLS
633 1.2, 1.1 and DTLS to fix DoS attack.
635 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
636 fuzzing as a service testing platform.
640 *) Initialise tkeylen properly when encrypting CMS messages.
641 Thanks to Solar Designer of Openwall for reporting this issue.
644 *) In FIPS mode don't try to use composite ciphers as they are not
648 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
650 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
651 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
652 mean any application compiled against OpenSSL 1.0.0 headers setting
653 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
654 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
655 0x10000000L Any application which was previously compiled against
656 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
657 will need to be recompiled as a result. Letting be results in
658 inability to disable specifically TLS 1.1 and in client context,
659 in unlike event, limit maximum offered version to TLS 1.0 [see below].
662 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
663 disable just protocol X, but all protocols above X *if* there are
664 protocols *below* X still enabled. In more practical terms it means
665 that if application wants to disable TLS1.0 in favor of TLS1.1 and
666 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
667 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
671 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
673 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
674 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
675 in CRYPTO_realloc_clean.
677 Thanks to Tavis Ormandy, Google Security Team, for discovering this
678 issue and to Adam Langley <agl@chromium.org> for fixing it.
680 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
682 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
685 *) Workarounds for some broken servers that "hang" if a client hello
686 record length exceeds 255 bytes.
688 1. Do not use record version number > TLS 1.0 in initial client
689 hello: some (but not all) hanging servers will now work.
690 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
691 the number of ciphers sent in the client hello. This should be
692 set to an even number, such as 50, for example by passing:
693 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
694 Most broken servers should now work.
695 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
696 TLS 1.2 client support entirely.
699 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
702 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
704 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
705 STRING form instead of a DigestInfo.
708 *) The format used for MDC2 RSA signatures is inconsistent between EVP
709 and the RSA_sign/RSA_verify functions. This was made more apparent when
710 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
711 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
712 the correct format in RSA_verify so both forms transparently work.
715 *) Some servers which support TLS 1.0 can choke if we initially indicate
716 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
717 encrypted premaster secret. As a workaround use the maximum pemitted
718 client version in client hello, this should keep such servers happy
719 and still work with previous versions of OpenSSL.
722 *) Add support for TLS/DTLS heartbeats.
723 [Robin Seggelmann <seggelmann@fh-muenster.de>]
725 *) Add support for SCTP.
726 [Robin Seggelmann <seggelmann@fh-muenster.de>]
728 *) Improved PRNG seeding for VOS.
729 [Paul Green <Paul.Green@stratus.com>]
731 *) Extensive assembler packs updates, most notably:
733 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
734 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
735 - x86_64: bit-sliced AES implementation;
736 - ARM: NEON support, contemporary platforms optimizations;
737 - s390x: z196 support;
738 - *: GHASH and GF(2^m) multiplication implementations;
742 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
743 (removal of unnecessary code)
744 [Peter Sylvester <peter.sylvester@edelweb.fr>]
746 *) Add TLS key material exporter from RFC 5705.
749 *) Add DTLS-SRTP negotiation from RFC 5764.
752 *) Add Next Protocol Negotiation,
753 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
754 disabled with a no-npn flag to config or Configure. Code donated
756 [Adam Langley <agl@google.com> and Ben Laurie]
758 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
759 NIST-P256, NIST-P521, with constant-time single point multiplication on
760 typical inputs. Compiler support for the nonstandard type __uint128_t is
761 required to use this (present in gcc 4.4 and later, for 64-bit builds).
762 Code made available under Apache License version 2.0.
764 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
765 line to include this in your build of OpenSSL, and run "make depend" (or
766 "make update"). This enables the following EC_METHODs:
768 EC_GFp_nistp224_method()
769 EC_GFp_nistp256_method()
770 EC_GFp_nistp521_method()
772 EC_GROUP_new_by_curve_name() will automatically use these (while
773 EC_GROUP_new_curve_GFp() currently prefers the more flexible
775 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
777 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
778 all platforms. Move ssize_t definition from e_os.h to the public
779 header file e_os2.h as it now appears in public header file cms.h
782 *) New -sigopt option to the ca, req and x509 utilities. Additional
783 signature parameters can be passed using this option and in
787 *) Add RSA PSS signing function. This will generate and set the
788 appropriate AlgorithmIdentifiers for PSS based on those in the
789 corresponding EVP_MD_CTX structure. No application support yet.
792 *) Support for companion algorithm specific ASN1 signing routines.
793 New function ASN1_item_sign_ctx() signs a pre-initialised
794 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
795 the appropriate parameters.
798 *) Add new algorithm specific ASN1 verification initialisation function
799 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
800 handling will be the same no matter what EVP_PKEY_METHOD is used.
801 Add a PSS handler to support verification of PSS signatures: checked
802 against a number of sample certificates.
805 *) Add signature printing for PSS. Add PSS OIDs.
806 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
808 *) Add algorithm specific signature printing. An individual ASN1 method
809 can now print out signatures instead of the standard hex dump.
811 More complex signatures (e.g. PSS) can print out more meaningful
812 information. Include DSA version that prints out the signature
816 *) Password based recipient info support for CMS library: implementing
820 *) Split password based encryption into PBES2 and PBKDF2 functions. This
821 neatly separates the code into cipher and PBE sections and is required
822 for some algorithms that split PBES2 into separate pieces (such as
826 *) Session-handling fixes:
827 - Fix handling of connections that are resuming with a session ID,
828 but also support Session Tickets.
829 - Fix a bug that suppressed issuing of a new ticket if the client
830 presented a ticket with an expired session.
831 - Try to set the ticket lifetime hint to something reasonable.
832 - Make tickets shorter by excluding irrelevant information.
833 - On the client side, don't ignore renewed tickets.
834 [Adam Langley, Bodo Moeller (Google)]
836 *) Fix PSK session representation.
839 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
841 This work was sponsored by Intel.
844 *) Add GCM support to TLS library. Some custom code is needed to split
845 the IV between the fixed (from PRF) and explicit (from TLS record)
846 portions. This adds all GCM ciphersuites supported by RFC5288 and
847 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
848 add a special AESGCM string for GCM only.
851 *) Expand range of ctrls for AES GCM. Permit setting invocation
852 field on decrypt and retrieval of invocation field only on encrypt.
855 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
856 As required by RFC5289 these ciphersuites cannot be used if for
857 versions of TLS earlier than 1.2.
860 *) For FIPS capable OpenSSL interpret a NULL default public key method
861 as unset and return the appopriate default but do *not* set the default.
862 This means we can return the appopriate method in applications that
863 swicth between FIPS and non-FIPS modes.
866 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
867 ENGINE is used then we cannot handle that in the FIPS module so we
868 keep original code iff non-FIPS operations are allowed.
871 *) Add -attime option to openssl utilities.
872 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
874 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
877 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
878 FIPS EC methods unconditionally for now.
881 *) New build option no-ec2m to disable characteristic 2 code.
884 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
885 all cases can be covered as some introduce binary incompatibilities.
888 *) Redirect RSA operations to FIPS module including keygen,
889 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
892 *) Add similar low level API blocking to ciphers.
895 *) Low level digest APIs are not approved in FIPS mode: any attempt
896 to use these will cause a fatal error. Applications that *really* want
897 to use them can use the private_* version instead.
900 *) Redirect cipher operations to FIPS module for FIPS builds.
903 *) Redirect digest operations to FIPS module for FIPS builds.
906 *) Update build system to add "fips" flag which will link in fipscanister.o
907 for static and shared library builds embedding a signature if needed.
910 *) Output TLS supported curves in preference order instead of numerical
911 order. This is currently hardcoded for the highest order curves first.
912 This should be configurable so applications can judge speed vs strength.
915 *) Add TLS v1.2 server support for client authentication.
918 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
922 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
923 FIPS modules versions.
926 *) Add TLS v1.2 client side support for client authentication. Keep cache
927 of handshake records longer as we don't know the hash algorithm to use
928 until after the certificate request message is received.
931 *) Initial TLS v1.2 client support. Add a default signature algorithms
932 extension including all the algorithms we support. Parse new signature
933 format in client key exchange. Relax some ECC signing restrictions for
934 TLS v1.2 as indicated in RFC5246.
937 *) Add server support for TLS v1.2 signature algorithms extension. Switch
938 to new signature format when needed using client digest preference.
939 All server ciphersuites should now work correctly in TLS v1.2. No client
940 support yet and no support for client certificates.
943 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
944 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
945 ciphersuites. At present only RSA key exchange ciphersuites work with
946 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
947 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
948 and version checking.
951 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
952 with this defined it will not be affected by any changes to ssl internal
953 structures. Add several utility functions to allow openssl application
954 to work with OPENSSL_NO_SSL_INTERN defined.
958 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
960 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
963 *) Permit abbreviated handshakes when renegotiating using the function
964 SSL_renegotiate_abbreviated().
965 [Robin Seggelmann <seggelmann@fh-muenster.de>]
967 *) Add call to ENGINE_register_all_complete() to
968 ENGINE_load_builtin_engines(), so some implementations get used
969 automatically instead of needing explicit application support.
972 *) Add support for TLS key exporter as described in RFC5705.
973 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
975 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
976 a few changes are required:
978 Add SSL_OP_NO_TLSv1_1 flag.
980 Update version checking logic to handle version 1.1.
981 Add explicit IV handling (ported from DTLS code).
982 Add command line options to s_client/s_server.
985 Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
987 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
989 This addresses the flaw in CBC record processing discovered by
990 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
991 at: http://www.isg.rhul.ac.uk/tls/
993 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
994 Security Group at Royal Holloway, University of London
995 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
996 Emilia Käsper for the initial patch.
998 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1000 *) Return an error when checking OCSP signatures when key is NULL.
1001 This fixes a DoS attack. (CVE-2013-0166)
1004 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1005 the right response is stapled. Also change SSL_get_certificate()
1006 so it returns the certificate actually sent.
1007 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1008 (This is a backport)
1009 [Rob Stradling <rob.stradling@comodo.com>]
1011 *) Fix possible deadlock when decoding public keys.
1014 Changes between 1.0.0i and 1.0.0j [10 May 2012]
1016 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
1019 *) Sanity check record length before skipping explicit IV in DTLS
1022 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1023 fuzzing as a service testing platform.
1027 *) Initialise tkeylen properly when encrypting CMS messages.
1028 Thanks to Solar Designer of Openwall for reporting this issue.
1031 Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
1033 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1034 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1035 in CRYPTO_realloc_clean.
1037 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1038 issue and to Adam Langley <agl@chromium.org> for fixing it.
1040 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1042 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
1044 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1045 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1046 content decryption and always return the same error. Note: this attack
1047 needs on average 2^20 messages so it only affects automated senders. The
1048 old behaviour can be reenabled in the CMS code by setting the
1049 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1050 an MMA defence is not necessary.
1051 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1052 this issue. (CVE-2012-0884)
1055 *) Fix CVE-2011-4619: make sure we really are receiving a
1056 client hello before rejecting multiple SGC restarts. Thanks to
1057 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1060 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
1062 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1063 Thanks to Antonio Martin, Enterprise Secure Access Research and
1064 Development, Cisco Systems, Inc. for discovering this bug and
1065 preparing a fix. (CVE-2012-0050)
1068 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
1070 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1071 of the Vaudenay padding oracle attack on CBC mode encryption
1072 which enables an efficient plaintext recovery attack against
1073 the OpenSSL implementation of DTLS. Their attack exploits timing
1074 differences arising during decryption processing. A research
1075 paper describing this attack can be found at:
1076 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1077 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1078 Security Group at Royal Holloway, University of London
1079 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1080 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1081 for preparing the fix. (CVE-2011-4108)
1082 [Robin Seggelmann, Michael Tuexen]
1084 *) Clear bytes used for block padding of SSL 3.0 records.
1086 [Adam Langley (Google)]
1088 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1089 Kadianakis <desnacked@gmail.com> for discovering this issue and
1090 Adam Langley for preparing the fix. (CVE-2011-4619)
1091 [Adam Langley (Google)]
1093 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
1094 [Andrey Kulikov <amdeich@gmail.com>]
1096 *) Prevent malformed RFC3779 data triggering an assertion failure.
1097 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1098 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1099 [Rob Austein <sra@hactrn.net>]
1101 *) Improved PRNG seeding for VOS.
1102 [Paul Green <Paul.Green@stratus.com>]
1104 *) Fix ssl_ciph.c set-up race.
1105 [Adam Langley (Google)]
1107 *) Fix spurious failures in ecdsatest.c.
1108 [Emilia Käsper (Google)]
1110 *) Fix the BIO_f_buffer() implementation (which was mixing different
1111 interpretations of the '..._len' fields).
1112 [Adam Langley (Google)]
1114 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1115 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1116 threads won't reuse the same blinding coefficients.
1118 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1119 lock to call BN_BLINDING_invert_ex, and avoids one use of
1120 BN_BLINDING_update for each BN_BLINDING structure (previously,
1121 the last update always remained unused).
1122 [Emilia Käsper (Google)]
1124 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
1125 [Bob Buckholz (Google)]
1127 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
1129 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
1130 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
1131 [Kaspar Brand <ossl@velox.ch>]
1133 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1134 for multi-threaded use of ECDH. (CVE-2011-3210)
1135 [Adam Langley (Google)]
1137 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1140 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
1141 signature public key algorithm by using OID xref utilities instead.
1142 Before this you could only use some ECC ciphersuites with SHA1 only.
1145 *) Add protection against ECDSA timing attacks as mentioned in the paper
1146 by Billy Bob Brumley and Nicola Tuveri, see:
1148 http://eprint.iacr.org/2011/232.pdf
1150 [Billy Bob Brumley and Nicola Tuveri]
1152 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
1154 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1155 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1157 *) Fix bug in string printing code: if *any* escaping is enabled we must
1158 escape the escape character (backslash) or the resulting string is
1162 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
1164 *) Disable code workaround for ancient and obsolete Netscape browsers
1165 and servers: an attacker can use it in a ciphersuite downgrade attack.
1166 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1169 *) Fixed J-PAKE implementation error, originally discovered by
1170 Sebastien Martini, further info and confirmation from Stefan
1171 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1174 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
1176 *) Fix extension code to avoid race conditions which can result in a buffer
1177 overrun vulnerability: resumed sessions must not be modified as they can
1178 be shared by multiple threads. CVE-2010-3864
1181 *) Fix WIN32 build system to correctly link an ENGINE directory into
1185 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
1187 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
1189 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
1191 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
1193 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
1194 context. The operation can be customised via the ctrl mechanism in
1195 case ENGINEs want to include additional functionality.
1198 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
1201 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
1202 output hashes compatible with older versions of OpenSSL.
1203 [Willy Weisz <weisz@vcpc.univie.ac.at>]
1205 *) Fix compression algorithm handling: if resuming a session use the
1206 compression algorithm of the resumed session instead of determining
1207 it from client hello again. Don't allow server to change algorithm.
1210 *) Add load_crls() function to apps tidying load_certs() too. Add option
1211 to verify utility to allow additional CRLs to be included.
1214 *) Update OCSP request code to permit adding custom headers to the request:
1215 some responders need this.
1218 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
1220 [Julia Lawall <julia@diku.dk>]
1222 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
1223 needlessly dereferenced structures, used obsolete functions and
1224 didn't handle all updated verify codes correctly.
1227 *) Disable MD2 in the default configuration.
1230 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
1231 indicate the initial BIO being pushed or popped. This makes it possible
1232 to determine whether the BIO is the one explicitly called or as a result
1233 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
1234 it handles reference counts correctly and doesn't zero out the I/O bio
1235 when it is not being explicitly popped. WARNING: applications which
1236 included workarounds for the old buggy behaviour will need to be modified
1237 or they could free up already freed BIOs.
1240 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
1241 renaming to all platforms (within the 0.9.8 branch, this was
1242 done conditionally on Netware platforms to avoid a name clash).
1243 [Guenter <lists@gknw.net>]
1245 *) Add ECDHE and PSK support to DTLS.
1246 [Michael Tuexen <tuexen@fh-muenster.de>]
1248 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
1252 *) Add "missing" function EVP_MD_flags() (without this the only way to
1253 retrieve a digest flags is by accessing the structure directly. Update
1254 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
1255 or cipher is registered as in the "from" argument. Print out all
1256 registered digests in the dgst usage message instead of manually
1257 attempting to work them out.
1260 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
1261 this allows the use of compression and extensions. Change default cipher
1262 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
1263 by default unless an application cipher string requests it.
1266 *) Alter match criteria in PKCS12_parse(). It used to try to use local
1267 key ids to find matching certificates and keys but some PKCS#12 files
1268 don't follow the (somewhat unwritten) rules and this strategy fails.
1269 Now just gather all certificates together and the first private key
1270 then look for the first certificate that matches the key.
1273 *) Support use of registered digest and cipher names for dgst and cipher
1274 commands instead of having to add each one as a special case. So now
1281 openssl dgst -sha256 foo
1283 and this works for ENGINE based algorithms too.
1287 *) Update Gost ENGINE to support parameter files.
1288 [Victor B. Wagner <vitus@cryptocom.ru>]
1290 *) Support GeneralizedTime in ca utility.
1291 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
1293 *) Enhance the hash format used for certificate directory links. The new
1294 form uses the canonical encoding (meaning equivalent names will work
1295 even if they aren't identical) and uses SHA1 instead of MD5. This form
1296 is incompatible with the older format and as a result c_rehash should
1297 be used to rebuild symbolic links.
1300 *) Make PKCS#8 the default write format for private keys, replacing the
1301 traditional format. This form is standardised, more secure and doesn't
1302 include an implicit MD5 dependency.
1305 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
1306 committed to OpenSSL should pass this lot as a minimum.
1309 *) Add session ticket override functionality for use by EAP-FAST.
1310 [Jouni Malinen <j@w1.fi>]
1312 *) Modify HMAC functions to return a value. Since these can be implemented
1313 in an ENGINE errors can occur.
1316 *) Type-checked OBJ_bsearch_ex.
1319 *) Type-checked OBJ_bsearch. Also some constification necessitated
1320 by type-checking. Still to come: TXT_DB, bsearch(?),
1321 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
1325 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
1326 seconds to a tm structure directly, instead of going through OS
1327 specific date routines. This avoids any issues with OS routines such
1328 as the year 2038 bug. New *_adj() functions for ASN1 time structures
1329 and X509_time_adj_ex() to cover the extended range. The existing
1330 X509_time_adj() is still usable and will no longer have any date issues.
1333 *) Delta CRL support. New use deltas option which will attempt to locate
1334 and search any appropriate delta CRLs available.
1336 This work was sponsored by Google.
1339 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
1340 code and add additional score elements. Validate alternate CRL paths
1341 as part of the CRL checking and indicate a new error "CRL path validation
1342 error" in this case. Applications wanting additional details can use
1343 the verify callback and check the new "parent" field. If this is not
1344 NULL CRL path validation is taking place. Existing applications wont
1345 see this because it requires extended CRL support which is off by
1348 This work was sponsored by Google.
1351 *) Support for freshest CRL extension.
1353 This work was sponsored by Google.
1356 *) Initial indirect CRL support. Currently only supported in the CRLs
1357 passed directly and not via lookup. Process certificate issuer
1358 CRL entry extension and lookup CRL entries by bother issuer name
1359 and serial number. Check and process CRL issuer entry in IDP extension.
1361 This work was sponsored by Google.
1364 *) Add support for distinct certificate and CRL paths. The CRL issuer
1365 certificate is validated separately in this case. Only enabled if
1366 an extended CRL support flag is set: this flag will enable additional
1367 CRL functionality in future.
1369 This work was sponsored by Google.
1372 *) Add support for policy mappings extension.
1374 This work was sponsored by Google.
1377 *) Fixes to pathlength constraint, self issued certificate handling,
1378 policy processing to align with RFC3280 and PKITS tests.
1380 This work was sponsored by Google.
1383 *) Support for name constraints certificate extension. DN, email, DNS
1384 and URI types are currently supported.
1386 This work was sponsored by Google.
1389 *) To cater for systems that provide a pointer-based thread ID rather
1390 than numeric, deprecate the current numeric thread ID mechanism and
1391 replace it with a structure and associated callback type. This
1392 mechanism allows a numeric "hash" to be extracted from a thread ID in
1393 either case, and on platforms where pointers are larger than 'long',
1394 mixing is done to help ensure the numeric 'hash' is usable even if it
1395 can't be guaranteed unique. The default mechanism is to use "&errno"
1396 as a pointer-based thread ID to distinguish between threads.
1398 Applications that want to provide their own thread IDs should now use
1399 CRYPTO_THREADID_set_callback() to register a callback that will call
1400 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
1402 Note that ERR_remove_state() is now deprecated, because it is tied
1403 to the assumption that thread IDs are numeric. ERR_remove_state(0)
1404 to free the current thread's error state should be replaced by
1405 ERR_remove_thread_state(NULL).
1407 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
1408 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
1409 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
1410 application was previously providing a numeric thread callback that
1411 was inappropriate for distinguishing threads, then uniqueness might
1412 have been obtained with &errno that happened immediately in the
1413 intermediate development versions of OpenSSL; this is no longer the
1414 case, the numeric thread callback will now override the automatic use
1416 [Geoff Thorpe, with help from Bodo Moeller]
1418 *) Initial support for different CRL issuing certificates. This covers a
1419 simple case where the self issued certificates in the chain exist and
1420 the real CRL issuer is higher in the existing chain.
1422 This work was sponsored by Google.
1425 *) Removed effectively defunct crypto/store from the build.
1428 *) Revamp of STACK to provide stronger type-checking. Still to come:
1429 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
1430 ASN1_STRING, CONF_VALUE.
1433 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
1434 RAM on SSL connections. This option can save about 34k per idle SSL.
1437 *) Revamp of LHASH to provide stronger type-checking. Still to come:
1438 STACK, TXT_DB, bsearch, qsort.
1441 *) Initial support for Cryptographic Message Syntax (aka CMS) based
1442 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
1443 support for data, signedData, compressedData, digestedData and
1444 encryptedData, envelopedData types included. Scripts to check against
1445 RFC4134 examples draft and interop and consistency checks of many
1446 content types and variants.
1449 *) Add options to enc utility to support use of zlib compression BIO.
1452 *) Extend mk1mf to support importing of options and assembly language
1453 files from Configure script, currently only included in VC-WIN32.
1454 The assembly language rules can now optionally generate the source
1455 files from the associated perl scripts.
1458 *) Implement remaining functionality needed to support GOST ciphersuites.
1459 Interop testing has been performed using CryptoPro implementations.
1460 [Victor B. Wagner <vitus@cryptocom.ru>]
1462 *) s390x assembler pack.
1465 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
1469 *) Implement Opaque PRF Input TLS extension as specified in
1470 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
1471 official specification yet and no extension type assignment by
1472 IANA exists, this extension (for now) will have to be explicitly
1473 enabled when building OpenSSL by providing the extension number
1474 to use. For example, specify an option
1476 -DTLSEXT_TYPE_opaque_prf_input=0x9527
1478 to the "config" or "Configure" script to enable the extension,
1479 assuming extension number 0x9527 (which is a completely arbitrary
1480 and unofficial assignment based on the MD5 hash of the Internet
1481 Draft). Note that by doing so, you potentially lose
1482 interoperability with other TLS implementations since these might
1483 be using the same extension number for other purposes.
1485 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
1486 opaque PRF input value to use in the handshake. This will create
1487 an interal copy of the length-'len' string at 'src', and will
1488 return non-zero for success.
1490 To get more control and flexibility, provide a callback function
1493 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
1494 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
1498 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
1501 Callback function 'cb' will be called in handshakes, and is
1502 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
1503 Argument 'arg' is for application purposes (the value as given to
1504 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
1505 be provided to the callback function). The callback function
1506 has to return non-zero to report success: usually 1 to use opaque
1507 PRF input just if possible, or 2 to enforce use of the opaque PRF
1508 input. In the latter case, the library will abort the handshake
1509 if opaque PRF input is not successfully negotiated.
1511 Arguments 'peerinput' and 'len' given to the callback function
1512 will always be NULL and 0 in the case of a client. A server will
1513 see the client's opaque PRF input through these variables if
1514 available (NULL and 0 otherwise). Note that if the server
1515 provides an opaque PRF input, the length must be the same as the
1516 length of the client's opaque PRF input.
1518 Note that the callback function will only be called when creating
1519 a new session (session resumption can resume whatever was
1520 previously negotiated), and will not be called in SSL 2.0
1521 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
1522 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
1523 for applications that need to enforce opaque PRF input.
1527 *) Update ssl code to support digests other than SHA1+MD5 for handshake
1530 [Victor B. Wagner <vitus@cryptocom.ru>]
1532 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1533 RFC4507bis. The encrypted ticket format is an encrypted encoded
1534 SSL_SESSION structure, that way new session features are automatically
1537 If a client application caches session in an SSL_SESSION structure
1538 support is transparent because tickets are now stored in the encoded
1541 The SSL_CTX structure automatically generates keys for ticket
1542 protection in servers so again support should be possible
1543 with no application modification.
1545 If a client or server wishes to disable RFC4507 support then the option
1546 SSL_OP_NO_TICKET can be set.
1548 Add a TLS extension debugging callback to allow the contents of any client
1549 or server extensions to be examined.
1551 This work was sponsored by Google.
1554 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1555 OpenSSL should now compile cleanly on gcc 4.2
1556 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1558 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1559 support including streaming MAC support: this is required for GOST
1560 ciphersuite support.
1561 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1563 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1564 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1565 to output in BER and PEM format.
1568 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1569 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1570 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1571 ENGINE support for HMAC keys which are unextractable. New -mac and
1572 -macopt options to dgst utility.
1575 *) New option -sigopt to dgst utility. Update dgst to use
1576 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1577 alternative signing paramaters such as X9.31 or PSS in the dgst
1581 *) Change ssl_cipher_apply_rule(), the internal function that does
1582 the work each time a ciphersuite string requests enabling
1583 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1584 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1585 the order of disabled ciphersuites such that those ciphersuites
1586 that most recently went from enabled to disabled not only stay
1587 in order with respect to each other, but also have higher priority
1588 than other disabled ciphersuites the next time ciphersuites are
1591 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1592 the same ciphersuites as with "HIGH" alone, but in a specific
1593 order where the PSK ciphersuites come first (since they are the
1594 most recently disabled ciphersuites when "HIGH" is parsed).
1596 Also, change ssl_create_cipher_list() (using this new
1597 funcionality) such that between otherwise identical
1598 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
1602 *) Change ssl_create_cipher_list() so that it automatically
1603 arranges the ciphersuites in reasonable order before starting
1604 to process the rule string. Thus, the definition for "DEFAULT"
1605 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
1606 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
1607 This makes it much easier to arrive at a reasonable default order
1608 in applications for which anonymous ciphers are OK (meaning
1609 that you can't actually use DEFAULT).
1610 [Bodo Moeller; suggested by Victor Duchovni]
1612 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
1613 processing) into multiple integers instead of setting
1614 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
1615 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
1616 (These masks as well as the individual bit definitions are hidden
1617 away into the non-exported interface ssl/ssl_locl.h, so this
1618 change to the definition of the SSL_CIPHER structure shouldn't
1619 affect applications.) This give us more bits for each of these
1620 categories, so there is no longer a need to coagulate AES128 and
1621 AES256 into a single algorithm bit, and to coagulate Camellia128
1622 and Camellia256 into a single algorithm bit, which has led to all
1625 Thus, among other things, the kludge introduced in 0.9.7m and
1626 0.9.8e for masking out AES256 independently of AES128 or masking
1627 out Camellia256 independently of AES256 is not needed here in 0.9.9.
1629 With the change, we also introduce new ciphersuite aliases that
1630 so far were missing: "AES128", "AES256", "CAMELLIA128", and
1634 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
1635 Use the leftmost N bytes of the signature input if the input is
1636 larger than the prime q (with N being the size in bytes of q).
1639 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1640 it yet and it is largely untested.
1643 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1646 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1647 some compilers (gcc 4.2 and later) reject their use. Safestack is
1648 reimplemented. Update ASN1 to avoid use of legacy functions.
1651 *) Win32/64 targets are linked with Winsock2.
1654 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1655 to external functions. This can be used to increase CRL handling
1656 efficiency especially when CRLs are very large by (for example) storing
1657 the CRL revoked certificates in a database.
1660 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1661 new CRLs added to a directory can be used. New command line option
1662 -verify_return_error to s_client and s_server. This causes real errors
1663 to be returned by the verify callback instead of carrying on no matter
1664 what. This reflects the way a "real world" verify callback would behave.
1667 *) GOST engine, supporting several GOST algorithms and public key formats.
1668 Kindly donated by Cryptocom.
1671 *) Partial support for Issuing Distribution Point CRL extension. CRLs
1672 partitioned by DP are handled but no indirect CRL or reason partitioning
1673 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
1674 selected via a scoring technique which handles IDP and AKID in CRLs.
1677 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1678 will ultimately be used for all verify operations: this will remove the
1679 X509_STORE dependency on certificate verification and allow alternative
1680 lookup methods. X509_STORE based implementations of these two callbacks.
1683 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1684 Modify get_crl() to find a valid (unexpired) CRL if possible.
1687 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1688 this would be called X509_CRL_cmp() but that name is already used by
1689 a function that just compares CRL issuer names. Cache several CRL
1690 extensions in X509_CRL structure and cache CRLDP in X509.
1693 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1694 this maps equivalent X509_NAME structures into a consistent structure.
1695 Name comparison can then be performed rapidly using memcmp().
1698 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1702 *) Allow digests to supply their own micalg string for S/MIME type using
1703 the ctrl EVP_MD_CTRL_MICALG.
1706 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1707 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1708 ctrl. It can then customise the structure before and/or after signing
1712 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1713 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1714 to free up any added signature OIDs.
1717 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1718 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1719 digest and cipher tables. New options added to openssl utility:
1720 list-message-digest-algorithms and list-cipher-algorithms.
1723 *) Change the array representation of binary polynomials: the list
1724 of degrees of non-zero coefficients is now terminated with -1.
1725 Previously it was terminated with 0, which was also part of the
1726 value; thus, the array representation was not applicable to
1727 polynomials where t^0 has coefficient zero. This change makes
1728 the array representation useful in a more general context.
1731 *) Various modifications and fixes to SSL/TLS cipher string
1732 handling. For ECC, the code now distinguishes between fixed ECDH
1733 with RSA certificates on the one hand and with ECDSA certificates
1734 on the other hand, since these are separate ciphersuites. The
1735 unused code for Fortezza ciphersuites has been removed.
1737 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1738 (not "ECDHE"). For consistency with the code for DH
1739 certificates, use of ECDH certificates is now considered ECDH
1740 authentication, not RSA or ECDSA authentication (the latter is
1741 merely the CA's signing algorithm and not actively used in the
1744 The temporary ciphersuite alias "ECCdraft" is no longer
1745 available, and ECC ciphersuites are no longer excluded from "ALL"
1746 and "DEFAULT". The following aliases now exist for RFC 4492
1747 ciphersuites, most of these by analogy with the DH case:
1749 kECDHr - ECDH cert, signed with RSA
1750 kECDHe - ECDH cert, signed with ECDSA
1751 kECDH - ECDH cert (signed with either RSA or ECDSA)
1752 kEECDH - ephemeral ECDH
1753 ECDH - ECDH cert or ephemeral ECDH
1759 AECDH - anonymous ECDH
1760 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1764 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1765 Use correct micalg parameters depending on digest(s) in signed message.
1768 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1769 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1772 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1773 an engine to register a method. Add ENGINE lookups for methods and
1774 functional reference processing.
1777 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1778 EVP_{Sign,Verify}* which allow an application to customise the signature
1782 *) New -resign option to smime utility. This adds one or more signers
1783 to an existing PKCS#7 signedData structure. Also -md option to use an
1784 alternative message digest algorithm for signing.
1787 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1788 create PKCS7 structures containing multiple signers. Update smime
1789 application to support multiple signers.
1792 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1796 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1797 Reorganize PBE internals to lookup from a static table using NIDs,
1798 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1799 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1800 PRF which will be automatically used with PBES2.
1803 *) Replace the algorithm specific calls to generate keys in "req" with the
1807 *) Update PKCS#7 enveloped data routines to use new API. This is now
1808 supported by any public key method supporting the encrypt operation. A
1809 ctrl is added to allow the public key algorithm to examine or modify
1810 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1814 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1815 a default digest type to use. In most cases this will be SHA1 but some
1816 algorithms (such as GOST) need to specify an alternative digest. The
1817 return value indicates how strong the prefernce is 1 means optional and
1818 2 is mandatory (that is it is the only supported type). Modify
1819 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1820 use the default md. Update openssl utilities to use the default digest
1821 type for signing if it is not explicitly indicated.
1824 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1825 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1826 signing method from the key type. This effectively removes the link
1827 between digests and public key types.
1830 *) Add an OID cross reference table and utility functions. Its purpose is to
1831 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1832 rsaEncryption. This will allow some of the algorithm specific hackery
1833 needed to use the correct OID to be removed.
1836 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1837 structures for PKCS7_sign(). They are now set up by the relevant public
1841 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1844 *) Add support for key derivation (agreement) in the API, DH method and
1848 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1849 public and private key formats. As a side effect these add additional
1850 command line functionality not previously available: DSA signatures can be
1851 generated and verified using pkeyutl and DH key support and generation in
1856 [Oliver Tappe <zooey@hirschkaefer.de>]
1858 *) New make target "install_html_docs" installs HTML renditions of the
1860 [Oliver Tappe <zooey@hirschkaefer.de>]
1862 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1863 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1864 support key and parameter generation and add initial key generation
1865 functionality for RSA.
1868 *) Add functions for main EVP_PKEY_method operations. The undocumented
1869 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1870 EVP_PKEY_{encrypt,decrypt}_old.
1873 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1874 key API, doesn't do much yet.
1877 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1878 public key algorithms. New option to openssl utility:
1879 "list-public-key-algorithms" to print out info.
1882 *) Implement the Supported Elliptic Curves Extension for
1883 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1886 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1887 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1890 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1891 utilities such as rsa, dsa, dsaparam etc except they process any key
1895 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1896 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1897 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1901 *) Initial support for pluggable public key ASN1.
1902 De-spaghettify the public key ASN1 handling. Move public and private
1903 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1904 algorithm specific handling to a single module within the relevant
1905 algorithm directory. Add functions to allow (near) opaque processing
1906 of public and private key structures.
1909 *) Implement the Supported Point Formats Extension for
1910 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1913 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1914 for the psk identity [hint] and the psk callback functions to the
1915 SSL_SESSION, SSL and SSL_CTX structure.
1918 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1922 SSL_CTX_use_psk_identity_hint
1923 SSL_get_psk_identity_hint
1924 SSL_get_psk_identity
1925 SSL_use_psk_identity_hint
1927 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1929 *) Add RFC 3161 compliant time stamp request creation, response generation
1930 and response verification functionality.
1931 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1933 *) Add initial support for TLS extensions, specifically for the server_name
1934 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1935 have new members for a host name. The SSL data structure has an
1936 additional member SSL_CTX *initial_ctx so that new sessions can be
1937 stored in that context to allow for session resumption, even after the
1938 SSL has been switched to a new SSL_CTX in reaction to a client's
1939 server_name extension.
1941 New functions (subject to change):
1943 SSL_get_servername()
1944 SSL_get_servername_type()
1947 New CTRL codes and macros (subject to change):
1949 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1950 - SSL_CTX_set_tlsext_servername_callback()
1951 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1952 - SSL_CTX_set_tlsext_servername_arg()
1953 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1955 openssl s_client has a new '-servername ...' option.
1957 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1958 '-key2 ...', '-servername_fatal' (subject to change). This allows
1959 testing the HostName extension for a specific single host name ('-cert'
1960 and '-key' remain fallbacks for handshakes without HostName
1961 negotiation). If the unrecogninzed_name alert has to be sent, this by
1962 default is a warning; it becomes fatal with the '-servername_fatal'
1965 [Peter Sylvester, Remy Allais, Christophe Renou]
1967 *) Whirlpool hash implementation is added.
1970 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1971 bn(64,32). Because of instruction set limitations it doesn't have
1972 any negative impact on performance. This was done mostly in order
1973 to make it possible to share assembler modules, such as bn_mul_mont
1974 implementations, between 32- and 64-bit builds without hassle.
1977 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1978 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1982 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1983 dedicated Montgomery multiplication procedure, is introduced.
1984 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1985 "64-bit" performance on certain 32-bit targets.
1988 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1989 in SSL structures. New SSL ctrl to set maximum send fragment size.
1990 Save memory by seeting the I/O buffer sizes dynamically instead of
1991 using the maximum available value.
1994 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1995 in addition to the text details.
1998 *) Very, very preliminary EXPERIMENTAL support for printing of general
1999 ASN1 structures. This currently produces rather ugly output and doesn't
2000 handle several customised structures at all.
2003 *) Integrated support for PVK file format and some related formats such
2004 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
2005 these in the 'rsa' and 'dsa' utilities.
2008 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
2011 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
2012 place for the (very old) "NETSCAPE" format certificates which are now
2013 handled using new ASN1 code equivalents.
2016 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
2017 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
2018 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
2021 *) Modify CRL distribution points extension code to print out previously
2022 unsupported fields. Enhance extension setting code to allow setting of
2026 *) Add print and set support for Issuing Distribution Point CRL extension.
2029 *) Change 'Configure' script to enable Camellia by default.
2032 Changes between 0.9.8x and 0.9.8y [5 Feb 2013]
2034 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2036 This addresses the flaw in CBC record processing discovered by
2037 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2038 at: http://www.isg.rhul.ac.uk/tls/
2040 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2041 Security Group at Royal Holloway, University of London
2042 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2043 Emilia Käsper for the initial patch.
2045 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2047 *) Return an error when checking OCSP signatures when key is NULL.
2048 This fixes a DoS attack. (CVE-2013-0166)
2051 *) Return an error when checking OCSP signatures when key is NULL.
2052 This fixes a DoS attack. (CVE-2013-0166)
2055 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2056 the right response is stapled. Also change SSL_get_certificate()
2057 so it returns the certificate actually sent.
2058 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2059 (This is a backport)
2060 [Rob Stradling <rob.stradling@comodo.com>]
2062 *) Fix possible deadlock when decoding public keys.
2065 Changes between 0.9.8w and 0.9.8x [10 May 2012]
2067 *) Sanity check record length before skipping explicit IV in DTLS
2070 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2071 fuzzing as a service testing platform.
2075 *) Initialise tkeylen properly when encrypting CMS messages.
2076 Thanks to Solar Designer of Openwall for reporting this issue.
2079 Changes between 0.9.8v and 0.9.8w [23 Apr 2012]
2081 *) The fix for CVE-2012-2110 did not take into account that the
2082 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an
2083 int in OpenSSL 0.9.8, making it still vulnerable. Fix by
2084 rejecting negative len parameter. (CVE-2012-2131)
2085 [Tomas Hoger <thoger@redhat.com>]
2087 Changes between 0.9.8u and 0.9.8v [19 Apr 2012]
2089 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2090 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2091 in CRYPTO_realloc_clean.
2093 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2094 issue and to Adam Langley <agl@chromium.org> for fixing it.
2096 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2098 Changes between 0.9.8t and 0.9.8u [12 Mar 2012]
2100 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2101 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2102 content decryption and always return the same error. Note: this attack
2103 needs on average 2^20 messages so it only affects automated senders. The
2104 old behaviour can be reenabled in the CMS code by setting the
2105 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2106 an MMA defence is not necessary.
2107 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2108 this issue. (CVE-2012-0884)
2111 *) Fix CVE-2011-4619: make sure we really are receiving a
2112 client hello before rejecting multiple SGC restarts. Thanks to
2113 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2116 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
2118 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2119 Thanks to Antonio Martin, Enterprise Secure Access Research and
2120 Development, Cisco Systems, Inc. for discovering this bug and
2121 preparing a fix. (CVE-2012-0050)
2124 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
2126 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2127 of the Vaudenay padding oracle attack on CBC mode encryption
2128 which enables an efficient plaintext recovery attack against
2129 the OpenSSL implementation of DTLS. Their attack exploits timing
2130 differences arising during decryption processing. A research
2131 paper describing this attack can be found at:
2132 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2133 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2134 Security Group at Royal Holloway, University of London
2135 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2136 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2137 for preparing the fix. (CVE-2011-4108)
2138 [Robin Seggelmann, Michael Tuexen]
2140 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
2141 [Ben Laurie, Kasper <ekasper@google.com>]
2143 *) Clear bytes used for block padding of SSL 3.0 records.
2145 [Adam Langley (Google)]
2147 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2148 Kadianakis <desnacked@gmail.com> for discovering this issue and
2149 Adam Langley for preparing the fix. (CVE-2011-4619)
2150 [Adam Langley (Google)]
2152 *) Prevent malformed RFC3779 data triggering an assertion failure.
2153 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2154 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2155 [Rob Austein <sra@hactrn.net>]
2157 *) Fix ssl_ciph.c set-up race.
2158 [Adam Langley (Google)]
2160 *) Fix spurious failures in ecdsatest.c.
2161 [Emilia Käsper (Google)]
2163 *) Fix the BIO_f_buffer() implementation (which was mixing different
2164 interpretations of the '..._len' fields).
2165 [Adam Langley (Google)]
2167 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2168 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2169 threads won't reuse the same blinding coefficients.
2171 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2172 lock to call BN_BLINDING_invert_ex, and avoids one use of
2173 BN_BLINDING_update for each BN_BLINDING structure (previously,
2174 the last update always remained unused).
2175 [Emilia Käsper (Google)]
2177 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2178 for multi-threaded use of ECDH.
2179 [Adam Langley (Google)]
2181 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2184 *) Add protection against ECDSA timing attacks as mentioned in the paper
2185 by Billy Bob Brumley and Nicola Tuveri, see:
2187 http://eprint.iacr.org/2011/232.pdf
2189 [Billy Bob Brumley and Nicola Tuveri]
2191 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
2193 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2194 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2196 *) Fix bug in string printing code: if *any* escaping is enabled we must
2197 escape the escape character (backslash) or the resulting string is
2201 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
2203 *) Disable code workaround for ancient and obsolete Netscape browsers
2204 and servers: an attacker can use it in a ciphersuite downgrade attack.
2205 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2208 *) Fixed J-PAKE implementation error, originally discovered by
2209 Sebastien Martini, further info and confirmation from Stefan
2210 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2213 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
2215 *) Fix extension code to avoid race conditions which can result in a buffer
2216 overrun vulnerability: resumed sessions must not be modified as they can
2217 be shared by multiple threads. CVE-2010-3864
2220 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
2223 *) Don't reencode certificate when calculating signature: cache and use
2224 the original encoding instead. This makes signature verification of
2225 some broken encodings work correctly.
2228 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
2229 is also one of the inputs.
2230 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
2232 *) Don't repeatedly append PBE algorithms to table if they already exist.
2233 Sort table on each new add. This effectively makes the table read only
2234 after all algorithms are added and subsequent calls to PKCS12_pbe_add
2238 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
2240 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
2243 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
2244 access or freeing data twice (CVE-2010-0742)
2245 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
2247 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
2248 common in certificates and some applications which only call
2249 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
2253 Reduce copying into .apps and .test in makevms.com
2254 Don't try to use blank CA certificate in CA.com
2255 Allow use of C files from original directories in maketests.com
2256 [Steven M. Schweda" <sms@antinode.info>]
2258 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2260 *) When rejecting SSL/TLS records due to an incorrect version number, never
2261 update s->server with a new major version number. As of
2262 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2263 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2264 the previous behavior could result in a read attempt at NULL when
2265 receiving specific incorrect SSL/TLS records once record payload
2266 protection is active. (CVE-2010-0740)
2267 [Bodo Moeller, Adam Langley <agl@chromium.org>]
2269 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
2270 could be crashed if the relevant tables were not present (e.g. chrooted).
2271 [Tomas Hoger <thoger@redhat.com>]
2273 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
2275 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
2276 [Martin Olsson, Neel Mehta]
2278 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
2279 accommodate for stack sorting, always a write lock!).
2282 *) On some versions of WIN32 Heap32Next is very slow. This can cause
2283 excessive delays in the RAND_poll(): over a minute. As a workaround
2284 include a time check in the inner Heap32Next loop too.
2287 *) The code that handled flushing of data in SSL/TLS originally used the
2288 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
2289 the problem outlined in PR#1949. The fix suggested there however can
2290 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
2291 of Apache). So instead simplify the code to flush unconditionally.
2292 This should be fine since flushing with no data to flush is a no op.
2295 *) Handle TLS versions 2.0 and later properly and correctly use the
2296 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
2297 off ancient servers have a habit of sticking around for a while...
2300 *) Modify compression code so it frees up structures without using the
2301 ex_data callbacks. This works around a problem where some applications
2302 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
2303 restarting) then use compression (e.g. SSL with compression) later.
2304 This results in significant per-connection memory leaks and
2305 has caused some security issues including CVE-2008-1678 and
2309 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
2310 change when encrypting or decrypting.
2313 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
2314 connect and renegotiate with servers which do not support RI.
2315 Until RI is more widely deployed this option is enabled by default.
2318 *) Add "missing" ssl ctrls to clear options and mode.
2321 *) If client attempts to renegotiate and doesn't support RI respond with
2322 a no_renegotiation alert as required by RFC5746. Some renegotiating
2323 TLS clients will continue a connection gracefully when they receive
2324 the alert. Unfortunately OpenSSL mishandled this alert and would hang
2325 waiting for a server hello which it will never receive. Now we treat a
2326 received no_renegotiation alert as a fatal error. This is because
2327 applications requesting a renegotiation might well expect it to succeed
2328 and would have no code in place to handle the server denying it so the
2329 only safe thing to do is to terminate the connection.
2332 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
2333 peer supports secure renegotiation and 0 otherwise. Print out peer
2334 renegotiation support in s_client/s_server.
2337 *) Replace the highly broken and deprecated SPKAC certification method with
2338 the updated NID creation version. This should correctly handle UTF8.
2341 *) Implement RFC5746. Re-enable renegotiation but require the extension
2342 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2343 turns out to be a bad idea. It has been replaced by
2344 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2345 SSL_CTX_set_options(). This is really not recommended unless you
2346 know what you are doing.
2347 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2349 *) Fixes to stateless session resumption handling. Use initial_ctx when
2350 issuing and attempting to decrypt tickets in case it has changed during
2351 servername handling. Use a non-zero length session ID when attempting
2352 stateless session resumption: this makes it possible to determine if
2353 a resumption has occurred immediately after receiving server hello
2354 (several places in OpenSSL subtly assume this) instead of later in
2358 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2359 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
2360 fixes for a few places where the return code is not checked
2362 [Julia Lawall <julia@diku.dk>]
2364 *) Add --strict-warnings option to Configure script to include devteam
2365 warnings in other configurations.
2368 *) Add support for --libdir option and LIBDIR variable in makefiles. This
2369 makes it possible to install openssl libraries in locations which
2370 have names other than "lib", for example "/usr/lib64" which some
2372 [Steve Henson, based on patch from Jeremy Utley]
2374 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
2375 X690 8.9.12 and can produce some misleading textual output of OIDs.
2376 [Steve Henson, reported by Dan Kaminsky]
2378 *) Delete MD2 from algorithm tables. This follows the recommendation in
2379 several standards that it is not used in new applications due to
2380 several cryptographic weaknesses. For binary compatibility reasons
2381 the MD2 API is still compiled in by default.
2384 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
2388 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
2389 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
2391 [Guenter <lists@gknw.net>]
2393 *) Fix the server certificate chain building code to use X509_verify_cert(),
2394 it used to have an ad-hoc builder which was unable to cope with anything
2395 other than a simple chain.
2396 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
2398 *) Don't check self signed certificate signatures in X509_verify_cert()
2399 by default (a flag can override this): it just wastes time without
2400 adding any security. As a useful side effect self signed root CAs
2401 with non-FIPS digests are now usable in FIPS mode.
2404 *) In dtls1_process_out_of_seq_message() the check if the current message
2405 is already buffered was missing. For every new message was memory
2406 allocated, allowing an attacker to perform an denial of service attack
2407 with sending out of seq handshake messages until there is no memory
2408 left. Additionally every future messege was buffered, even if the
2409 sequence number made no sense and would be part of another handshake.
2410 So only messages with sequence numbers less than 10 in advance will be
2411 buffered. (CVE-2009-1378)
2412 [Robin Seggelmann, discovered by Daniel Mentz]
2414 *) Records are buffered if they arrive with a future epoch to be
2415 processed after finishing the corresponding handshake. There is
2416 currently no limitation to this buffer allowing an attacker to perform
2417 a DOS attack with sending records with future epochs until there is no
2418 memory left. This patch adds the pqueue_size() function to detemine
2419 the size of a buffer and limits the record buffer to 100 entries.
2421 [Robin Seggelmann, discovered by Daniel Mentz]
2423 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2424 parent structure is freed. (CVE-2009-1379)
2427 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2428 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2431 [Ilya O. <vrghost@gmail.com>]
2433 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2435 *) Disable renegotiation completely - this fixes a severe security
2436 problem (CVE-2009-3555) at the cost of breaking all
2437 renegotiation. Renegotiation can be re-enabled by setting
2438 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2439 run-time. This is really not recommended unless you know what
2443 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2445 *) Don't set val to NULL when freeing up structures, it is freed up by
2446 underlying code. If sizeof(void *) > sizeof(long) this can result in
2447 zeroing past the valid field. (CVE-2009-0789)
2448 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2450 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2451 checked correctly. This would allow some invalid signed attributes to
2452 appear to verify correctly. (CVE-2009-0591)
2453 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2455 *) Reject UniversalString and BMPString types with invalid lengths. This
2456 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2457 a legal length. (CVE-2009-0590)
2460 *) Set S/MIME signing as the default purpose rather than setting it
2461 unconditionally. This allows applications to override it at the store
2465 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2466 to handle some structures.
2469 *) Improve efficiency of mem_gets: don't search whole buffer each time
2471 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2473 *) New -hex option for openssl rand.
2476 *) Print out UTF8String and NumericString when parsing ASN1.
2479 *) Support NumericString type for name components.
2482 *) Allow CC in the environment to override the automatically chosen
2483 compiler. Note that nothing is done to ensure flags work with the
2487 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2489 *) Properly check EVP_VerifyFinal() and similar return values
2491 [Ben Laurie, Bodo Moeller, Google Security Team]
2493 *) Enable TLS extensions by default.
2496 *) Allow the CHIL engine to be loaded, whether the application is
2497 multithreaded or not. (This does not release the developer from the
2498 obligation to set up the dynamic locking callbacks.)
2499 [Sander Temme <sander@temme.net>]
2501 *) Use correct exit code if there is an error in dgst command.
2502 [Steve Henson; problem pointed out by Roland Dirlewanger]
2504 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2505 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2508 *) Add experimental JPAKE support, including demo authentication in
2509 s_client and s_server.
2512 *) Set the comparison function in v3_addr_canonize().
2513 [Rob Austein <sra@hactrn.net>]
2515 *) Add support for XMPP STARTTLS in s_client.
2516 [Philip Paeps <philip@freebsd.org>]
2518 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2519 to ensure that even with this option, only ciphersuites in the
2520 server's preference list will be accepted. (Note that the option
2521 applies only when resuming a session, so the earlier behavior was
2522 just about the algorithm choice for symmetric cryptography.)
2525 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2527 *) Fix NULL pointer dereference if a DTLS server received
2528 ChangeCipherSpec as first record (CVE-2009-1386).
2531 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
2532 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2535 *) The fix in 0.9.8c that supposedly got rid of unsafe
2536 double-checked locking was incomplete for RSA blinding,
2537 addressing just one layer of what turns out to have been
2538 doubly unsafe triple-checked locking.
2540 So now fix this for real by retiring the MONT_HELPER macro
2541 in crypto/rsa/rsa_eay.c.
2543 [Bodo Moeller; problem pointed out by Marius Schilder]
2545 *) Various precautionary measures:
2547 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2549 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2550 (NB: This would require knowledge of the secret session ticket key
2551 to exploit, in which case you'd be SOL either way.)
2553 - Change bn_nist.c so that it will properly handle input BIGNUMs
2554 outside the expected range.
2556 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2559 [Neel Mehta, Bodo Moeller]
2561 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2562 the load fails. Useful for distros.
2563 [Ben Laurie and the FreeBSD team]
2565 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2568 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2571 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2573 This work was sponsored by Logica.
2576 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2577 keystores. Support for SSL/TLS client authentication too.
2578 Not compiled unless enable-capieng specified to Configure.
2580 This work was sponsored by Logica.
2583 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2584 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2585 attribute creation routines such as certifcate requests and PKCS#12
2589 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2591 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2592 handshake which could lead to a cilent crash as found using the
2593 Codenomicon TLS test suite (CVE-2008-1672)
2594 [Steve Henson, Mark Cox]
2596 *) Fix double free in TLS server name extensions which could lead to
2597 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
2600 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
2602 Clear the error queue to ensure that error entries left from
2603 older function calls do not interfere with the correct operation.
2604 [Lutz Jaenicke, Erik de Castro Lopo]
2606 *) Remove root CA certificates of commercial CAs:
2608 The OpenSSL project does not recommend any specific CA and does not
2609 have any policy with respect to including or excluding any CA.
2610 Therefore it does not make any sense to ship an arbitrary selection
2611 of root CA certificates with the OpenSSL software.
2614 *) RSA OAEP patches to fix two separate invalid memory reads.
2615 The first one involves inputs when 'lzero' is greater than
2616 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
2617 before the beginning of from). The second one involves inputs where
2618 the 'db' section contains nothing but zeroes (there is a one-byte
2619 invalid read after the end of 'db').
2620 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2622 *) Partial backport from 0.9.9-dev:
2624 Introduce bn_mul_mont (dedicated Montgomery multiplication
2625 procedure) as a candidate for BIGNUM assembler implementation.
2626 While 0.9.9-dev uses assembler for various architectures, only
2627 x86_64 is available by default here in the 0.9.8 branch, and
2628 32-bit x86 is available through a compile-time setting.
2630 To try the 32-bit x86 assembler implementation, use Configure
2631 option "enable-montasm" (which exists only for this backport).
2633 As "enable-montasm" for 32-bit x86 disclaims code stability
2634 anyway, in this constellation we activate additional code
2635 backported from 0.9.9-dev for further performance improvements,
2636 namely BN_from_montgomery_word. (To enable this otherwise,
2637 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
2639 [Andy Polyakov (backport partially by Bodo Moeller)]
2641 *) Add TLS session ticket callback. This allows an application to set
2642 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
2643 values. This is useful for key rollover for example where several key
2644 sets may exist with different names.
2647 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
2648 This was broken until now in 0.9.8 releases, such that the only way
2649 a registered ENGINE could be used (assuming it initialises
2650 successfully on the host) was to explicitly set it as the default
2651 for the relevant algorithms. This is in contradiction with 0.9.7
2652 behaviour and the documentation. With this fix, when an ENGINE is
2653 registered into a given algorithm's table of implementations, the
2654 'uptodate' flag is reset so that auto-discovery will be used next
2655 time a new context for that algorithm attempts to select an
2657 [Ian Lister (tweaked by Geoff Thorpe)]
2659 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
2660 implemention in the following ways:
2662 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
2665 Lack of BER streaming support means one pass streaming processing is
2666 only supported if data is detached: setting the streaming flag is
2667 ignored for embedded content.
2669 CMS support is disabled by default and must be explicitly enabled
2670 with the enable-cms configuration option.
2673 *) Update the GMP engine glue to do direct copies between BIGNUM and
2674 mpz_t when openssl and GMP use the same limb size. Otherwise the
2675 existing "conversion via a text string export" trick is still used.
2676 [Paul Sheer <paulsheer@gmail.com>]
2678 *) Zlib compression BIO. This is a filter BIO which compressed and
2679 uncompresses any data passed through it.
2682 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
2683 RFC3394 compatible AES key wrapping.
2686 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
2687 sets string data without copying. X509_ALGOR_set0() and
2688 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
2689 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
2690 from an X509_ATTRIBUTE structure optionally checking it occurs only
2691 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
2695 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
2696 to get the expected BN_FLG_CONSTTIME behavior.
2697 [Bodo Moeller (Google)]
2701 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
2702 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
2703 - added some more tests to do_tests.pl
2704 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
2705 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
2706 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
2707 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
2708 - various changes to netware.pl to enable gcc-cross builds on Win32
2710 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
2711 - various changes to fix missing prototype warnings
2712 - fixed x86nasm.pl to create correct asm files for NASM COFF output
2713 - added AES, WHIRLPOOL and CPUID assembler code to build files
2714 - added missing AES assembler make rules to mk1mf.pl
2715 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
2716 [Guenter Knauf <eflash@gmx.net>]
2718 *) Implement certificate status request TLS extension defined in RFC3546.
2719 A client can set the appropriate parameters and receive the encoded
2720 OCSP response via a callback. A server can query the supplied parameters
2721 and set the encoded OCSP response in the callback. Add simplified examples
2722 to s_client and s_server.
2725 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
2727 *) Fix various bugs:
2728 + Binary incompatibility of ssl_ctx_st structure
2729 + DTLS interoperation with non-compliant servers
2730 + Don't call get_session_cb() without proposed session
2731 + Fix ia64 assembler code
2732 [Andy Polyakov, Steve Henson]
2734 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
2736 *) DTLS Handshake overhaul. There were longstanding issues with
2737 OpenSSL DTLS implementation, which were making it impossible for
2738 RFC 4347 compliant client to communicate with OpenSSL server.
2739 Unfortunately just fixing these incompatibilities would "cut off"
2740 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
2741 server keeps tolerating non RFC compliant syntax. The opposite is
2742 not true, 0.9.8f client can not communicate with earlier server.
2743 This update even addresses CVE-2007-4995.
2746 *) Changes to avoid need for function casts in OpenSSL: some compilers
2747 (gcc 4.2 and later) reject their use.
2748 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
2751 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2752 RFC4507bis. The encrypted ticket format is an encrypted encoded
2753 SSL_SESSION structure, that way new session features are automatically
2756 If a client application caches session in an SSL_SESSION structure
2757 support is transparent because tickets are now stored in the encoded
2760 The SSL_CTX structure automatically generates keys for ticket
2761 protection in servers so again support should be possible
2762 with no application modification.
2764 If a client or server wishes to disable RFC4507 support then the option
2765 SSL_OP_NO_TICKET can be set.
2767 Add a TLS extension debugging callback to allow the contents of any client
2768 or server extensions to be examined.
2770 This work was sponsored by Google.
2773 *) Add initial support for TLS extensions, specifically for the server_name
2774 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2775 have new members for a host name. The SSL data structure has an
2776 additional member SSL_CTX *initial_ctx so that new sessions can be
2777 stored in that context to allow for session resumption, even after the
2778 SSL has been switched to a new SSL_CTX in reaction to a client's
2779 server_name extension.
2781 New functions (subject to change):
2783 SSL_get_servername()
2784 SSL_get_servername_type()
2787 New CTRL codes and macros (subject to change):
2789 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2790 - SSL_CTX_set_tlsext_servername_callback()
2791 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2792 - SSL_CTX_set_tlsext_servername_arg()
2793 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2795 openssl s_client has a new '-servername ...' option.
2797 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2798 '-key2 ...', '-servername_fatal' (subject to change). This allows
2799 testing the HostName extension for a specific single host name ('-cert'
2800 and '-key' remain fallbacks for handshakes without HostName
2801 negotiation). If the unrecogninzed_name alert has to be sent, this by
2802 default is a warning; it becomes fatal with the '-servername_fatal'
2805 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2807 *) Add AES and SSE2 assembly language support to VC++ build.
2810 *) Mitigate attack on final subtraction in Montgomery reduction.
2813 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2814 (which previously caused an internal error).
2817 *) Squeeze another 10% out of IGE mode when in != out.
2820 *) AES IGE mode speedup.
2821 [Dean Gaudet (Google)]
2823 *) Add the Korean symmetric 128-bit cipher SEED (see
2824 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2825 add SEED ciphersuites from RFC 4162:
2827 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2828 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2829 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2830 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2832 To minimize changes between patchlevels in the OpenSSL 0.9.8
2833 series, SEED remains excluded from compilation unless OpenSSL
2834 is configured with 'enable-seed'.
2835 [KISA, Bodo Moeller]
2837 *) Mitigate branch prediction attacks, which can be practical if a
2838 single processor is shared, allowing a spy process to extract
2839 information. For detailed background information, see
2840 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2841 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2842 and Necessary Software Countermeasures"). The core of the change
2843 are new versions BN_div_no_branch() and
2844 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2845 respectively, which are slower, but avoid the security-relevant
2846 conditional branches. These are automatically called by BN_div()
2847 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2848 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2849 remove a conditional branch.
2851 BN_FLG_CONSTTIME is the new name for the previous
2852 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2853 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2854 in the exponent causes BN_mod_exp_mont() to use the alternative
2855 implementation in BN_mod_exp_mont_consttime().) The old name
2856 remains as a deprecated alias.
2858 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2859 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2860 constant-time implementations for more than just exponentiation.
2861 Here too the old name is kept as a deprecated alias.
2863 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2864 the BN_BLINDING structure gets an independent copy of the
2865 modulus. This means that the previous "BIGNUM *m" argument to
2866 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2867 essentially becomes "const BIGNUM *m", although we can't actually
2868 change this in the header file before 0.9.9. It allows
2869 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2870 enable BN_FLG_CONSTTIME.
2872 [Matthew D Wood (Intel Corp)]
2874 *) In the SSL/TLS server implementation, be strict about session ID
2875 context matching (which matters if an application uses a single
2876 external cache for different purposes). Previously,
2877 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2878 set. This did ensure strict client verification, but meant that,
2879 with applications using a single external cache for quite
2880 different requirements, clients could circumvent ciphersuite
2881 restrictions for a given session ID context by starting a session
2882 in a different context.
2885 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2886 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2887 authentication-only ciphersuites.
2890 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2891 not complete and could lead to a possible single byte overflow
2892 (CVE-2007-5135) [Ben Laurie]
2894 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2896 *) Since AES128 and AES256 (and similarly Camellia128 and
2897 Camellia256) share a single mask bit in the logic of
2898 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2899 kludge to work properly if AES128 is available and AES256 isn't
2900 (or if Camellia128 is available and Camellia256 isn't).
2903 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2904 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2905 When a point or a seed is encoded in a BIT STRING, we need to
2906 prevent the removal of trailing zero bits to get the proper DER
2907 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2908 of a NamedBitList, for which trailing 0 bits need to be removed.)
2911 *) Have SSL/TLS server implementation tolerate "mismatched" record
2912 protocol version while receiving ClientHello even if the
2913 ClientHello is fragmented. (The server can't insist on the
2914 particular protocol version it has chosen before the ServerHello
2915 message has informed the client about his choice.)
2918 *) Add RFC 3779 support.
2919 [Rob Austein for ARIN, Ben Laurie]
2921 *) Load error codes if they are not already present instead of using a
2922 static variable. This allows them to be cleanly unloaded and reloaded.
2923 Improve header file function name parsing.
2926 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2927 or CAPABILITY handshake as required by RFCs.
2930 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2932 *) Introduce limits to prevent malicious keys being able to
2933 cause a denial of service. (CVE-2006-2940)
2934 [Steve Henson, Bodo Moeller]
2936 *) Fix ASN.1 parsing of certain invalid structures that can result
2937 in a denial of service. (CVE-2006-2937) [Steve Henson]
2939 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2940 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2942 *) Fix SSL client code which could crash if connecting to a
2943 malicious SSLv2 server. (CVE-2006-4343)
2944 [Tavis Ormandy and Will Drewry, Google Security Team]
2946 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2947 match only those. Before that, "AES256-SHA" would be interpreted
2948 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2949 the same strength classification in 0.9.7h) as we currently only
2950 have a single AES bit in the ciphersuite description bitmap.
2951 That change, however, also applied to ciphersuite strings such as
2952 "RC4-MD5" that intentionally matched multiple ciphersuites --
2953 namely, SSL 2.0 ciphersuites in addition to the more common ones
2954 from SSL 3.0/TLS 1.0.
2956 So we change the selection algorithm again: Naming an explicit
2957 ciphersuite selects this one ciphersuite, and any other similar
2958 ciphersuite (same bitmap) from *other* protocol versions.
2959 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2960 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2962 Since SSL 2.0 does not have any ciphersuites for which the
2963 128/256 bit distinction would be relevant, this works for now.
2964 The proper fix will be to use different bits for AES128 and
2965 AES256, which would have avoided the problems from the beginning;
2966 however, bits are scarce, so we can only do this in a new release
2967 (not just a patchlevel) when we can change the SSL_CIPHER
2968 definition to split the single 'unsigned long mask' bitmap into
2969 multiple values to extend the available space.
2973 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2975 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2976 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2978 *) Add AES IGE and biIGE modes.
2981 *) Change the Unix randomness entropy gathering to use poll() when
2982 possible instead of select(), since the latter has some
2983 undesirable limitations.
2984 [Darryl Miles via Richard Levitte and Bodo Moeller]
2986 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2987 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2988 cannot be implicitly activated as part of, e.g., the "AES" alias.
2989 However, please upgrade to OpenSSL 0.9.9[-dev] for
2990 non-experimental use of the ECC ciphersuites to get TLS extension
2991 support, which is required for curve and point format negotiation
2992 to avoid potential handshake problems.
2995 *) Disable rogue ciphersuites:
2997 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2998 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2999 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3001 The latter two were purportedly from
3002 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3005 Also deactivate the remaining ciphersuites from
3006 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3007 unofficial, and the ID has long expired.
3010 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3011 dual-core machines) and other potential thread-safety issues.
3014 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3015 versions), which is now available for royalty-free use
3016 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3017 Also, add Camellia TLS ciphersuites from RFC 4132.
3019 To minimize changes between patchlevels in the OpenSSL 0.9.8
3020 series, Camellia remains excluded from compilation unless OpenSSL
3021 is configured with 'enable-camellia'.
3024 *) Disable the padding bug check when compression is in use. The padding
3025 bug check assumes the first packet is of even length, this is not
3026 necessarily true if compresssion is enabled and can result in false
3027 positives causing handshake failure. The actual bug test is ancient
3028 code so it is hoped that implementations will either have fixed it by
3029 now or any which still have the bug do not support compression.
3032 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3034 *) When applying a cipher rule check to see if string match is an explicit
3035 cipher suite and only match that one cipher suite if it is.
3038 *) Link in manifests for VC++ if needed.
3039 [Austin Ziegler <halostatue@gmail.com>]
3041 *) Update support for ECC-based TLS ciphersuites according to
3042 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3043 TLS extensions, which are supported starting with the 0.9.9
3044 branch, not in the OpenSSL 0.9.8 branch).
3047 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3048 opaque EVP_CIPHER_CTX handling.
3051 *) Fixes and enhancements to zlib compression code. We now only use
3052 "zlib1.dll" and use the default __cdecl calling convention on Win32
3053 to conform with the standards mentioned here:
3054 http://www.zlib.net/DLL_FAQ.txt
3055 Static zlib linking now works on Windows and the new --with-zlib-include
3056 --with-zlib-lib options to Configure can be used to supply the location
3057 of the headers and library. Gracefully handle case where zlib library
3061 *) Several fixes and enhancements to the OID generation code. The old code
3062 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3063 handle numbers larger than ULONG_MAX, truncated printing and had a
3064 non standard OBJ_obj2txt() behaviour.
3067 *) Add support for building of engines under engine/ as shared libraries
3068 under VC++ build system.
3071 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3072 Hopefully, we will not see any false combination of paths any more.
3075 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3077 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3078 (part of SSL_OP_ALL). This option used to disable the
3079 countermeasure against man-in-the-middle protocol-version
3080 rollback in the SSL 2.0 server implementation, which is a bad
3081 idea. (CVE-2005-2969)
3083 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3084 for Information Security, National Institute of Advanced Industrial
3085 Science and Technology [AIST], Japan)]
3087 *) Add two function to clear and return the verify parameter flags.
3090 *) Keep cipherlists sorted in the source instead of sorting them at
3091 runtime, thus removing the need for a lock.
3094 *) Avoid some small subgroup attacks in Diffie-Hellman.
3095 [Nick Mathewson and Ben Laurie]
3097 *) Add functions for well-known primes.
3100 *) Extended Windows CE support.
3101 [Satoshi Nakamura and Andy Polyakov]
3103 *) Initialize SSL_METHOD structures at compile time instead of during
3104 runtime, thus removing the need for a lock.
3107 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3108 attempting to decrypt each encrypted key in turn. Add support to
3112 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3114 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3117 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3120 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3121 key into the same file any more.
3124 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3127 *) Add -utf8 command line and config file option to 'ca'.
3128 [Stefan <stf@udoma.org]
3130 *) Removed the macro des_crypt(), as it seems to conflict with some
3131 libraries. Use DES_crypt().
3134 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3135 involves renaming the source and generated shared-libs for
3136 both. The engines will accept the corrected or legacy ids
3137 ('ncipher' and '4758_cca' respectively) when binding. NB,
3138 this only applies when building 'shared'.
3139 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3141 *) Add attribute functions to EVP_PKEY structure. Modify
3142 PKCS12_create() to recognize a CSP name attribute and
3143 use it. Make -CSP option work again in pkcs12 utility.
3146 *) Add new functionality to the bn blinding code:
3147 - automatic re-creation of the BN_BLINDING parameters after
3148 a fixed number of uses (currently 32)
3149 - add new function for parameter creation
3150 - introduce flags to control the update behaviour of the
3151 BN_BLINDING parameters
3152 - hide BN_BLINDING structure
3153 Add a second BN_BLINDING slot to the RSA structure to improve
3154 performance when a single RSA object is shared among several
3158 *) Add support for DTLS.
3159 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3161 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3162 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3165 *) Remove buggy and incompletet DH cert support from
3166 ssl/ssl_rsa.c and ssl/s3_both.c
3169 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3170 the apps/openssl applications.
3173 *) Compile clean with "-Wall -Wmissing-prototypes
3174 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
3175 DEBUG_SAFESTACK must also be set.
3178 *) Change ./Configure so that certain algorithms can be disabled by default.
3179 The new counterpiece to "no-xxx" is "enable-xxx".
3181 The patented RC5 and MDC2 algorithms will now be disabled unless
3182 "enable-rc5" and "enable-mdc2", respectively, are specified.
3184 (IDEA remains enabled despite being patented. This is because IDEA
3185 is frequently required for interoperability, and there is no license
3186 fee for non-commercial use. As before, "no-idea" can be used to
3187 avoid this algorithm.)
3191 *) Add processing of proxy certificates (see RFC 3820). This work was
3192 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
3193 EGEE (Enabling Grids for E-science in Europe).
3196 *) RC4 performance overhaul on modern architectures/implementations, such
3197 as Intel P4, IA-64 and AMD64.
3200 *) New utility extract-section.pl. This can be used specify an alternative
3201 section number in a pod file instead of having to treat each file as
3202 a separate case in Makefile. This can be done by adding two lines to the
3205 =for comment openssl_section:XXX
3207 The blank line is mandatory.
3211 *) New arguments -certform, -keyform and -pass for s_client and s_server
3212 to allow alternative format key and certificate files and passphrase
3216 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
3217 update associated structures and add various utility functions.
3219 Add new policy related verify parameters, include policy checking in
3220 standard verify code. Enhance 'smime' application with extra parameters
3221 to support policy checking and print out.
3224 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
3225 Nehemiah processors. These extensions support AES encryption in hardware
3226 as well as RNG (though RNG support is currently disabled).
3227 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
3229 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
3232 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
3233 [Andy Polyakov and a number of other people]
3235 *) Improved PowerPC platform support. Most notably BIGNUM assembler
3236 implementation contributed by IBM.
3237 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
3239 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
3240 exponent rather than 'unsigned long'. There is a corresponding change to
3241 the new 'rsa_keygen' element of the RSA_METHOD structure.
3242 [Jelte Jansen, Geoff Thorpe]
3244 *) Functionality for creating the initial serial number file is now
3245 moved from CA.pl to the 'ca' utility with a new option -create_serial.
3247 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
3248 number file to 1, which is bound to cause problems. To avoid
3249 the problems while respecting compatibility between different 0.9.7
3250 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
3251 CA.pl for serial number initialization. With the new release 0.9.8,
3252 we can fix the problem directly in the 'ca' utility.)
3255 *) Reduced header interdepencies by declaring more opaque objects in
3256 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
3257 give fewer recursive includes, which could break lazy source code - so
3258 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
3259 developers should define this symbol when building and using openssl to
3260 ensure they track the recommended behaviour, interfaces, [etc], but
3261 backwards-compatible behaviour prevails when this isn't defined.
3264 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
3267 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
3268 This will generate a random key of the appropriate length based on the
3269 cipher context. The EVP_CIPHER can provide its own random key generation
3270 routine to support keys of a specific form. This is used in the des and
3271 3des routines to generate a key of the correct parity. Update S/MIME
3272 code to use new functions and hence generate correct parity DES keys.
3273 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
3274 valid (weak or incorrect parity).
3277 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
3278 as looking them up. This is useful when the verified structure may contain
3279 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
3280 present unless the new PKCS7_NO_CRL flag is asserted.
3283 *) Extend ASN1 oid configuration module. It now additionally accepts the
3286 shortName = some long name, 1.2.3.4
3289 *) Reimplemented the BN_CTX implementation. There is now no more static
3290 limitation on the number of variables it can handle nor the depth of the
3291 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
3292 information can now expand as required, and rather than having a single
3293 static array of bignums, BN_CTX now uses a linked-list of such arrays
3294 allowing it to expand on demand whilst maintaining the usefulness of
3295 BN_CTX's "bundling".
3298 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
3299 to allow all RSA operations to function using a single BN_CTX.
3302 *) Preliminary support for certificate policy evaluation and checking. This
3303 is initially intended to pass the tests outlined in "Conformance Testing
3304 of Relying Party Client Certificate Path Processing Logic" v1.07.
3307 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
3308 remained unused and not that useful. A variety of other little bignum
3309 tweaks and fixes have also been made continuing on from the audit (see
3313 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
3314 associated ASN1, EVP and SSL functions and old ASN1 macros.
3317 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
3318 and this should never fail. So the return value from the use of
3319 BN_set_word() (which can fail due to needless expansion) is now deprecated;
3320 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
3323 *) BN_CTX_get() should return zero-valued bignums, providing the same
3324 initialised value as BN_new().
3325 [Geoff Thorpe, suggested by Ulf Möller]
3327 *) Support for inhibitAnyPolicy certificate extension.
3330 *) An audit of the BIGNUM code is underway, for which debugging code is
3331 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
3332 is considered valid when processing BIGNUMs, and causes execution to
3333 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
3334 further steps are taken to deliberately pollute unused data in BIGNUM
3335 structures to try and expose faulty code further on. For now, openssl will
3336 (in its default mode of operation) continue to tolerate the inconsistent
3337 forms that it has tolerated in the past, but authors and packagers should
3338 consider trying openssl and their own applications when compiled with
3339 these debugging symbols defined. It will help highlight potential bugs in
3340 their own code, and will improve the test coverage for OpenSSL itself. At
3341 some point, these tighter rules will become openssl's default to improve
3342 maintainability, though the assert()s and other overheads will remain only
3343 in debugging configurations. See bn.h for more details.
3344 [Geoff Thorpe, Nils Larsch, Ulf Möller]
3346 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
3347 that can only be obtained through BN_CTX_new() (which implicitly
3348 initialises it). The presence of this function only made it possible
3349 to overwrite an existing structure (and cause memory leaks).
3352 *) Because of the callback-based approach for implementing LHASH as a
3353 template type, lh_insert() adds opaque objects to hash-tables and
3354 lh_doall() or lh_doall_arg() are typically used with a destructor callback
3355 to clean up those corresponding objects before destroying the hash table
3356 (and losing the object pointers). So some over-zealous constifications in
3357 LHASH have been relaxed so that lh_insert() does not take (nor store) the
3358 objects as "const" and the lh_doall[_arg] callback wrappers are not
3359 prototyped to have "const" restrictions on the object pointers they are
3360 given (and so aren't required to cast them away any more).
3363 *) The tmdiff.h API was so ugly and minimal that our own timing utility
3364 (speed) prefers to use its own implementation. The two implementations
3365 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
3366 its object type properly exposed (MS_TM) instead of casting to/from "char
3367 *". This may still change yet if someone realises MS_TM and "ms_time_***"
3368 aren't necessarily the greatest nomenclatures - but this is what was used
3369 internally to the implementation so I've used that for now.
3372 *) Ensure that deprecated functions do not get compiled when
3373 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
3374 the self-tests were still using deprecated key-generation functions so
3375 these have been updated also.
3378 *) Reorganise PKCS#7 code to separate the digest location functionality
3379 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
3380 New function PKCS7_set_digest() to set the digest type for PKCS#7
3381 digestedData type. Add additional code to correctly generate the
3382 digestedData type and add support for this type in PKCS7 initialization
3386 *) New function PKCS7_set0_type_other() this initializes a PKCS7
3387 structure of type "other".
3390 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
3391 sure the loop does correctly stop and breaking ("division by zero")
3392 modulus operations are not performed. The (pre-generated) prime
3393 table crypto/bn/bn_prime.h was already correct, but it could not be
3394 re-generated on some platforms because of the "division by zero"
3395 situation in the script.
3396 [Ralf S. Engelschall]
3398 *) Update support for ECC-based TLS ciphersuites according to
3399 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
3400 SHA-1 now is only used for "small" curves (where the
3401 representation of a field element takes up to 24 bytes); for
3402 larger curves, the field element resulting from ECDH is directly
3403 used as premaster secret.
3404 [Douglas Stebila (Sun Microsystems Laboratories)]
3406 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
3407 curve secp160r1 to the tests.
3408 [Douglas Stebila (Sun Microsystems Laboratories)]
3410 *) Add the possibility to load symbols globally with DSO.
3411 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
3413 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
3414 control of the error stack.
3417 *) Add support for STORE in ENGINE.
3420 *) Add the STORE type. The intention is to provide a common interface
3421 to certificate and key stores, be they simple file-based stores, or
3422 HSM-type store, or LDAP stores, or...
3423 NOTE: The code is currently UNTESTED and isn't really used anywhere.
3426 *) Add a generic structure called OPENSSL_ITEM. This can be used to
3427 pass a list of arguments to any function as well as provide a way
3428 for a function to pass data back to the caller.
3431 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
3432 works like BUF_strdup() but can be used to duplicate a portion of
3433 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
3437 *) Add the function sk_find_ex() which works like sk_find(), but will
3438 return an index to an element even if an exact match couldn't be
3439 found. The index is guaranteed to point at the element where the
3440 searched-for key would be inserted to preserve sorting order.
3443 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
3444 takes an extra flags argument for optional functionality. Currently,
3445 the following flags are defined:
3447 OBJ_BSEARCH_VALUE_ON_NOMATCH
3448 This one gets OBJ_bsearch_ex() to return a pointer to the first
3449 element where the comparing function returns a negative or zero
3452 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
3453 This one gets OBJ_bsearch_ex() to return a pointer to the first
3454 element where the comparing function returns zero. This is useful
3455 if there are more than one element where the comparing function
3459 *) Make it possible to create self-signed certificates with 'openssl ca'
3460 in such a way that the self-signed certificate becomes part of the
3461 CA database and uses the same mechanisms for serial number generation
3462 as all other certificate signing. The new flag '-selfsign' enables
3463 this functionality. Adapt CA.sh and CA.pl.in.
3466 *) Add functionality to check the public key of a certificate request
3467 against a given private. This is useful to check that a certificate
3468 request can be signed by that key (self-signing).
3471 *) Make it possible to have multiple active certificates with the same
3472 subject in the CA index file. This is done only if the keyword
3473 'unique_subject' is set to 'no' in the main CA section (default
3474 if 'CA_default') of the configuration file. The value is saved
3475 with the database itself in a separate index attribute file,
3476 named like the index file with '.attr' appended to the name.
3479 *) Generate muti valued AVAs using '+' notation in config files for
3483 *) Support for nameConstraints certificate extension.
3486 *) Support for policyConstraints certificate extension.
3489 *) Support for policyMappings certificate extension.
3492 *) Make sure the default DSA_METHOD implementation only uses its
3493 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3494 and change its own handlers to be NULL so as to remove unnecessary
3495 indirection. This lets alternative implementations fallback to the
3496 default implementation more easily.
3499 *) Support for directoryName in GeneralName related extensions
3503 *) Make it possible to link applications using Makefile.shared.
3504 Make that possible even when linking against static libraries!
3507 *) Support for single pass processing for S/MIME signing. This now
3508 means that S/MIME signing can be done from a pipe, in addition
3509 cleartext signing (multipart/signed type) is effectively streaming
3510 and the signed data does not need to be all held in memory.
3512 This is done with a new flag PKCS7_STREAM. When this flag is set
3513 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3514 is done after the data is output (and digests calculated) in
3515 SMIME_write_PKCS7().
3518 *) Add full support for -rpath/-R, both in shared libraries and
3519 applications, at least on the platforms where it's known how
3523 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3524 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3525 will now compute a table of multiples of the generator that
3526 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3527 faster (notably in the case of a single point multiplication,
3528 scalar * generator).
3529 [Nils Larsch, Bodo Moeller]
3531 *) IPv6 support for certificate extensions. The various extensions
3532 which use the IP:a.b.c.d can now take IPv6 addresses using the
3533 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3537 *) Added an ENGINE that implements RSA by performing private key
3538 exponentiations with the GMP library. The conversions to and from
3539 GMP's mpz_t format aren't optimised nor are any montgomery forms
3540 cached, and on x86 it appears OpenSSL's own performance has caught up.
3541 However there are likely to be other architectures where GMP could
3542 provide a boost. This ENGINE is not built in by default, but it can be
3543 specified at Configure time and should be accompanied by the necessary
3544 linker additions, eg;
3545 ./config -DOPENSSL_USE_GMP -lgmp
3548 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3549 testing availability of engines with "-t" - the old behaviour is
3550 produced by increasing the feature's verbosity with "-tt".
3553 *) ECDSA routines: under certain error conditions uninitialized BN objects
3554 could be freed. Solution: make sure initialization is performed early
3555 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3559 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3560 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3561 software implementations. For DSA and DH, parameter generation can
3562 also be overriden by providing the appropriate method callbacks.
3565 *) Change the "progress" mechanism used in key-generation and
3566 primality testing to functions that take a new BN_GENCB pointer in
3567 place of callback/argument pairs. The new API functions have "_ex"
3568 postfixes and the older functions are reimplemented as wrappers for
3569 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3570 declarations of the old functions to help (graceful) attempts to
3571 migrate to the new functions. Also, the new key-generation API
3572 functions operate on a caller-supplied key-structure and return
3573 success/failure rather than returning a key or NULL - this is to
3574 help make "keygen" another member function of RSA_METHOD etc.
3576 Example for using the new callback interface:
3578 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3582 BN_GENCB_set(&my_cb, my_callback, my_arg);
3584 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3585 /* For the meaning of a, b in calls to my_callback(), see the
3586 * documentation of the function that calls the callback.
3587 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3588 * my_callback should return 1 if it wants BN_is_prime_ex()
3589 * to continue, or 0 to stop.
3594 *) Change the ZLIB compression method to be stateful, and make it
3595 available to TLS with the number defined in
3596 draft-ietf-tls-compression-04.txt.
3599 *) Add the ASN.1 structures and functions for CertificatePair, which
3600 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
3602 CertificatePair ::= SEQUENCE {
3603 forward [0] Certificate OPTIONAL,
3604 reverse [1] Certificate OPTIONAL,
3605 -- at least one of the pair shall be present -- }
3607 Also implement the PEM functions to read and write certificate
3608 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
3610 This needed to be defined, mostly for the sake of the LDAP
3611 attribute crossCertificatePair, but may prove useful elsewhere as
3615 *) Make it possible to inhibit symlinking of shared libraries in
3616 Makefile.shared, for Cygwin's sake.
3619 *) Extend the BIGNUM API by creating a function
3620 void BN_set_negative(BIGNUM *a, int neg);
3621 and a macro that behave like
3622 int BN_is_negative(const BIGNUM *a);
3624 to avoid the need to access 'a->neg' directly in applications.
3627 *) Implement fast modular reduction for pseudo-Mersenne primes
3628 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
3629 EC_GROUP_new_curve_GFp() will now automatically use this
3631 [Nils Larsch <nla@trustcenter.de>]
3633 *) Add new lock type (CRYPTO_LOCK_BN).
3636 *) Change the ENGINE framework to automatically load engines
3637 dynamically from specific directories unless they could be
3638 found to already be built in or loaded. Move all the
3639 current engines except for the cryptodev one to a new
3641 The engines in engines/ are built as shared libraries if
3642 the "shared" options was given to ./Configure or ./config.
3643 Otherwise, they are inserted in libcrypto.a.
3644 /usr/local/ssl/engines is the default directory for dynamic
3645 engines, but that can be overriden at configure time through
3646 the usual use of --prefix and/or --openssldir, and at run
3647 time with the environment variable OPENSSL_ENGINES.
3648 [Geoff Thorpe and Richard Levitte]
3650 *) Add Makefile.shared, a helper makefile to build shared
3651 libraries. Addapt Makefile.org.
3654 *) Add version info to Win32 DLLs.
3655 [Peter 'Luna' Runestig" <peter@runestig.com>]
3657 *) Add new 'medium level' PKCS#12 API. Certificates and keys
3658 can be added using this API to created arbitrary PKCS#12
3659 files while avoiding the low level API.
3661 New options to PKCS12_create(), key or cert can be NULL and
3662 will then be omitted from the output file. The encryption
3663 algorithm NIDs can be set to -1 for no encryption, the mac
3664 iteration count can be set to 0 to omit the mac.
3666 Enhance pkcs12 utility by making the -nokeys and -nocerts
3667 options work when creating a PKCS#12 file. New option -nomac
3668 to omit the mac, NONE can be set for an encryption algorithm.
3669 New code is modified to use the enhanced PKCS12_create()
3670 instead of the low level API.
3673 *) Extend ASN1 encoder to support indefinite length constructed
3674 encoding. This can output sequences tags and octet strings in
3675 this form. Modify pk7_asn1.c to support indefinite length
3676 encoding. This is experimental and needs additional code to
3677 be useful, such as an ASN1 bio and some enhanced streaming
3680 Extend template encode functionality so that tagging is passed
3681 down to the template encoder.
3684 *) Let 'openssl req' fail if an argument to '-newkey' is not
3685 recognized instead of using RSA as a default.
3688 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
3689 As these are not official, they are not included in "ALL";
3690 the "ECCdraft" ciphersuite group alias can be used to select them.
3691 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
3693 *) Add ECDH engine support.
3694 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
3696 *) Add ECDH in new directory crypto/ecdh/.
3697 [Douglas Stebila (Sun Microsystems Laboratories)]
3699 *) Let BN_rand_range() abort with an error after 100 iterations
3700 without success (which indicates a broken PRNG).
3703 *) Change BN_mod_sqrt() so that it verifies that the input value
3704 is really the square of the return value. (Previously,
3705 BN_mod_sqrt would show GIGO behaviour.)
3708 *) Add named elliptic curves over binary fields from X9.62, SECG,
3709 and WAP/WTLS; add OIDs that were still missing.
3711 [Sheueling Chang Shantz and Douglas Stebila
3712 (Sun Microsystems Laboratories)]
3714 *) Extend the EC library for elliptic curves over binary fields
3715 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
3718 EC_GF2m_simple_method
3722 EC_GROUP_new_curve_GF2m
3723 EC_GROUP_set_curve_GF2m
3724 EC_GROUP_get_curve_GF2m
3725 EC_POINT_set_affine_coordinates_GF2m
3726 EC_POINT_get_affine_coordinates_GF2m
3727 EC_POINT_set_compressed_coordinates_GF2m
3729 Point compression for binary fields is disabled by default for
3730 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
3733 As binary polynomials are represented as BIGNUMs, various members
3734 of the EC_GROUP and EC_POINT data structures can be shared
3735 between the implementations for prime fields and binary fields;
3736 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
3737 are essentially identical to their ..._GFp counterparts.
3738 (For simplicity, the '..._GFp' prefix has been dropped from
3739 various internal method names.)
3741 An internal 'field_div' method (similar to 'field_mul' and
3742 'field_sqr') has been added; this is used only for binary fields.
3744 [Sheueling Chang Shantz and Douglas Stebila
3745 (Sun Microsystems Laboratories)]
3747 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
3748 through methods ('mul', 'precompute_mult').
3750 The generic implementations (now internally called 'ec_wNAF_mul'
3751 and 'ec_wNAF_precomputed_mult') remain the default if these
3752 methods are undefined.
3754 [Sheueling Chang Shantz and Douglas Stebila
3755 (Sun Microsystems Laboratories)]
3757 *) New function EC_GROUP_get_degree, which is defined through
3758 EC_METHOD. For curves over prime fields, this returns the bit
3759 length of the modulus.
3761 [Sheueling Chang Shantz and Douglas Stebila
3762 (Sun Microsystems Laboratories)]
3764 *) New functions EC_GROUP_dup, EC_POINT_dup.
3765 (These simply call ..._new and ..._copy).
3767 [Sheueling Chang Shantz and Douglas Stebila
3768 (Sun Microsystems Laboratories)]
3770 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3771 Polynomials are represented as BIGNUMs (where the sign bit is not
3772 used) in the following functions [macros]:
3775 BN_GF2m_sub [= BN_GF2m_add]
3776 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3777 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3778 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3780 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3781 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3782 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3783 BN_GF2m_cmp [= BN_ucmp]
3785 (Note that only the 'mod' functions are actually for fields GF(2^m).
3786 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3788 For some functions, an the irreducible polynomial defining a
3789 field can be given as an 'unsigned int[]' with strictly
3790 decreasing elements giving the indices of those bits that are set;
3791 i.e., p[] represents the polynomial
3792 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3794 p[0] > p[1] > ... > p[k] = 0.
3795 This applies to the following functions:
3800 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3801 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3803 BN_GF2m_mod_sqrt_arr
3804 BN_GF2m_mod_solve_quad_arr
3808 Conversion can be performed by the following functions:
3813 bntest.c has additional tests for binary polynomial arithmetic.
3815 Two implementations for BN_GF2m_mod_div() are available.
3816 The default algorithm simply uses BN_GF2m_mod_inv() and
3817 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3818 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3819 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3821 [Sheueling Chang Shantz and Douglas Stebila
3822 (Sun Microsystems Laboratories)]
3824 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3825 functionality is disabled at compile-time.
3826 [Douglas Stebila <douglas.stebila@sun.com>]
3828 *) Change default behaviour of 'openssl asn1parse' so that more
3829 information is visible when viewing, e.g., a certificate:
3831 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3832 mode the content of non-printable OCTET STRINGs is output in a
3833 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3834 avoid the appearance of a printable string.
3835 [Nils Larsch <nla@trustcenter.de>]
3837 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3839 EC_GROUP_set_asn1_flag()
3840 EC_GROUP_get_asn1_flag()
3841 EC_GROUP_set_point_conversion_form()
3842 EC_GROUP_get_point_conversion_form()
3843 These control ASN1 encoding details:
3844 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3845 has been set to OPENSSL_EC_NAMED_CURVE.
3846 - Points are encoded in uncompressed form by default; options for
3847 asn1_for are as for point2oct, namely
3848 POINT_CONVERSION_COMPRESSED
3849 POINT_CONVERSION_UNCOMPRESSED
3850 POINT_CONVERSION_HYBRID
3852 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3855 EC_GROUP_get0_seed()
3856 EC_GROUP_get_seed_len()
3857 This is used only for ASN1 purposes (so far).
3858 [Nils Larsch <nla@trustcenter.de>]
3860 *) Add 'field_type' member to EC_METHOD, which holds the NID
3861 of the appropriate field type OID. The new function
3862 EC_METHOD_get_field_type() returns this value.
3863 [Nils Larsch <nla@trustcenter.de>]
3868 EC_POINT_point2hex()
3869 EC_POINT_hex2point()
3870 providing useful interfaces to EC_POINT_point2oct() and
3871 EC_POINT_oct2point().
3872 [Nils Larsch <nla@trustcenter.de>]
3874 *) Change internals of the EC library so that the functions
3875 EC_GROUP_set_generator()
3876 EC_GROUP_get_generator()
3877 EC_GROUP_get_order()
3878 EC_GROUP_get_cofactor()
3879 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3880 to methods, which would lead to unnecessary code duplication when
3881 adding different types of curves.
3882 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3884 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3885 arithmetic, and such that modified wNAFs are generated
3886 (which avoid length expansion in many cases).
3889 *) Add a function EC_GROUP_check_discriminant() (defined via
3890 EC_METHOD) that verifies that the curve discriminant is non-zero.
3892 Add a function EC_GROUP_check() that makes some sanity tests
3893 on a EC_GROUP, its generator and order. This includes
3894 EC_GROUP_check_discriminant().
3895 [Nils Larsch <nla@trustcenter.de>]
3897 *) Add ECDSA in new directory crypto/ecdsa/.
3899 Add applications 'openssl ecparam' and 'openssl ecdsa'
3900 (these are based on 'openssl dsaparam' and 'openssl dsa').
3902 ECDSA support is also included in various other files across the
3903 library. Most notably,
3904 - 'openssl req' now has a '-newkey ecdsa:file' option;
3905 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3906 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3907 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3908 them suitable for ECDSA where domain parameters must be
3909 extracted before the specific public key;
3910 - ECDSA engine support has been added.
3911 [Nils Larsch <nla@trustcenter.de>]
3913 *) Include some named elliptic curves, and add OIDs from X9.62,
3914 SECG, and WAP/WTLS. Each curve can be obtained from the new
3916 EC_GROUP_new_by_curve_name(),
3917 and the list of available named curves can be obtained with
3918 EC_get_builtin_curves().
3919 Also add a 'curve_name' member to EC_GROUP objects, which can be
3921 EC_GROUP_set_curve_name()
3922 EC_GROUP_get_curve_name()
3923 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3925 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3926 was actually never needed) and in BN_mul(). The removal in BN_mul()
3927 required a small change in bn_mul_part_recursive() and the addition
3928 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3929 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3930 bn_sub_words() and bn_add_words() except they take arrays with
3934 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3936 *) Cleanse PEM buffers before freeing them since they may contain
3938 [Benjamin Bennett <ben@psc.edu>]
3940 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3941 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3942 authentication-only ciphersuites.
3945 *) Since AES128 and AES256 share a single mask bit in the logic of
3946 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3947 kludge to work properly if AES128 is available and AES256 isn't.
3950 *) Expand security boundary to match 1.1.1 module.
3953 *) Remove redundant features: hash file source, editing of test vectors
3954 modify fipsld to use external fips_premain.c signature.
3957 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3958 run algorithm test programs.
3961 *) Make algorithm test programs more tolerant of whitespace.
3964 *) Have SSL/TLS server implementation tolerate "mismatched" record
3965 protocol version while receiving ClientHello even if the
3966 ClientHello is fragmented. (The server can't insist on the
3967 particular protocol version it has chosen before the ServerHello
3968 message has informed the client about his choice.)
3971 *) Load error codes if they are not already present instead of using a
3972 static variable. This allows them to be cleanly unloaded and reloaded.
3975 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3977 *) Introduce limits to prevent malicious keys being able to
3978 cause a denial of service. (CVE-2006-2940)
3979 [Steve Henson, Bodo Moeller]
3981 *) Fix ASN.1 parsing of certain invalid structures that can result
3982 in a denial of service. (CVE-2006-2937) [Steve Henson]
3984 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3985 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3987 *) Fix SSL client code which could crash if connecting to a
3988 malicious SSLv2 server. (CVE-2006-4343)
3989 [Tavis Ormandy and Will Drewry, Google Security Team]
3991 *) Change ciphersuite string processing so that an explicit
3992 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3993 will no longer include "AES128-SHA"), and any other similar
3994 ciphersuite (same bitmap) from *other* protocol versions (so that
3995 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3996 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3997 changes from 0.9.8b and 0.9.8d.
4000 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4002 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4003 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4005 *) Change the Unix randomness entropy gathering to use poll() when
4006 possible instead of select(), since the latter has some
4007 undesirable limitations.
4008 [Darryl Miles via Richard Levitte and Bodo Moeller]
4010 *) Disable rogue ciphersuites:
4012 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4013 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4014 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4016 The latter two were purportedly from
4017 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4020 Also deactive the remaining ciphersuites from
4021 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4022 unofficial, and the ID has long expired.
4025 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4026 dual-core machines) and other potential thread-safety issues.
4029 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4031 *) Adapt fipsld and the build system to link against the validated FIPS
4032 module in FIPS mode.
4035 *) Fixes for VC++ 2005 build under Windows.
4038 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4039 from a Windows bash shell such as MSYS. It is autodetected from the
4040 "config" script when run from a VC++ environment. Modify standard VC++
4041 build to use fipscanister.o from the GNU make build.
4044 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4046 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4047 The value now differs depending on if you build for FIPS or not.
4048 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4049 safely run with a non-FIPSed libcrypto, as it may crash because of
4050 the difference induced by this change.
4053 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4055 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4056 (part of SSL_OP_ALL). This option used to disable the
4057 countermeasure against man-in-the-middle protocol-version
4058 rollback in the SSL 2.0 server implementation, which is a bad
4059 idea. (CVE-2005-2969)
4061 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4062 for Information Security, National Institute of Advanced Industrial
4063 Science and Technology [AIST], Japan)]
4065 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4066 mainly for FIPS compliance and not fully integrated at this stage.
4069 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4070 the exponentiation using a fixed-length exponent. (Otherwise,
4071 the information leaked through timing could expose the secret key
4072 after many signatures; cf. Bleichenbacher's attack on DSA with
4076 *) Make a new fixed-window mod_exp implementation the default for
4077 RSA, DSA, and DH private-key operations so that the sequence of
4078 squares and multiplies and the memory access pattern are
4079 independent of the particular secret key. This will mitigate
4080 cache-timing and potential related attacks.
4082 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4083 and this is automatically used by BN_mod_exp_mont() if the new flag
4084 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4085 will use this BN flag for private exponents unless the flag
4086 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4087 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4089 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4091 *) Change the client implementation for SSLv23_method() and
4092 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4093 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4094 (Previously, the SSL 2.0 backwards compatible Client Hello
4095 message format would be used even with SSL_OP_NO_SSLv2.)
4098 *) Add support for smime-type MIME parameter in S/MIME messages which some
4102 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4103 a threadsafe manner. Modify rsa code to use new function and add calls
4104 to dsa and dh code (which had race conditions before).
4107 *) Include the fixed error library code in the C error file definitions
4108 instead of fixing them up at runtime. This keeps the error code
4109 structures constant.
4112 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4114 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4117 *) Fixes for newer kerberos headers. NB: the casts are needed because
4118 the 'length' field is signed on one version and unsigned on another
4119 with no (?) obvious way to tell the difference, without these VC++
4120 complains. Also the "definition" of FAR (blank) is no longer included
4121 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4122 some needed definitions.
4125 *) Undo Cygwin change.
4128 *) Added support for proxy certificates according to RFC 3820.
4129 Because they may be a security thread to unaware applications,
4130 they must be explicitely allowed in run-time. See
4131 docs/HOWTO/proxy_certificates.txt for further information.
4134 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4136 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4137 server and client random values. Previously
4138 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4139 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4141 This change has negligible security impact because:
4143 1. Server and client random values still have 24 bytes of pseudo random
4146 2. Server and client random values are sent in the clear in the initial
4149 3. The master secret is derived using the premaster secret (48 bytes in
4150 size for static RSA ciphersuites) as well as client server and random
4153 The OpenSSL team would like to thank the UK NISCC for bringing this issue
4156 [Stephen Henson, reported by UK NISCC]
4158 *) Use Windows randomness collection on Cygwin.
4161 *) Fix hang in EGD/PRNGD query when communication socket is closed
4162 prematurely by EGD/PRNGD.
4163 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
4165 *) Prompt for pass phrases when appropriate for PKCS12 input format.
4168 *) Back-port of selected performance improvements from development
4169 branch, as well as improved support for PowerPC platforms.
4172 *) Add lots of checks for memory allocation failure, error codes to indicate
4173 failure and freeing up memory if a failure occurs.
4174 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
4176 *) Add new -passin argument to dgst.
4179 *) Perform some character comparisons of different types in X509_NAME_cmp:
4180 this is needed for some certificates that reencode DNs into UTF8Strings
4181 (in violation of RFC3280) and can't or wont issue name rollover
4185 *) Make an explicit check during certificate validation to see that
4186 the CA setting in each certificate on the chain is correct. As a
4187 side effect always do the following basic checks on extensions,
4188 not just when there's an associated purpose to the check:
4190 - if there is an unhandled critical extension (unless the user
4191 has chosen to ignore this fault)
4192 - if the path length has been exceeded (if one is set at all)
4193 - that certain extensions fit the associated purpose (if one has
4197 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
4199 *) Avoid a race condition when CRLs are checked in a multi threaded
4200 environment. This would happen due to the reordering of the revoked
4201 entries during signature checking and serial number lookup. Now the
4202 encoding is cached and the serial number sort performed under a lock.
4203 Add new STACK function sk_is_sorted().
4206 *) Add Delta CRL to the extension code.
4209 *) Various fixes to s3_pkt.c so alerts are sent properly.
4210 [David Holmes <d.holmes@f5.com>]
4212 *) Reduce the chances of duplicate issuer name and serial numbers (in
4213 violation of RFC3280) using the OpenSSL certificate creation utilities.
4214 This is done by creating a random 64 bit value for the initial serial
4215 number when a serial number file is created or when a self signed
4216 certificate is created using 'openssl req -x509'. The initial serial
4217 number file is created using 'openssl x509 -next_serial' in CA.pl
4218 rather than being initialized to 1.
4221 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
4223 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4224 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4225 [Joe Orton, Steve Henson]
4227 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
4229 [Joe Orton, Steve Henson]
4231 *) Make it possible to have multiple active certificates with the same
4232 subject in the CA index file. This is done only if the keyword
4233 'unique_subject' is set to 'no' in the main CA section (default
4234 if 'CA_default') of the configuration file. The value is saved
4235 with the database itself in a separate index attribute file,
4236 named like the index file with '.attr' appended to the name.
4239 *) X509 verify fixes. Disable broken certificate workarounds when
4240 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
4241 keyUsage extension present. Don't accept CRLs with unhandled critical
4242 extensions: since verify currently doesn't process CRL extensions this
4243 rejects a CRL with *any* critical extensions. Add new verify error codes
4247 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
4248 A clarification of RFC2560 will require the use of OCTET STRINGs and
4249 some implementations cannot handle the current raw format. Since OpenSSL
4250 copies and compares OCSP nonces as opaque blobs without any attempt at
4251 parsing them this should not create any compatibility issues.
4254 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
4255 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
4256 this HMAC (and other) operations are several times slower than OpenSSL
4260 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
4261 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4263 *) Use the correct content when signing type "other".
4266 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
4268 *) Fix various bugs revealed by running the NISCC test suite:
4270 Stop out of bounds reads in the ASN1 code when presented with
4271 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4273 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
4275 If verify callback ignores invalid public key errors don't try to check
4276 certificate signature with the NULL public key.
4280 *) New -ignore_err option in ocsp application to stop the server
4281 exiting on the first error in a request.
4284 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4285 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4289 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4290 extra data after the compression methods not only for TLS 1.0
4291 but also for SSL 3.0 (as required by the specification).
4292 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4294 *) Change X509_certificate_type() to mark the key as exported/exportable
4295 when it's 512 *bits* long, not 512 bytes.
4298 *) Change AES_cbc_encrypt() so it outputs exact multiple of
4299 blocks during encryption.
4302 *) Various fixes to base64 BIO and non blocking I/O. On write
4303 flushes were not handled properly if the BIO retried. On read
4304 data was not being buffered properly and had various logic bugs.
4305 This also affects blocking I/O when the data being decoded is a
4309 *) Various S/MIME bugfixes and compatibility changes:
4310 output correct application/pkcs7 MIME type if
4311 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
4312 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
4313 of files as .eml work). Correctly handle very long lines in MIME
4317 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
4319 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4320 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4321 a protocol version number mismatch like a decryption error
4322 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4325 *) Turn on RSA blinding by default in the default implementation
4326 to avoid a timing attack. Applications that don't want it can call
4327 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4328 They would be ill-advised to do so in most cases.
4329 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4331 *) Change RSA blinding code so that it works when the PRNG is not
4332 seeded (in this case, the secret RSA exponent is abused as
4333 an unpredictable seed -- if it is not unpredictable, there
4334 is no point in blinding anyway). Make RSA blinding thread-safe
4335 by remembering the creator's thread ID in rsa->blinding and
4336 having all other threads use local one-time blinding factors
4337 (this requires more computation than sharing rsa->blinding, but
4338 avoids excessive locking; and if an RSA object is not shared
4339 between threads, blinding will still be very fast).
4342 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
4343 ENGINE as defaults for all supported algorithms irrespective of
4344 the 'flags' parameter. 'flags' is now honoured, so applications
4345 should make sure they are passing it correctly.
4348 *) Target "mingw" now allows native Windows code to be generated in
4349 the Cygwin environment as well as with the MinGW compiler.
4352 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
4354 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4355 via timing by performing a MAC computation even if incorrrect
4356 block cipher padding has been found. This is a countermeasure
4357 against active attacks where the attacker has to distinguish
4358 between bad padding and a MAC verification error. (CVE-2003-0078)
4360 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4361 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4362 Martin Vuagnoux (EPFL, Ilion)]
4364 *) Make the no-err option work as intended. The intention with no-err
4365 is not to have the whole error stack handling routines removed from
4366 libcrypto, it's only intended to remove all the function name and
4367 reason texts, thereby removing some of the footprint that may not
4368 be interesting if those errors aren't displayed anyway.
4370 NOTE: it's still possible for any application or module to have it's
4371 own set of error texts inserted. The routines are there, just not
4372 used by default when no-err is given.
4375 *) Add support for FreeBSD on IA64.
4376 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
4378 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
4379 Kerberos function mit_des_cbc_cksum(). Before this change,
4380 the value returned by DES_cbc_cksum() was like the one from
4381 mit_des_cbc_cksum(), except the bytes were swapped.
4382 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
4384 *) Allow an application to disable the automatic SSL chain building.
4385 Before this a rather primitive chain build was always performed in
4386 ssl3_output_cert_chain(): an application had no way to send the
4387 correct chain if the automatic operation produced an incorrect result.
4389 Now the chain builder is disabled if either:
4391 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
4393 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
4395 The reasoning behind this is that an application would not want the
4396 auto chain building to take place if extra chain certificates are
4397 present and it might also want a means of sending no additional
4398 certificates (for example the chain has two certificates and the
4402 *) Add the possibility to build without the ENGINE framework.
4403 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4405 *) Under Win32 gmtime() can return NULL: check return value in
4406 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
4409 *) DSA routines: under certain error conditions uninitialized BN objects
4410 could be freed. Solution: make sure initialization is performed early
4411 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
4412 Nils Larsch <nla@trustcenter.de> via PR#459)
4415 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
4416 checked on reconnect on the client side, therefore session resumption
4417 could still fail with a "ssl session id is different" error. This
4418 behaviour is masked when SSL_OP_ALL is used due to
4419 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
4420 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4421 followup to PR #377.
4424 *) IA-32 assembler support enhancements: unified ELF targets, support
4425 for SCO/Caldera platforms, fix for Cygwin shared build.
4428 *) Add support for FreeBSD on sparc64. As a consequence, support for
4429 FreeBSD on non-x86 processors is separate from x86 processors on
4430 the config script, much like the NetBSD support.
4431 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
4433 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
4435 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
4438 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
4439 code (06) was taken as the first octet of the session ID and the last
4440 octet was ignored consequently. As a result SSLv2 client side session
4441 caching could not have worked due to the session ID mismatch between
4443 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4447 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
4448 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
4452 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
4453 seems that in spite of existing for more than a year, many application
4454 author have done nothing to provide the necessary callbacks, which
4455 means that this particular engine will not work properly anywhere.
4456 This is a very unfortunate situation which forces us, in the name
4457 of usability, to give the hw_ncipher.c a static lock, which is part
4459 NOTE: This is for the 0.9.7 series ONLY. This hack will never
4460 appear in 0.9.8 or later. We EXPECT application authors to have
4461 dealt properly with this when 0.9.8 is released (unless we actually
4462 make such changes in the libcrypto locking code that changes will
4463 have to be made anyway).
4466 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
4467 octets have been read, EOF or an error occurs. Without this change
4468 some truncated ASN1 structures will not produce an error.
4471 *) Disable Heimdal support, since it hasn't been fully implemented.
4472 Still give the possibility to force the use of Heimdal, but with
4473 warnings and a request that patches get sent to openssl-dev.
4476 *) Add the VC-CE target, introduce the WINCE sysname, and add
4477 INSTALL.WCE and appropriate conditionals to make it build.
4478 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4480 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4481 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4482 edit numbers of the version.
4483 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4485 *) Introduce safe string copy and catenation functions
4486 (BUF_strlcpy() and BUF_strlcat()).
4487 [Ben Laurie (CHATS) and Richard Levitte]
4489 *) Avoid using fixed-size buffers for one-line DNs.
4490 [Ben Laurie (CHATS)]
4492 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4493 resizing buffers containing secrets, and use where appropriate.
4494 [Ben Laurie (CHATS)]
4496 *) Avoid using fixed size buffers for configuration file location.
4497 [Ben Laurie (CHATS)]
4499 *) Avoid filename truncation for various CA files.
4500 [Ben Laurie (CHATS)]
4502 *) Use sizeof in preference to magic numbers.
4503 [Ben Laurie (CHATS)]
4505 *) Avoid filename truncation in cert requests.
4506 [Ben Laurie (CHATS)]
4508 *) Add assertions to check for (supposedly impossible) buffer
4510 [Ben Laurie (CHATS)]
4512 *) Don't cache truncated DNS entries in the local cache (this could
4513 potentially lead to a spoofing attack).
4514 [Ben Laurie (CHATS)]
4516 *) Fix various buffers to be large enough for hex/decimal
4517 representations in a platform independent manner.
4518 [Ben Laurie (CHATS)]
4520 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4521 resizing buffers containing secrets, and use where appropriate.
4522 [Ben Laurie (CHATS)]
4524 *) Add BIO_indent() to avoid much slightly worrying code to do
4526 [Ben Laurie (CHATS)]
4528 *) Convert sprintf()/BIO_puts() to BIO_printf().
4529 [Ben Laurie (CHATS)]
4531 *) buffer_gets() could terminate with the buffer only half
4533 [Ben Laurie (CHATS)]
4535 *) Add assertions to prevent user-supplied crypto functions from
4536 overflowing internal buffers by having large block sizes, etc.
4537 [Ben Laurie (CHATS)]
4539 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4541 [Ben Laurie (CHATS)]
4543 *) Eliminate unused copy of key in RC4.
4544 [Ben Laurie (CHATS)]
4546 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4547 [Ben Laurie (CHATS)]
4549 *) Fix off-by-one error in EGD path.
4550 [Ben Laurie (CHATS)]
4552 *) If RANDFILE path is too long, ignore instead of truncating.
4553 [Ben Laurie (CHATS)]
4555 *) Eliminate unused and incorrectly sized X.509 structure
4557 [Ben Laurie (CHATS)]
4559 *) Eliminate unused and dangerous function knumber().
4560 [Ben Laurie (CHATS)]
4562 *) Eliminate unused and dangerous structure, KSSL_ERR.
4563 [Ben Laurie (CHATS)]
4565 *) Protect against overlong session ID context length in an encoded
4566 session object. Since these are local, this does not appear to be
4568 [Ben Laurie (CHATS)]
4570 *) Change from security patch (see 0.9.6e below) that did not affect
4571 the 0.9.6 release series:
4573 Remote buffer overflow in SSL3 protocol - an attacker could
4574 supply an oversized master key in Kerberos-enabled versions.
4576 [Ben Laurie (CHATS)]
4578 *) Change the SSL kerb5 codes to match RFC 2712.
4581 *) Make -nameopt work fully for req and add -reqopt switch.
4582 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4584 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4585 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4587 *) Make sure tests can be performed even if the corresponding algorithms
4588 have been removed entirely. This was also the last step to make
4589 OpenSSL compilable with DJGPP under all reasonable conditions.
4590 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
4592 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
4593 to allow version independent disabling of normally unselected ciphers,
4594 which may be activated as a side-effect of selecting a single cipher.
4596 (E.g., cipher list string "RSA" enables ciphersuites that are left
4597 out of "ALL" because they do not provide symmetric encryption.
4598 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
4599 [Lutz Jaenicke, Bodo Moeller]
4601 *) Add appropriate support for separate platform-dependent build
4602 directories. The recommended way to make a platform-dependent
4603 build directory is the following (tested on Linux), maybe with
4606 # Place yourself outside of the OpenSSL source tree. In
4607 # this example, the environment variable OPENSSL_SOURCE
4608 # is assumed to contain the absolute OpenSSL source directory.
4609 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
4610 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
4611 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
4612 mkdir -p `dirname $F`
4613 ln -s $OPENSSL_SOURCE/$F $F
4616 To be absolutely sure not to disturb the source tree, a "make clean"
4617 is a good thing. If it isn't successfull, don't worry about it,
4618 it probably means the source directory is very clean.
4621 *) Make sure any ENGINE control commands make local copies of string
4622 pointers passed to them whenever necessary. Otherwise it is possible
4623 the caller may have overwritten (or deallocated) the original string
4624 data when a later ENGINE operation tries to use the stored values.
4625 [Götz Babin-Ebell <babinebell@trustcenter.de>]
4627 *) Improve diagnostics in file reading and command-line digests.
4628 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
4630 *) Add AES modes CFB and OFB to the object database. Correct an
4631 error in AES-CFB decryption.
4634 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
4635 allows existing EVP_CIPHER_CTX structures to be reused after
4636 calling EVP_*Final(). This behaviour is used by encryption
4637 BIOs and some applications. This has the side effect that
4638 applications must explicitly clean up cipher contexts with
4639 EVP_CIPHER_CTX_cleanup() or they will leak memory.
4642 *) Check the values of dna and dnb in bn_mul_recursive before calling
4643 bn_mul_comba (a non zero value means the a or b arrays do not contain
4644 n2 elements) and fallback to bn_mul_normal if either is not zero.
4647 *) Fix escaping of non-ASCII characters when using the -subj option
4648 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
4651 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
4652 form for "surname", serialNumber has no short form.
4653 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
4654 therefore remove "mail" short name for "internet 7".
4655 The OID for unique identifiers in X509 certificates is
4656 x500UniqueIdentifier, not uniqueIdentifier.
4657 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
4660 *) Add an "init" command to the ENGINE config module and auto initialize
4661 ENGINEs. Without any "init" command the ENGINE will be initialized
4662 after all ctrl commands have been executed on it. If init=1 the
4663 ENGINE is initailized at that point (ctrls before that point are run
4664 on the uninitialized ENGINE and after on the initialized one). If
4665 init=0 then the ENGINE will not be iniatialized at all.
4668 *) Fix the 'app_verify_callback' interface so that the user-defined
4669 argument is actually passed to the callback: In the
4670 SSL_CTX_set_cert_verify_callback() prototype, the callback
4671 declaration has been changed from
4674 int (*cb)(X509_STORE_CTX *,void *);
4675 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
4676 i=s->ctx->app_verify_callback(&ctx)
4677 has been changed into
4678 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
4680 To update applications using SSL_CTX_set_cert_verify_callback(),
4681 a dummy argument can be added to their callback functions.
4682 [D. K. Smetters <smetters@parc.xerox.com>]
4684 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
4685 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
4687 *) Add and OPENSSL_LOAD_CONF define which will cause
4688 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
4689 This allows older applications to transparently support certain
4690 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
4691 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
4692 load the config file and OPENSSL_add_all_algorithms_conf() which will
4693 always load it have also been added.
4696 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
4697 Adjust NIDs and EVP layer.
4698 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4700 *) Config modules support in openssl utility.
4702 Most commands now load modules from the config file,
4703 though in a few (such as version) this isn't done
4704 because it couldn't be used for anything.
4706 In the case of ca and req the config file used is
4707 the same as the utility itself: that is the -config
4708 command line option can be used to specify an
4712 *) Move default behaviour from OPENSSL_config(). If appname is NULL
4713 use "openssl_conf" if filename is NULL use default openssl config file.
4716 *) Add an argument to OPENSSL_config() to allow the use of an alternative
4717 config section name. Add a new flag to tolerate a missing config file
4718 and move code to CONF_modules_load_file().
4721 *) Support for crypto accelerator cards from Accelerated Encryption
4722 Processing, www.aep.ie. (Use engine 'aep')
4723 The support was copied from 0.9.6c [engine] and adapted/corrected
4724 to work with the new engine framework.
4725 [AEP Inc. and Richard Levitte]
4727 *) Support for SureWare crypto accelerator cards from Baltimore
4728 Technologies. (Use engine 'sureware')
4729 The support was copied from 0.9.6c [engine] and adapted
4730 to work with the new engine framework.
4733 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
4734 make the newer ENGINE framework commands for the CHIL engine work.
4735 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
4737 *) Make it possible to produce shared libraries on ReliantUNIX.
4738 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
4740 *) Add the configuration target debug-linux-ppro.
4741 Make 'openssl rsa' use the general key loading routines
4742 implemented in apps.c, and make those routines able to
4743 handle the key format FORMAT_NETSCAPE and the variant
4745 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4747 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4748 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4750 *) Add -keyform to rsautl, and document -engine.
4751 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
4753 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
4754 BIO_R_NO_SUCH_FILE error code rather than the generic
4755 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
4758 *) Add new functions
4760 ERR_peek_last_error_line
4761 ERR_peek_last_error_line_data.
4762 These are similar to
4765 ERR_peek_error_line_data,
4766 but report on the latest error recorded rather than the first one
4767 still in the error queue.
4768 [Ben Laurie, Bodo Moeller]
4770 *) default_algorithms option in ENGINE config module. This allows things
4772 default_algorithms = ALL
4773 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4776 *) Prelminary ENGINE config module.
4779 *) New experimental application configuration code.
4782 *) Change the AES code to follow the same name structure as all other
4783 symmetric ciphers, and behave the same way. Move everything to
4784 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4785 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4787 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4788 [Ben Laurie and Theo de Raadt]
4790 *) Add option to output public keys in req command.
4791 [Massimiliano Pala madwolf@openca.org]
4793 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4794 (up to about 10% better than before for P-192 and P-224).
4797 *) New functions/macros
4799 SSL_CTX_set_msg_callback(ctx, cb)
4800 SSL_CTX_set_msg_callback_arg(ctx, arg)
4801 SSL_set_msg_callback(ssl, cb)
4802 SSL_set_msg_callback_arg(ssl, arg)
4804 to request calling a callback function
4806 void cb(int write_p, int version, int content_type,
4807 const void *buf, size_t len, SSL *ssl, void *arg)
4809 whenever a protocol message has been completely received
4810 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4811 protocol version according to which the SSL library interprets
4812 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4813 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4814 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4815 specification (change_cipher_spec(20), alert(21), handshake(22)).
4816 'buf' and 'len' point to the actual message, 'ssl' to the
4817 SSL object, and 'arg' is the application-defined value set by
4818 SSL[_CTX]_set_msg_callback_arg().
4820 'openssl s_client' and 'openssl s_server' have new '-msg' options
4821 to enable a callback that displays all protocol messages.
4824 *) Change the shared library support so shared libraries are built as
4825 soon as the corresponding static library is finished, and thereby get
4826 openssl and the test programs linked against the shared library.
4827 This still only happens when the keyword "shard" has been given to
4828 the configuration scripts.
4830 NOTE: shared library support is still an experimental thing, and
4831 backward binary compatibility is still not guaranteed.
4832 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4834 *) Add support for Subject Information Access extension.
4835 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4837 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4838 additional bytes when new memory had to be allocated, not just
4839 when reusing an existing buffer.
4842 *) New command line and configuration option 'utf8' for the req command.
4843 This allows field values to be specified as UTF8 strings.
4846 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4847 runs for the former and machine-readable output for the latter.
4850 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4851 of the e-mail address in the DN (i.e., it will go into a certificate
4852 extension only). The new configuration file option 'email_in_dn = no'
4853 has the same effect.
4854 [Massimiliano Pala madwolf@openca.org]
4856 *) Change all functions with names starting with des_ to be starting
4857 with DES_ instead. Add wrappers that are compatible with libdes,
4858 but are named _ossl_old_des_*. Finally, add macros that map the
4859 des_* symbols to the corresponding _ossl_old_des_* if libdes
4860 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4861 desired, the des_* symbols will be mapped to DES_*, with one
4864 Since we provide two compatibility mappings, the user needs to
4865 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4866 compatibility is desired. The default (i.e., when that macro
4867 isn't defined) is OpenSSL 0.9.6c compatibility.
4869 There are also macros that enable and disable the support of old
4870 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4871 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4872 are defined, the default will apply: to support the old des routines.
4874 In either case, one must include openssl/des.h to get the correct
4875 definitions. Do not try to just include openssl/des_old.h, that
4878 NOTE: This is a major break of an old API into a new one. Software
4879 authors are encouraged to switch to the DES_ style functions. Some
4880 time in the future, des_old.h and the libdes compatibility functions
4881 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4882 default), and then completely removed.
4885 *) Test for certificates which contain unsupported critical extensions.
4886 If such a certificate is found during a verify operation it is
4887 rejected by default: this behaviour can be overridden by either
4888 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4889 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4890 X509_supported_extension() has also been added which returns 1 if a
4891 particular extension is supported.
4894 *) Modify the behaviour of EVP cipher functions in similar way to digests
4895 to retain compatibility with existing code.
4898 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4899 compatibility with existing code. In particular the 'ctx' parameter does
4900 not have to be to be initialized before the call to EVP_DigestInit() and
4901 it is tidied up after a call to EVP_DigestFinal(). New function
4902 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4903 EVP_MD_CTX_copy() changed to not require the destination to be
4904 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4905 requires the destination to be valid.
4907 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4908 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4911 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4912 so that complete 'Handshake' protocol structures are kept in memory
4913 instead of overwriting 'msg_type' and 'length' with 'body' data.
4916 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4917 [Massimo Santin via Richard Levitte]
4919 *) Major restructuring to the underlying ENGINE code. This includes
4920 reduction of linker bloat, separation of pure "ENGINE" manipulation
4921 (initialisation, etc) from functionality dealing with implementations
4922 of specific crypto iterfaces. This change also introduces integrated
4923 support for symmetric ciphers and digest implementations - so ENGINEs
4924 can now accelerate these by providing EVP_CIPHER and EVP_MD
4925 implementations of their own. This is detailed in crypto/engine/README
4926 as it couldn't be adequately described here. However, there are a few
4927 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4928 were changed in the original introduction of ENGINE code have now
4929 reverted back - the hooking from this code to ENGINE is now a good
4930 deal more passive and at run-time, operations deal directly with
4931 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4932 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4933 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4934 they were not being used by the framework as there is no concept of a
4935 BIGNUM_METHOD and they could not be generalised to the new
4936 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4937 ENGINE_cpy() has been removed as it cannot be consistently defined in
4941 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4944 *) Change mkdef.pl to sort symbols that get the same entry number,
4945 and make sure the automatically generated functions ERR_load_*
4946 become part of libeay.num as well.
4949 *) New function SSL_renegotiate_pending(). This returns true once
4950 renegotiation has been requested (either SSL_renegotiate() call
4951 or HelloRequest/ClientHello receveived from the peer) and becomes
4952 false once a handshake has been completed.
4953 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4954 sends a HelloRequest, but does not ensure that a handshake takes
4955 place. SSL_renegotiate_pending() is useful for checking if the
4956 client has followed the request.)
4959 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4960 By default, clients may request session resumption even during
4961 renegotiation (if session ID contexts permit); with this option,
4962 session resumption is possible only in the first handshake.
4964 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4965 more bits available for options that should not be part of
4966 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4969 *) Add some demos for certificate and certificate request creation.
4972 *) Make maximum certificate chain size accepted from the peer application
4973 settable (SSL*_get/set_max_cert_list()), as proposed by
4974 "Douglas E. Engert" <deengert@anl.gov>.
4977 *) Add support for shared libraries for Unixware-7
4978 (Boyd Lynn Gerber <gerberb@zenez.com>).
4981 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4982 be done prior to destruction. Use this to unload error strings from
4983 ENGINEs that load their own error strings. NB: This adds two new API
4984 functions to "get" and "set" this destroy handler in an ENGINE.
4987 *) Alter all existing ENGINE implementations (except "openssl" and
4988 "openbsd") to dynamically instantiate their own error strings. This
4989 makes them more flexible to be built both as statically-linked ENGINEs
4990 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4991 Also, add stub code to each that makes building them as self-contained
4992 shared-libraries easier (see README.ENGINE).
4995 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4996 implementations into applications that are completely implemented in
4997 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4998 commands that can be used to configure what shared-library to load and
4999 to control aspects of the way it is handled. Also, made an update to
5000 the README.ENGINE file that brings its information up-to-date and
5001 provides some information and instructions on the "dynamic" ENGINE
5002 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5005 *) Make it possible to unload ranges of ERR strings with a new
5006 "ERR_unload_strings" function.
5009 *) Add a copy() function to EVP_MD.
5012 *) Make EVP_MD routines take a context pointer instead of just the
5013 md_data void pointer.
5016 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5017 that the digest can only process a single chunk of data
5018 (typically because it is provided by a piece of
5019 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5020 is only going to provide a single chunk of data, and hence the
5021 framework needn't accumulate the data for oneshot drivers.
5024 *) As with "ERR", make it possible to replace the underlying "ex_data"
5025 functions. This change also alters the storage and management of global
5026 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5027 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5028 index counters. The API functions that use this state have been changed
5029 to take a "class_index" rather than pointers to the class's local STACK
5030 and counter, and there is now an API function to dynamically create new
5031 classes. This centralisation allows us to (a) plug a lot of the
5032 thread-safety problems that existed, and (b) makes it possible to clean
5033 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5034 such data would previously have always leaked in application code and
5035 workarounds were in place to make the memory debugging turn a blind eye
5036 to it. Application code that doesn't use this new function will still
5037 leak as before, but their memory debugging output will announce it now
5038 rather than letting it slide.
5040 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5041 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5042 has a return value to indicate success or failure.
5045 *) Make it possible to replace the underlying "ERR" functions such that the
5046 global state (2 LHASH tables and 2 locks) is only used by the "default"
5047 implementation. This change also adds two functions to "get" and "set"
5048 the implementation prior to it being automatically set the first time
5049 any other ERR function takes place. Ie. an application can call "get",
5050 pass the return value to a module it has just loaded, and that module
5051 can call its own "set" function using that value. This means the
5052 module's "ERR" operations will use (and modify) the error state in the
5053 application and not in its own statically linked copy of OpenSSL code.
5056 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5057 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5058 the operation, and provides a more encapsulated way for external code
5059 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5060 to use these functions rather than manually incrementing the counts.
5062 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5065 *) Add EVP test program.
5068 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5071 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5072 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5073 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5074 These allow a CRL to be built without having to access X509_CRL fields
5075 directly. Modify 'ca' application to use new functions.
5078 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5079 bug workarounds. Rollback attack detection is a security feature.
5080 The problem will only arise on OpenSSL servers when TLSv1 is not
5081 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5082 Software authors not wanting to support TLSv1 will have special reasons
5083 for their choice and can explicitly enable this option.
5084 [Bodo Moeller, Lutz Jaenicke]
5086 *) Rationalise EVP so it can be extended: don't include a union of
5087 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5088 (similar to those existing for EVP_CIPHER_CTX).
5093 EVP_MD_CTX_init(&md); /* new function call */
5094 EVP_DigestInit(&md, EVP_sha1());
5095 EVP_DigestUpdate(&md, in, len);
5096 EVP_DigestFinal(&md, out, NULL);
5097 EVP_MD_CTX_cleanup(&md); /* new function call */
5101 *) Make DES key schedule conform to the usual scheme, as well as
5102 correcting its structure. This means that calls to DES functions
5103 now have to pass a pointer to a des_key_schedule instead of a
5104 plain des_key_schedule (which was actually always a pointer
5107 des_key_schedule ks;
5109 des_set_key_checked(..., &ks);
5110 des_ncbc_encrypt(..., &ks, ...);
5112 (Note that a later change renames 'des_...' into 'DES_...'.)
5115 *) Initial reduction of linker bloat: the use of some functions, such as
5116 PEM causes large amounts of unused functions to be linked in due to
5117 poor organisation. For example pem_all.c contains every PEM function
5118 which has a knock on effect of linking in large amounts of (unused)
5119 ASN1 code. Grouping together similar functions and splitting unrelated
5120 functions prevents this.
5123 *) Cleanup of EVP macros.
5126 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5127 correct _ecb suffix.
5130 *) Add initial OCSP responder support to ocsp application. The
5131 revocation information is handled using the text based index
5132 use by the ca application. The responder can either handle
5133 requests generated internally, supplied in files (for example
5134 via a CGI script) or using an internal minimal server.
5137 *) Add configuration choices to get zlib compression for TLS.
5140 *) Changes to Kerberos SSL for RFC 2712 compliance:
5141 1. Implemented real KerberosWrapper, instead of just using
5142 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5143 2. Implemented optional authenticator field of KerberosWrapper.
5145 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5146 and authenticator structs; see crypto/krb5/.
5148 Generalized Kerberos calls to support multiple Kerberos libraries.
5149 [Vern Staats <staatsvr@asc.hpc.mil>,
5150 Jeffrey Altman <jaltman@columbia.edu>
5151 via Richard Levitte]
5153 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
5154 already does with RSA. testdsa.h now has 'priv_key/pub_key'
5155 values for each of the key sizes rather than having just
5156 parameters (and 'speed' generating keys each time).
5159 *) Speed up EVP routines.
5162 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
5163 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
5164 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
5165 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
5167 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
5168 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
5169 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
5172 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
5174 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
5177 *) Added the OS2-EMX target.
5178 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
5180 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
5181 to support NCONF routines in extension code. New function CONF_set_nconf()
5182 to allow functions which take an NCONF to also handle the old LHASH
5183 structure: this means that the old CONF compatible routines can be
5184 retained (in particular wrt extensions) without having to duplicate the
5185 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
5188 *) Enhance the general user interface with mechanisms for inner control
5189 and with possibilities to have yes/no kind of prompts.
5192 *) Change all calls to low level digest routines in the library and
5193 applications to use EVP. Add missing calls to HMAC_cleanup() and
5194 don't assume HMAC_CTX can be copied using memcpy().
5195 [Verdon Walker <VWalker@novell.com>, Steve Henson]
5197 *) Add the possibility to control engines through control names but with
5198 arbitrary arguments instead of just a string.
5199 Change the key loaders to take a UI_METHOD instead of a callback
5200 function pointer. NOTE: this breaks binary compatibility with earlier
5201 versions of OpenSSL [engine].
5202 Adapt the nCipher code for these new conditions and add a card insertion
5206 *) Enhance the general user interface with mechanisms to better support
5207 dialog box interfaces, application-defined prompts, the possibility
5208 to use defaults (for example default passwords from somewhere else)
5209 and interrupts/cancellations.
5212 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
5213 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
5216 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
5217 tidy up some unnecessarily weird code in 'sk_new()').
5218 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
5220 *) Change the key loading routines for ENGINEs to use the same kind
5221 callback (pem_password_cb) as all other routines that need this
5225 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
5226 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
5227 than this minimum value is recommended.
5230 *) New random seeder for OpenVMS, using the system process statistics
5231 that are easily reachable.
5234 *) Windows apparently can't transparently handle global
5235 variables defined in DLLs. Initialisations such as:
5237 const ASN1_ITEM *it = &ASN1_INTEGER_it;
5239 wont compile. This is used by the any applications that need to
5240 declare their own ASN1 modules. This was fixed by adding the option
5241 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
5242 needed for static libraries under Win32.
5245 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
5246 setting of purpose and trust fields. New X509_STORE trust and
5247 purpose functions and tidy up setting in other SSL functions.
5250 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
5251 structure. These are inherited by X509_STORE_CTX when it is
5252 initialised. This allows various defaults to be set in the
5253 X509_STORE structure (such as flags for CRL checking and custom
5254 purpose or trust settings) for functions which only use X509_STORE_CTX
5255 internally such as S/MIME.
5257 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
5258 trust settings if they are not set in X509_STORE. This allows X509_STORE
5259 purposes and trust (in S/MIME for example) to override any set by default.
5261 Add command line options for CRL checking to smime, s_client and s_server
5265 *) Initial CRL based revocation checking. If the CRL checking flag(s)
5266 are set then the CRL is looked up in the X509_STORE structure and
5267 its validity and signature checked, then if the certificate is found
5268 in the CRL the verify fails with a revoked error.
5270 Various new CRL related callbacks added to X509_STORE_CTX structure.
5272 Command line options added to 'verify' application to support this.
5274 This needs some additional work, such as being able to handle multiple
5275 CRLs with different times, extension based lookup (rather than just
5276 by subject name) and ultimately more complete V2 CRL extension
5280 *) Add a general user interface API (crypto/ui/). This is designed
5281 to replace things like des_read_password and friends (backward
5282 compatibility functions using this new API are provided).
5283 The purpose is to remove prompting functions from the DES code
5284 section as well as provide for prompting through dialog boxes in
5285 a window system and the like.
5288 *) Add "ex_data" support to ENGINE so implementations can add state at a
5289 per-structure level rather than having to store it globally.
5292 *) Make it possible for ENGINE structures to be copied when retrieved by
5293 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
5294 This causes the "original" ENGINE structure to act like a template,
5295 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
5296 operational state can be localised to each ENGINE structure, despite the
5297 fact they all share the same "methods". New ENGINE structures returned in
5298 this case have no functional references and the return value is the single
5299 structural reference. This matches the single structural reference returned
5300 by ENGINE_by_id() normally, when it is incremented on the pre-existing
5304 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
5305 needs to match any other type at all we need to manually clear the
5309 *) Changes to the "openssl engine" utility to include;
5310 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
5311 about an ENGINE's available control commands.
5312 - executing control commands from command line arguments using the
5313 '-pre' and '-post' switches. '-post' is only used if '-t' is
5314 specified and the ENGINE is successfully initialised. The syntax for
5315 the individual commands are colon-separated, for example;
5316 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
5319 *) New dynamic control command support for ENGINEs. ENGINEs can now
5320 declare their own commands (numbers), names (strings), descriptions,
5321 and input types for run-time discovery by calling applications. A
5322 subset of these commands are implicitly classed as "executable"
5323 depending on their input type, and only these can be invoked through
5324 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
5325 can be based on user input, config files, etc). The distinction is
5326 that "executable" commands cannot return anything other than a boolean
5327 result and can only support numeric or string input, whereas some
5328 discoverable commands may only be for direct use through
5329 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
5330 pointers, or other custom uses. The "executable" commands are to
5331 support parameterisations of ENGINE behaviour that can be
5332 unambiguously defined by ENGINEs and used consistently across any
5333 OpenSSL-based application. Commands have been added to all the
5334 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
5335 control over shared-library paths without source code alterations.
5338 *) Changed all ENGINE implementations to dynamically allocate their
5339 ENGINEs rather than declaring them statically. Apart from this being
5340 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
5341 this also allows the implementations to compile without using the
5342 internal engine_int.h header.
5345 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
5346 'const' value. Any code that should be able to modify a RAND_METHOD
5347 should already have non-const pointers to it (ie. they should only
5348 modify their own ones).
5351 *) Made a variety of little tweaks to the ENGINE code.
5352 - "atalla" and "ubsec" string definitions were moved from header files
5353 to C code. "nuron" string definitions were placed in variables
5354 rather than hard-coded - allowing parameterisation of these values
5355 later on via ctrl() commands.
5356 - Removed unused "#if 0"'d code.
5357 - Fixed engine list iteration code so it uses ENGINE_free() to release
5358 structural references.
5359 - Constified the RAND_METHOD element of ENGINE structures.
5360 - Constified various get/set functions as appropriate and added
5361 missing functions (including a catch-all ENGINE_cpy that duplicates
5362 all ENGINE values onto a new ENGINE except reference counts/state).
5363 - Removed NULL parameter checks in get/set functions. Setting a method
5364 or function to NULL is a way of cancelling out a previously set
5365 value. Passing a NULL ENGINE parameter is just plain stupid anyway
5366 and doesn't justify the extra error symbols and code.
5367 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
5368 flags from engine_int.h to engine.h.
5369 - Changed prototypes for ENGINE handler functions (init(), finish(),
5370 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
5373 *) Implement binary inversion algorithm for BN_mod_inverse in addition
5374 to the algorithm using long division. The binary algorithm can be
5375 used only if the modulus is odd. On 32-bit systems, it is faster
5376 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
5377 roughly 5-15% for 256-bit moduli), so we use it only for moduli
5378 up to 450 bits. In 64-bit environments, the binary algorithm
5379 appears to be advantageous for much longer moduli; here we use it
5380 for moduli up to 2048 bits.
5383 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
5384 could not support the combine flag in choice fields.
5387 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
5388 extensions from a certificate request to the certificate.
5391 *) Allow multiple 'certopt' and 'nameopt' options to be separated
5392 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
5393 file: this allows the display of the certificate about to be
5394 signed to be customised, to allow certain fields to be included
5395 or excluded and extension details. The old system didn't display
5396 multicharacter strings properly, omitted fields not in the policy
5397 and couldn't display additional details such as extensions.
5400 *) Function EC_POINTs_mul for multiple scalar multiplication
5401 of an arbitrary number of elliptic curve points
5402 \sum scalars[i]*points[i],
5403 optionally including the generator defined for the EC_GROUP:
5404 scalar*generator + \sum scalars[i]*points[i].
5406 EC_POINT_mul is a simple wrapper function for the typical case
5407 that the point list has just one item (besides the optional
5411 *) First EC_METHODs for curves over GF(p):
5413 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
5414 operations and provides various method functions that can also
5415 operate with faster implementations of modular arithmetic.
5417 EC_GFp_mont_method() reuses most functions that are part of
5418 EC_GFp_simple_method, but uses Montgomery arithmetic.
5420 [Bodo Moeller; point addition and point doubling
5421 implementation directly derived from source code provided by
5422 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
5424 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
5425 crypto/ec/ec_lib.c):
5427 Curves are EC_GROUP objects (with an optional group generator)
5428 based on EC_METHODs that are built into the library.
5430 Points are EC_POINT objects based on EC_GROUP objects.
5432 Most of the framework would be able to handle curves over arbitrary
5433 finite fields, but as there are no obvious types for fields other
5434 than GF(p), some functions are limited to that for now.
5437 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
5438 that the file contains a complete HTTP response.
5441 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
5442 change the def and num file printf format specifier from "%-40sXXX"
5443 to "%-39s XXX". The latter will always guarantee a space after the
5444 field while the former will cause them to run together if the field
5445 is 40 of more characters long.
5448 *) Constify the cipher and digest 'method' functions and structures
5449 and modify related functions to take constant EVP_MD and EVP_CIPHER
5453 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
5454 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
5457 *) Modify EVP_Digest*() routines so they now return values. Although the
5458 internal software routines can never fail additional hardware versions
5462 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
5464 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
5465 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
5468 ERR_R_NESTED_ASN1_ERROR
5470 ERR_R_MISSING_ASN1_EOS
5471 were 4 .. 9, conflicting with
5472 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5474 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5475 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5477 Add new error code 'ERR_R_INTERNAL_ERROR'.
5480 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5484 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5485 sets the subject name for a new request or supersedes the
5486 subject name in a given request. Formats that can be parsed are
5487 'CN=Some Name, OU=myOU, C=IT'
5489 'CN=Some Name/OU=myOU/C=IT'.
5491 Add options '-batch' and '-verbose' to 'openssl req'.
5492 [Massimiliano Pala <madwolf@hackmasters.net>]
5494 *) Introduce the possibility to access global variables through
5495 functions on platform were that's the best way to handle exporting
5496 global variables in shared libraries. To enable this functionality,
5497 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5498 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5499 is normally done by Configure or something similar).
5501 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5502 in the source file (foo.c) like this:
5504 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5505 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5507 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5508 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5510 OPENSSL_DECLARE_GLOBAL(int,foo);
5511 #define foo OPENSSL_GLOBAL_REF(foo)
5512 OPENSSL_DECLARE_GLOBAL(double,bar);
5513 #define bar OPENSSL_GLOBAL_REF(bar)
5515 The #defines are very important, and therefore so is including the
5516 header file everywhere where the defined globals are used.
5518 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5519 of ASN.1 items, but that structure is a bit different.
5521 The largest change is in util/mkdef.pl which has been enhanced with
5522 better and easier to understand logic to choose which symbols should
5523 go into the Windows .def files as well as a number of fixes and code
5524 cleanup (among others, algorithm keywords are now sorted
5525 lexicographically to avoid constant rewrites).
5528 *) In BN_div() keep a copy of the sign of 'num' before writing the
5529 result to 'rm' because if rm==num the value will be overwritten
5530 and produce the wrong result if 'num' is negative: this caused
5531 problems with BN_mod() and BN_nnmod().
5534 *) Function OCSP_request_verify(). This checks the signature on an
5535 OCSP request and verifies the signer certificate. The signer
5536 certificate is just checked for a generic purpose and OCSP request
5540 *) Add OCSP_check_validity() function to check the validity of OCSP
5541 responses. OCSP responses are prepared in real time and may only
5542 be a few seconds old. Simply checking that the current time lies
5543 between thisUpdate and nextUpdate max reject otherwise valid responses
5544 caused by either OCSP responder or client clock inaccuracy. Instead
5545 we allow thisUpdate and nextUpdate to fall within a certain period of
5546 the current time. The age of the response can also optionally be
5547 checked. Two new options -validity_period and -status_age added to
5551 *) If signature or public key algorithm is unrecognized print out its
5552 OID rather that just UNKNOWN.
5555 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5556 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5557 ID to be generated from the issuer certificate alone which can then be
5558 passed to OCSP_id_issuer_cmp().
5561 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5562 ASN1 modules to export functions returning ASN1_ITEM pointers
5563 instead of the ASN1_ITEM structures themselves. This adds several
5564 new macros which allow the underlying ASN1 function/structure to
5565 be accessed transparently. As a result code should not use ASN1_ITEM
5566 references directly (such as &X509_it) but instead use the relevant
5567 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5568 use of the new ASN1 code on platforms where exporting structures
5569 is problematical (for example in shared libraries) but exporting
5570 functions returning pointers to structures is not.
5573 *) Add support for overriding the generation of SSL/TLS session IDs.
5574 These callbacks can be registered either in an SSL_CTX or per SSL.
5575 The purpose of this is to allow applications to control, if they wish,
5576 the arbitrary values chosen for use as session IDs, particularly as it
5577 can be useful for session caching in multiple-server environments. A
5578 command-line switch for testing this (and any client code that wishes
5579 to use such a feature) has been added to "s_server".
5580 [Geoff Thorpe, Lutz Jaenicke]
5582 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5583 of the form '#if defined(...) || defined(...) || ...' and
5584 '#if !defined(...) && !defined(...) && ...'. This also avoids
5585 the growing number of special cases it was previously handling.
5588 *) Make all configuration macros available for application by making
5589 sure they are available in opensslconf.h, by giving them names starting
5590 with "OPENSSL_" to avoid conflicts with other packages and by making
5591 sure e_os2.h will cover all platform-specific cases together with
5593 Additionally, it is now possible to define configuration/platform-
5594 specific names (called "system identities"). In the C code, these
5595 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
5596 macro with the name beginning with "OPENSSL_SYS_", which is determined
5597 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
5601 *) New option -set_serial to 'req' and 'x509' this allows the serial
5602 number to use to be specified on the command line. Previously self
5603 signed certificates were hard coded with serial number 0 and the
5604 CA options of 'x509' had to use a serial number in a file which was
5608 *) New options to 'ca' utility to support V2 CRL entry extensions.
5609 Currently CRL reason, invalidity date and hold instruction are
5610 supported. Add new CRL extensions to V3 code and some new objects.
5613 *) New function EVP_CIPHER_CTX_set_padding() this is used to
5614 disable standard block padding (aka PKCS#5 padding) in the EVP
5615 API, which was previously mandatory. This means that the data is
5616 not padded in any way and so the total length much be a multiple
5617 of the block size, otherwise an error occurs.
5620 *) Initial (incomplete) OCSP SSL support.
5623 *) New function OCSP_parse_url(). This splits up a URL into its host,
5624 port and path components: primarily to parse OCSP URLs. New -url
5625 option to ocsp utility.
5628 *) New nonce behavior. The return value of OCSP_check_nonce() now
5629 reflects the various checks performed. Applications can decide
5630 whether to tolerate certain situations such as an absent nonce
5631 in a response when one was present in a request: the ocsp application
5632 just prints out a warning. New function OCSP_add1_basic_nonce()
5633 this is to allow responders to include a nonce in a response even if
5634 the request is nonce-less.
5637 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
5638 skipped when using openssl x509 multiple times on a single input file,
5639 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
5642 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
5643 set string type: to handle setting ASN1_TIME structures. Fix ca
5644 utility to correctly initialize revocation date of CRLs.
5647 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
5648 the clients preferred ciphersuites and rather use its own preferences.
5649 Should help to work around M$ SGC (Server Gated Cryptography) bug in
5650 Internet Explorer by ensuring unchanged hash method during stepup.
5651 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
5654 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
5655 to aes and add a new 'exist' option to print out symbols that don't
5659 *) Additional options to ocsp utility to allow flags to be set and
5660 additional certificates supplied.
5663 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
5664 OCSP client a number of certificate to only verify the response
5668 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
5669 handle the new API. Currently only ECB, CBC modes supported. Add new
5672 Add TLS AES ciphersuites as described in RFC3268, "Advanced
5673 Encryption Standard (AES) Ciphersuites for Transport Layer
5674 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
5675 not enabled by default and were not part of the "ALL" ciphersuite
5676 alias because they were not yet official; they could be
5677 explicitly requested by specifying the "AESdraft" ciphersuite
5678 group alias. In the final release of OpenSSL 0.9.7, the group
5679 alias is called "AES" and is part of "ALL".)
5680 [Ben Laurie, Steve Henson, Bodo Moeller]
5682 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
5683 request to response.
5686 *) Functions for OCSP responders. OCSP_request_onereq_count(),
5687 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
5688 extract information from a certificate request. OCSP_response_create()
5689 creates a response and optionally adds a basic response structure.
5690 OCSP_basic_add1_status() adds a complete single response to a basic
5691 response and returns the OCSP_SINGLERESP structure just added (to allow
5692 extensions to be included for example). OCSP_basic_add1_cert() adds a
5693 certificate to a basic response and OCSP_basic_sign() signs a basic
5694 response with various flags. New helper functions ASN1_TIME_check()
5695 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
5696 (converts ASN1_TIME to GeneralizedTime).
5699 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
5700 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
5701 structure from a certificate. X509_pubkey_digest() digests the public_key
5702 contents: this is used in various key identifiers.
5705 *) Make sk_sort() tolerate a NULL argument.
5706 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
5708 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
5709 passed by the function are trusted implicitly. If any of them signed the
5710 response then it is assumed to be valid and is not verified.
5713 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
5714 to data. This was previously part of the PKCS7 ASN1 code. This
5715 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
5716 [Steve Henson, reported by Kenneth R. Robinette
5717 <support@securenetterm.com>]
5719 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
5720 routines: without these tracing memory leaks is very painful.
5721 Fix leaks in PKCS12 and PKCS7 routines.
5724 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
5725 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
5726 effectively meant GeneralizedTime would never be used. Now it
5727 is initialised to -1 but X509_time_adj() now has to check the value
5728 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
5729 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
5730 [Steve Henson, reported by Kenneth R. Robinette
5731 <support@securenetterm.com>]
5733 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
5734 result in a zero length in the ASN1_INTEGER structure which was
5735 not consistent with the structure when d2i_ASN1_INTEGER() was used
5736 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
5737 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
5738 where it did not print out a minus for negative ASN1_INTEGER.
5741 *) Add summary printout to ocsp utility. The various functions which
5742 convert status values to strings have been renamed to:
5743 OCSP_response_status_str(), OCSP_cert_status_str() and
5744 OCSP_crl_reason_str() and are no longer static. New options
5745 to verify nonce values and to disable verification. OCSP response
5746 printout format cleaned up.
5749 *) Add additional OCSP certificate checks. These are those specified
5750 in RFC2560. This consists of two separate checks: the CA of the
5751 certificate being checked must either be the OCSP signer certificate
5752 or the issuer of the OCSP signer certificate. In the latter case the
5753 OCSP signer certificate must contain the OCSP signing extended key
5754 usage. This check is performed by attempting to match the OCSP
5755 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
5756 in the OCSP_CERTID structures of the response.
5759 *) Initial OCSP certificate verification added to OCSP_basic_verify()
5760 and related routines. This uses the standard OpenSSL certificate
5761 verify routines to perform initial checks (just CA validity) and
5762 to obtain the certificate chain. Then additional checks will be
5763 performed on the chain. Currently the root CA is checked to see
5764 if it is explicitly trusted for OCSP signing. This is used to set
5765 a root CA as a global signing root: that is any certificate that
5766 chains to that CA is an acceptable OCSP signing certificate.
5769 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5770 extensions from a separate configuration file.
5771 As when reading extensions from the main configuration file,
5772 the '-extensions ...' option may be used for specifying the
5774 [Massimiliano Pala <madwolf@comune.modena.it>]
5776 *) New OCSP utility. Allows OCSP requests to be generated or
5777 read. The request can be sent to a responder and the output
5778 parsed, outputed or printed in text form. Not complete yet:
5779 still needs to check the OCSP response validity.
5782 *) New subcommands for 'openssl ca':
5783 'openssl ca -status <serial>' prints the status of the cert with
5784 the given serial number (according to the index file).
5785 'openssl ca -updatedb' updates the expiry status of certificates
5787 [Massimiliano Pala <madwolf@comune.modena.it>]
5789 *) New '-newreq-nodes' command option to CA.pl. This is like
5790 '-newreq', but calls 'openssl req' with the '-nodes' option
5791 so that the resulting key is not encrypted.
5792 [Damien Miller <djm@mindrot.org>]
5794 *) New configuration for the GNU Hurd.
5795 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5797 *) Initial code to implement OCSP basic response verify. This
5798 is currently incomplete. Currently just finds the signer's
5799 certificate and verifies the signature on the response.
5802 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5803 value of OPENSSLDIR. This is available via the new '-d' option
5804 to 'openssl version', and is also included in 'openssl version -a'.
5807 *) Allowing defining memory allocation callbacks that will be given
5808 file name and line number information in additional arguments
5809 (a const char* and an int). The basic functionality remains, as
5810 well as the original possibility to just replace malloc(),
5811 realloc() and free() by functions that do not know about these
5812 additional arguments. To register and find out the current
5813 settings for extended allocation functions, the following
5814 functions are provided:
5816 CRYPTO_set_mem_ex_functions
5817 CRYPTO_set_locked_mem_ex_functions
5818 CRYPTO_get_mem_ex_functions
5819 CRYPTO_get_locked_mem_ex_functions
5821 These work the same way as CRYPTO_set_mem_functions and friends.
5822 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5823 extended allocation function is enabled.
5824 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5825 a conventional allocation function is enabled.
5826 [Richard Levitte, Bodo Moeller]
5828 *) Finish off removing the remaining LHASH function pointer casts.
5829 There should no longer be any prototype-casting required when using
5830 the LHASH abstraction, and any casts that remain are "bugs". See
5831 the callback types and macros at the head of lhash.h for details
5832 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5835 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5836 If /dev/[u]random devices are not available or do not return enough
5837 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5839 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5840 /etc/entropy will be queried once each in this sequence, quering stops
5841 when enough entropy was collected without querying more sockets.
5844 *) Change the Unix RAND_poll() variant to be able to poll several
5845 random devices, as specified by DEVRANDOM, until a sufficient amount
5846 of data has been collected. We spend at most 10 ms on each file
5847 (select timeout) and read in non-blocking mode. DEVRANDOM now
5848 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5849 (previously it was just the string "/dev/urandom"), so on typical
5850 platforms the 10 ms delay will never occur.
5851 Also separate out the Unix variant to its own file, rand_unix.c.
5852 For VMS, there's a currently-empty rand_vms.c.
5855 *) Move OCSP client related routines to ocsp_cl.c. These
5856 provide utility functions which an application needing
5857 to issue a request to an OCSP responder and analyse the
5858 response will typically need: as opposed to those which an
5859 OCSP responder itself would need which will be added later.
5861 OCSP_request_sign() signs an OCSP request with an API similar
5862 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5863 response. OCSP_response_get1_basic() extracts basic response
5864 from response. OCSP_resp_find_status(): finds and extracts status
5865 information from an OCSP_CERTID structure (which will be created
5866 when the request structure is built). These are built from lower
5867 level functions which work on OCSP_SINGLERESP structures but
5868 wont normally be used unless the application wishes to examine
5869 extensions in the OCSP response for example.
5871 Replace nonce routines with a pair of functions.
5872 OCSP_request_add1_nonce() adds a nonce value and optionally
5873 generates a random value. OCSP_check_nonce() checks the
5874 validity of the nonce in an OCSP response.
5877 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5878 This doesn't copy the supplied OCSP_CERTID and avoids the
5879 need to free up the newly created id. Change return type
5880 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5881 This can then be used to add extensions to the request.
5882 Deleted OCSP_request_new(), since most of its functionality
5883 is now in OCSP_REQUEST_new() (and the case insensitive name
5884 clash) apart from the ability to set the request name which
5885 will be added elsewhere.
5888 *) Update OCSP API. Remove obsolete extensions argument from
5889 various functions. Extensions are now handled using the new
5890 OCSP extension code. New simple OCSP HTTP function which
5891 can be used to send requests and parse the response.
5894 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5895 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5896 uses the special reorder version of SET OF to sort the attributes
5897 and reorder them to match the encoded order. This resolves a long
5898 standing problem: a verify on a PKCS7 structure just after signing
5899 it used to fail because the attribute order did not match the
5900 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5901 it uses the received order. This is necessary to tolerate some broken
5902 software that does not order SET OF. This is handled by encoding
5903 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5904 to produce the required SET OF.
5907 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5908 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5909 files to get correct declarations of the ASN.1 item variables.
5912 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5913 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5914 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5915 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5916 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5917 ASN1_ITEM and no wrapper functions.
5920 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5921 replace the old function pointer based I/O routines. Change most of
5922 the *_d2i_bio() and *_d2i_fp() functions to use these.
5925 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5926 lines, recognice more "algorithms" that can be deselected, and make
5927 it complain about algorithm deselection that isn't recognised.
5930 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5931 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5932 to use new functions. Add NO_ASN1_OLD which can be set to remove
5933 some old style ASN1 functions: this can be used to determine if old
5934 code will still work when these eventually go away.
5937 *) New extension functions for OCSP structures, these follow the
5938 same conventions as certificates and CRLs.
5941 *) New function X509V3_add1_i2d(). This automatically encodes and
5942 adds an extension. Its behaviour can be customised with various
5943 flags to append, replace or delete. Various wrappers added for
5944 certifcates and CRLs.
5947 *) Fix to avoid calling the underlying ASN1 print routine when
5948 an extension cannot be parsed. Correct a typo in the
5949 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5952 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5953 entries for variables.
5956 *) Add functionality to apps/openssl.c for detecting locking
5957 problems: As the program is single-threaded, all we have
5958 to do is register a locking callback using an array for
5959 storing which locks are currently held by the program.
5962 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5963 SSL_get_ex_data_X509_STORE_idx(), which is used in
5964 ssl_verify_cert_chain() and thus can be called at any time
5965 during TLS/SSL handshakes so that thread-safety is essential.
5966 Unfortunately, the ex_data design is not at all suited
5967 for multi-threaded use, so it probably should be abolished.
5970 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5971 [Broadcom, tweaked and integrated by Geoff Thorpe]
5973 *) Move common extension printing code to new function
5974 X509V3_print_extensions(). Reorganise OCSP print routines and
5975 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5978 *) New function X509_signature_print() to remove duplication in some
5982 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5983 set (this was treated exactly the same as SET OF previously). This
5984 is used to reorder the STACK representing the structure to match the
5985 encoding. This will be used to get round a problem where a PKCS7
5986 structure which was signed could not be verified because the STACK
5987 order did not reflect the encoded order.
5990 *) Reimplement the OCSP ASN1 module using the new code.
5993 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5994 for its ASN1 operations. The old style function pointers still exist
5995 for now but they will eventually go away.
5998 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5999 completely replaces the old ASN1 functionality with a table driven
6000 encoder and decoder which interprets an ASN1_ITEM structure describing
6001 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6002 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6003 has also been converted to the new form.
6006 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6007 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6008 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6009 for negative moduli.
6012 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6013 of not touching the result's sign bit.
6016 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6020 *) Changed the LHASH code to use prototypes for callbacks, and created
6021 macros to declare and implement thin (optionally static) functions
6022 that provide type-safety and avoid function pointer casting for the
6023 type-specific callbacks.
6026 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6028 [Veers Staats <staatsvr@asc.hpc.mil>,
6029 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6031 *) Reformat the FAQ so the different questions and answers can be divided
6032 in sections depending on the subject.
6035 *) Have the zlib compression code load ZLIB.DLL dynamically under
6039 *) New function BN_mod_sqrt for computing square roots modulo a prime
6040 (using the probabilistic Tonelli-Shanks algorithm unless
6041 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6042 be handled deterministically).
6043 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6045 *) Make BN_mod_inverse faster by explicitly handling small quotients
6046 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6047 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6050 *) New function BN_kronecker.
6053 *) Fix BN_gcd so that it works on negative inputs; the result is
6054 positive unless both parameters are zero.
6055 Previously something reasonably close to an infinite loop was
6056 possible because numbers could be growing instead of shrinking
6057 in the implementation of Euclid's algorithm.
6060 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6061 sign of the number in question.
6063 Fix BN_is_word(a,w) to work correctly for w == 0.
6065 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6066 because its test if the absolute value of 'a' equals 'w'.
6067 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6068 it exists mostly for use in the implementations of BN_is_zero(),
6069 BN_is_one(), and BN_is_word().
6072 *) New function BN_swap.
6075 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6076 the exponentiation functions are more likely to produce reasonable
6077 results on negative inputs.
6080 *) Change BN_mod_mul so that the result is always non-negative.
6081 Previously, it could be negative if one of the factors was negative;
6082 I don't think anyone really wanted that behaviour.
6085 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6086 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6087 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6088 and add new functions:
6097 BN_mod_lshift1_quick
6101 These functions always generate non-negative results.
6103 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6104 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6106 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6107 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6108 be reduced modulo m.
6109 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6112 The following entry accidentily appeared in the CHANGES file
6113 distributed with OpenSSL 0.9.7. The modifications described in
6114 it do *not* apply to OpenSSL 0.9.7.
6116 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6117 was actually never needed) and in BN_mul(). The removal in BN_mul()
6118 required a small change in bn_mul_part_recursive() and the addition
6119 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6120 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6121 bn_sub_words() and bn_add_words() except they take arrays with
6126 *) In 'openssl passwd', verify passwords read from the terminal
6127 unless the '-salt' option is used (which usually means that
6128 verification would just waste user's time since the resulting
6129 hash is going to be compared with some given password hash)
6130 or the new '-noverify' option is used.
6132 This is an incompatible change, but it does not affect
6133 non-interactive use of 'openssl passwd' (passwords on the command
6134 line, '-stdin' option, '-in ...' option) and thus should not
6138 *) Remove all references to RSAref, since there's no more need for it.
6141 *) Make DSO load along a path given through an environment variable
6142 (SHLIB_PATH) with shl_load().
6145 *) Constify the ENGINE code as a result of BIGNUM constification.
6146 Also constify the RSA code and most things related to it. In a
6147 few places, most notable in the depth of the ASN.1 code, ugly
6148 casts back to non-const were required (to be solved at a later
6152 *) Make it so the openssl application has all engines loaded by default.
6155 *) Constify the BIGNUM routines a little more.
6158 *) Add the following functions:
6160 ENGINE_load_cswift()
6162 ENGINE_load_atalla()
6164 ENGINE_load_builtin_engines()
6166 That way, an application can itself choose if external engines that
6167 are built-in in OpenSSL shall ever be used or not. The benefit is
6168 that applications won't have to be linked with libdl or other dso
6169 libraries unless it's really needed.
6171 Changed 'openssl engine' to load all engines on demand.
6172 Changed the engine header files to avoid the duplication of some
6173 declarations (they differed!).
6176 *) 'openssl engine' can now list capabilities.
6179 *) Better error reporting in 'openssl engine'.
6182 *) Never call load_dh_param(NULL) in s_server.
6185 *) Add engine application. It can currently list engines by name and
6186 identity, and test if they are actually available.
6189 *) Improve RPM specification file by forcing symbolic linking and making
6190 sure the installed documentation is also owned by root.root.
6191 [Damien Miller <djm@mindrot.org>]
6193 *) Give the OpenSSL applications more possibilities to make use of
6194 keys (public as well as private) handled by engines.
6197 *) Add OCSP code that comes from CertCo.
6200 *) Add VMS support for the Rijndael code.
6203 *) Added untested support for Nuron crypto accelerator.
6206 *) Add support for external cryptographic devices. This code was
6207 previously distributed separately as the "engine" branch.
6208 [Geoff Thorpe, Richard Levitte]
6210 *) Rework the filename-translation in the DSO code. It is now possible to
6211 have far greater control over how a "name" is turned into a filename
6212 depending on the operating environment and any oddities about the
6213 different shared library filenames on each system.
6216 *) Support threads on FreeBSD-elf in Configure.
6219 *) Fix for SHA1 assembly problem with MASM: it produces
6220 warnings about corrupt line number information when assembling
6221 with debugging information. This is caused by the overlapping
6223 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
6226 NCONF_get_number() has no error checking at all. As a replacement,
6227 NCONF_get_number_e() is defined (_e for "error checking") and is
6228 promoted strongly. The old NCONF_get_number is kept around for
6229 binary backward compatibility.
6230 Make it possible for methods to load from something other than a BIO,
6231 by providing a function pointer that is given a name instead of a BIO.
6232 For example, this could be used to load configuration data from an
6236 *) Fix for non blocking accept BIOs. Added new I/O special reason
6237 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
6238 with non blocking I/O was not possible because no retry code was
6239 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
6243 *) Added the beginnings of Rijndael support.
6246 *) Fix for bug in DirectoryString mask setting. Add support for
6247 X509_NAME_print_ex() in 'req' and X509_print_ex() function
6248 to allow certificate printing to more controllable, additional
6249 'certopt' option to 'x509' to allow new printing options to be
6253 *) Clean old EAY MD5 hack from e_os.h.
6256 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
6258 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6259 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6260 [Joe Orton, Steve Henson]
6262 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
6264 *) Fix additional bug revealed by the NISCC test suite:
6266 Stop bug triggering large recursion when presented with
6267 certain ASN.1 tags (CVE-2003-0851)
6270 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
6272 *) Fix various bugs revealed by running the NISCC test suite:
6274 Stop out of bounds reads in the ASN1 code when presented with
6275 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6277 If verify callback ignores invalid public key errors don't try to check
6278 certificate signature with the NULL public key.
6282 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6283 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6287 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6288 extra data after the compression methods not only for TLS 1.0
6289 but also for SSL 3.0 (as required by the specification).
6290 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6292 *) Change X509_certificate_type() to mark the key as exported/exportable
6293 when it's 512 *bits* long, not 512 bytes.
6296 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
6298 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6299 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6300 a protocol version number mismatch like a decryption error
6301 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6304 *) Turn on RSA blinding by default in the default implementation
6305 to avoid a timing attack. Applications that don't want it can call
6306 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6307 They would be ill-advised to do so in most cases.
6308 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6310 *) Change RSA blinding code so that it works when the PRNG is not
6311 seeded (in this case, the secret RSA exponent is abused as
6312 an unpredictable seed -- if it is not unpredictable, there
6313 is no point in blinding anyway). Make RSA blinding thread-safe
6314 by remembering the creator's thread ID in rsa->blinding and
6315 having all other threads use local one-time blinding factors
6316 (this requires more computation than sharing rsa->blinding, but
6317 avoids excessive locking; and if an RSA object is not shared
6318 between threads, blinding will still be very fast).
6321 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
6323 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6324 via timing by performing a MAC computation even if incorrrect
6325 block cipher padding has been found. This is a countermeasure
6326 against active attacks where the attacker has to distinguish
6327 between bad padding and a MAC verification error. (CVE-2003-0078)
6329 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6330 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6331 Martin Vuagnoux (EPFL, Ilion)]
6333 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
6335 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
6336 memory from it's contents. This is done with a counter that will
6337 place alternating values in each byte. This can be used to solve
6338 two issues: 1) the removal of calls to memset() by highly optimizing
6339 compilers, and 2) cleansing with other values than 0, since those can
6340 be read through on certain media, for example a swap space on disk.
6343 *) Bugfix: client side session caching did not work with external caching,
6344 because the session->cipher setting was not restored when reloading
6345 from the external cache. This problem was masked, when
6346 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
6347 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
6350 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
6351 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
6352 [Zeev Lieber <zeev-l@yahoo.com>]
6354 *) Undo an undocumented change introduced in 0.9.6e which caused
6355 repeated calls to OpenSSL_add_all_ciphers() and
6356 OpenSSL_add_all_digests() to be ignored, even after calling
6360 *) Change the default configuration reader to deal with last line not
6361 being properly terminated.
6364 *) Change X509_NAME_cmp() so it applies the special rules on handling
6365 DN values that are of type PrintableString, as well as RDNs of type
6366 emailAddress where the value has the type ia5String.
6367 [stefank@valicert.com via Richard Levitte]
6369 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
6370 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
6371 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
6372 the bitwise-OR of the two for use by the majority of applications
6373 wanting this behaviour, and update the docs. The documented
6374 behaviour and actual behaviour were inconsistent and had been
6375 changing anyway, so this is more a bug-fix than a behavioural
6377 [Geoff Thorpe, diagnosed by Nadav Har'El]
6379 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
6380 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
6383 *) Fix initialization code race conditions in
6384 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
6385 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
6386 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
6387 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
6388 ssl2_get_cipher_by_char(),
6389 ssl3_get_cipher_by_char().
6390 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
6392 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
6393 the cached sessions are flushed, as the remove_cb() might use ex_data
6394 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
6395 (see [openssl.org #212]).
6396 [Geoff Thorpe, Lutz Jaenicke]
6398 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
6399 length, instead of the encoding length to d2i_ASN1_OBJECT.
6402 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
6404 *) [In 0.9.6g-engine release:]
6405 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
6406 [Lynn Gazis <lgazis@rainbow.com>]
6408 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
6410 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
6411 and get fix the header length calculation.
6412 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
6413 Alon Kantor <alonk@checkpoint.com> (and others),
6416 *) Use proper error handling instead of 'assertions' in buffer
6417 overflow checks added in 0.9.6e. This prevents DoS (the
6418 assertions could call abort()).
6419 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
6421 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
6423 *) Add various sanity checks to asn1_get_length() to reject
6424 the ASN1 length bytes if they exceed sizeof(long), will appear
6425 negative or the content length exceeds the length of the
6427 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6429 *) Fix cipher selection routines: ciphers without encryption had no flags
6430 for the cipher strength set and where therefore not handled correctly
6431 by the selection routines (PR #130).
6434 *) Fix EVP_dsa_sha macro.
6438 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
6439 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
6440 that was added in OpenSSL 0.9.6d.
6442 As the countermeasure turned out to be incompatible with some
6443 broken SSL implementations, the new option is part of SSL_OP_ALL.
6444 SSL_OP_ALL is usually employed when compatibility with weird SSL
6445 implementations is desired (e.g. '-bugs' option to 's_client' and
6446 's_server'), so the new option is automatically set in many
6450 *) Changes in security patch:
6452 Changes marked "(CHATS)" were sponsored by the Defense Advanced
6453 Research Projects Agency (DARPA) and Air Force Research Laboratory,
6454 Air Force Materiel Command, USAF, under agreement number
6457 *) Add various sanity checks to asn1_get_length() to reject
6458 the ASN1 length bytes if they exceed sizeof(long), will appear
6459 negative or the content length exceeds the length of the
6460 supplied buffer. (CVE-2002-0659)
6461 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6463 *) Assertions for various potential buffer overflows, not known to
6465 [Ben Laurie (CHATS)]
6467 *) Various temporary buffers to hold ASCII versions of integers were
6468 too small for 64 bit platforms. (CVE-2002-0655)
6469 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6471 *) Remote buffer overflow in SSL3 protocol - an attacker could
6472 supply an oversized session ID to a client. (CVE-2002-0656)
6473 [Ben Laurie (CHATS)]
6475 *) Remote buffer overflow in SSL2 protocol - an attacker could
6476 supply an oversized client master key. (CVE-2002-0656)
6477 [Ben Laurie (CHATS)]
6479 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6481 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6482 encoded as NULL) with id-dsa-with-sha1.
6483 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6485 *) Check various X509_...() return values in apps/req.c.
6486 [Nils Larsch <nla@trustcenter.de>]
6488 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6489 an end-of-file condition would erronously be flagged, when the CRLF
6490 was just at the end of a processed block. The bug was discovered when
6491 processing data through a buffering memory BIO handing the data to a
6492 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6493 <ptsekov@syntrex.com> and Nedelcho Stanev.
6496 *) Implement a countermeasure against a vulnerability recently found
6497 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6498 before application data chunks to avoid the use of known IVs
6499 with data potentially chosen by the attacker.
6502 *) Fix length checks in ssl3_get_client_hello().
6505 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6506 to prevent ssl3_read_internal() from incorrectly assuming that
6507 ssl3_read_bytes() found application data while handshake
6508 processing was enabled when in fact s->s3->in_read_app_data was
6509 merely automatically cleared during the initial handshake.
6510 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6512 *) Fix object definitions for Private and Enterprise: they were not
6513 recognized in their shortname (=lowercase) representation. Extend
6514 obj_dat.pl to issue an error when using undefined keywords instead
6515 of silently ignoring the problem (Svenning Sorensen
6516 <sss@sss.dnsalias.net>).
6519 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6520 generators, i.e. generators other than 2 and 5. (Previously, the
6521 code did not properly initialise the 'add' and 'rem' values to
6522 BN_generate_prime().)
6524 In the new general case, we do not insist that 'generator' is
6525 actually a primitive root: This requirement is rather pointless;
6526 a generator of the order-q subgroup is just as good, if not
6530 *) Map new X509 verification errors to alerts. Discovered and submitted by
6531 Tom Wu <tom@arcot.com>.
6534 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6535 returning non-zero before the data has been completely received
6536 when using non-blocking I/O.
6537 [Bodo Moeller; problem pointed out by John Hughes]
6539 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6540 [Ben Laurie, Lutz Jaenicke]
6542 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6543 Yoram Zahavi <YoramZ@gilian.com>).
6546 *) Add information about CygWin 1.3 and on, and preserve proper
6547 configuration for the versions before that.
6548 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6550 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6551 check whether we deal with a copy of a session and do not delete from
6552 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6553 <izhar@checkpoint.com>.
6556 *) Do not store session data into the internal session cache, if it
6557 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6558 flag is set). Proposed by Aslam <aslam@funk.com>.
6561 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6565 *) [In 0.9.6d-engine release:]
6566 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6567 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6569 *) Add the configuration target linux-s390x.
6570 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6572 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6573 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6574 variable as an indication that a ClientHello message has been
6575 received. As the flag value will be lost between multiple
6576 invocations of ssl3_accept when using non-blocking I/O, the
6577 function may not be aware that a handshake has actually taken
6578 place, thus preventing a new session from being added to the
6581 To avoid this problem, we now set s->new_session to 2 instead of
6582 using a local variable.
6583 [Lutz Jaenicke, Bodo Moeller]
6585 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6586 if the SSL_R_LENGTH_MISMATCH error is detected.
6587 [Geoff Thorpe, Bodo Moeller]
6589 *) New 'shared_ldflag' column in Configure platform table.
6592 *) Fix EVP_CIPHER_mode macro.
6593 ["Dan S. Camper" <dan@bti.net>]
6595 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
6596 type, we must throw them away by setting rr->length to 0.
6597 [D P Chang <dpc@qualys.com>]
6599 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
6601 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
6602 <Dominikus.Scherkl@biodata.com>. (The previous implementation
6603 worked incorrectly for those cases where range = 10..._2 and
6604 3*range is two bits longer than range.)
6607 *) Only add signing time to PKCS7 structures if it is not already
6611 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
6612 OBJ_ld_ce should be OBJ_id_ce.
6613 Also some ip-pda OIDs in crypto/objects/objects.txt were
6614 incorrect (cf. RFC 3039).
6615 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
6617 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
6618 returns early because it has nothing to do.
6619 [Andy Schneider <andy.schneider@bjss.co.uk>]
6621 *) [In 0.9.6c-engine release:]
6622 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
6623 [Andy Schneider <andy.schneider@bjss.co.uk>]
6625 *) [In 0.9.6c-engine release:]
6626 Add support for Cryptographic Appliance's keyserver technology.
6627 (Use engine 'keyclient')
6628 [Cryptographic Appliances and Geoff Thorpe]
6630 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
6631 is called via tools/c89.sh because arguments have to be
6632 rearranged (all '-L' options must appear before the first object
6634 [Richard Shapiro <rshapiro@abinitio.com>]
6636 *) [In 0.9.6c-engine release:]
6637 Add support for Broadcom crypto accelerator cards, backported
6639 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
6641 *) [In 0.9.6c-engine release:]
6642 Add support for SureWare crypto accelerator cards from
6643 Baltimore Technologies. (Use engine 'sureware')
6644 [Baltimore Technologies and Mark Cox]
6646 *) [In 0.9.6c-engine release:]
6647 Add support for crypto accelerator cards from Accelerated
6648 Encryption Processing, www.aep.ie. (Use engine 'aep')
6649 [AEP Inc. and Mark Cox]
6651 *) Add a configuration entry for gcc on UnixWare.
6652 [Gary Benson <gbenson@redhat.com>]
6654 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
6655 messages are stored in a single piece (fixed-length part and
6656 variable-length part combined) and fix various bugs found on the way.
6659 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
6660 instead. BIO_gethostbyname() does not know what timeouts are
6661 appropriate, so entries would stay in cache even when they have
6663 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
6665 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
6666 faced with a pathologically small ClientHello fragment that does
6667 not contain client_version: Instead of aborting with an error,
6668 simply choose the highest available protocol version (i.e.,
6669 TLS 1.0 unless it is disabled). In practice, ClientHello
6670 messages are never sent like this, but this change gives us
6671 strictly correct behaviour at least for TLS.
6674 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
6675 never resets s->method to s->ctx->method when called from within
6676 one of the SSL handshake functions.
6677 [Bodo Moeller; problem pointed out by Niko Baric]
6679 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
6680 (sent using the client's version number) if client_version is
6681 smaller than the protocol version in use. Also change
6682 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
6683 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
6684 the client will at least see that alert.
6687 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
6691 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
6692 client receives HelloRequest while in a handshake.
6693 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
6695 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
6696 should end in 'break', not 'goto end' which circuments various
6697 cleanups done in state SSL_ST_OK. But session related stuff
6698 must be disabled for SSL_ST_OK in the case that we just sent a
6701 Also avoid some overhead by not calling ssl_init_wbio_buffer()
6702 before just sending a HelloRequest.
6703 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
6705 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
6706 reveal whether illegal block cipher padding was found or a MAC
6707 verification error occured. (Neither SSLerr() codes nor alerts
6708 are directly visible to potential attackers, but the information
6709 may leak via logfiles.)
6711 Similar changes are not required for the SSL 2.0 implementation
6712 because the number of padding bytes is sent in clear for SSL 2.0,
6713 and the extra bytes are just ignored. However ssl/s2_pkt.c
6714 failed to verify that the purported number of padding bytes is in
6718 *) Add OpenUNIX-8 support including shared libraries
6719 (Boyd Lynn Gerber <gerberb@zenez.com>).
6722 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
6723 'wristwatch attack' using huge encoding parameters (cf.
6724 James H. Manger's CRYPTO 2001 paper). Note that the
6725 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
6726 encoding parameters and hence was not vulnerable.
6729 *) BN_sqr() bug fix.
6730 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
6732 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
6733 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
6734 followed by modular reduction.
6735 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
6737 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
6738 equivalent based on BN_pseudo_rand() instead of BN_rand().
6741 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
6742 This function was broken, as the check for a new client hello message
6743 to handle SGC did not allow these large messages.
6744 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
6747 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
6750 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
6751 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
6754 *) Rework the configuration and shared library support for Tru64 Unix.
6755 The configuration part makes use of modern compiler features and
6756 still retains old compiler behavior for those that run older versions
6757 of the OS. The shared library support part includes a variant that
6758 uses the RPATH feature, and is available through the special
6759 configuration target "alpha-cc-rpath", which will never be selected
6761 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
6763 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
6764 with the same message size as in ssl3_get_certificate_request().
6765 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
6766 messages might inadvertently be reject as too long.
6767 [Petr Lampa <lampa@fee.vutbr.cz>]
6769 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6772 *) Modified SSL library such that the verify_callback that has been set
6773 specificly for an SSL object with SSL_set_verify() is actually being
6774 used. Before the change, a verify_callback set with this function was
6775 ignored and the verify_callback() set in the SSL_CTX at the time of
6776 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6777 to allow the necessary settings.
6780 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6781 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6782 done automatically (in contradiction to the requirements of the C
6783 standard). This made problems when used from OpenSSH.
6786 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6787 dh->length and always used
6789 BN_rand_range(priv_key, dh->p).
6791 BN_rand_range() is not necessary for Diffie-Hellman, and this
6792 specific range makes Diffie-Hellman unnecessarily inefficient if
6793 dh->length (recommended exponent length) is much smaller than the
6794 length of dh->p. We could use BN_rand_range() if the order of
6795 the subgroup was stored in the DH structure, but we only have
6800 BN_rand(priv_key, l, ...)
6802 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6808 RSA_eay_public_encrypt
6809 RSA_eay_private_decrypt
6810 RSA_eay_private_encrypt (signing)
6811 RSA_eay_public_decrypt (signature verification)
6813 (default implementations for RSA_public_encrypt,
6814 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6815 always reject numbers >= n.
6818 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6819 to synchronize access to 'locking_thread'. This is necessary on
6820 systems where access to 'locking_thread' (an 'unsigned long'
6821 variable) is not atomic.
6824 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6825 *before* setting the 'crypto_lock_rand' flag. The previous code had
6826 a race condition if 0 is a valid thread ID.
6827 [Travis Vitek <vitek@roguewave.com>]
6829 *) Add support for shared libraries under Irix.
6830 [Albert Chin-A-Young <china@thewrittenword.com>]
6832 *) Add configuration option to build on Linux on both big-endian and
6834 [Ralf Baechle <ralf@uni-koblenz.de>]
6836 *) Add the possibility to create shared libraries on HP-UX.
6839 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6841 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6842 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6843 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6844 PRNG state recovery was possible based on the output of
6845 one PRNG request appropriately sized to gain knowledge on
6846 'md' followed by enough consecutive 1-byte PRNG requests
6847 to traverse all of 'state'.
6849 1. When updating 'md_local' (the current thread's copy of 'md')
6850 during PRNG output generation, hash all of the previous
6851 'md_local' value, not just the half used for PRNG output.
6853 2. Make the number of bytes from 'state' included into the hash
6854 independent from the number of PRNG bytes requested.
6856 The first measure alone would be sufficient to avoid
6857 Markku-Juhani's attack. (Actually it had never occurred
6858 to me that the half of 'md_local' used for chaining was the
6859 half from which PRNG output bytes were taken -- I had always
6860 assumed that the secret half would be used.) The second
6861 measure makes sure that additional data from 'state' is never
6862 mixed into 'md_local' in small portions; this heuristically
6863 further strengthens the PRNG.
6866 *) Fix crypto/bn/asm/mips3.s.
6869 *) When only the key is given to "enc", the IV is undefined. Print out
6870 an error message in this case.
6873 *) Handle special case when X509_NAME is empty in X509 printing routines.
6876 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6877 positive and less than q.
6880 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6881 used: it isn't thread safe and the add_lock_callback should handle
6883 [Paul Rose <Paul.Rose@bridge.com>]
6885 *) Verify that incoming data obeys the block size in
6886 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6890 [Ulf Möller, Bodo Möller]
6892 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6893 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6894 when fixing the server behaviour for backwards-compatible 'client
6895 hello' messages. (Note that the attack is impractical against
6896 SSL 3.0 and TLS 1.0 anyway because length and version checking
6897 means that the probability of guessing a valid ciphertext is
6898 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6901 Before 0.9.5, the countermeasure (hide the error by generating a
6902 random 'decryption result') did not work properly because
6903 ERR_clear_error() was missing, meaning that SSL_get_error() would
6904 detect the supposedly ignored error.
6906 Both problems are now fixed.
6909 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6910 (previously it was 1024).
6913 *) Fix for compatibility mode trust settings: ignore trust settings
6914 unless some valid trust or reject settings are present.
6917 *) Fix for blowfish EVP: its a variable length cipher.
6920 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6921 parameters in DSA public key structures and return an error in the
6922 DSA routines if parameters are absent.
6925 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6926 in the current directory if neither $RANDFILE nor $HOME was set.
6927 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6928 caused some confusion to Windows users who haven't defined $HOME.
6929 Thus RAND_file_name() is changed again: e_os.h can define a
6930 DEFAULT_HOME, which will be used if $HOME is not set.
6931 For Windows, we use "C:"; on other platforms, we still require
6932 environment variables.
6934 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6935 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6936 having multiple threads call RAND_poll() concurrently.
6939 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6940 combination of a flag and a thread ID variable.
6941 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6942 flag), *other* threads can enter ssleay_add_bytes without obeying
6943 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6944 that they do not hold after the first thread unsets add_do_not_lock).
6947 *) Change bctest again: '-x' expressions are not available in all
6951 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6953 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6954 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6956 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6957 the default extension for executables, if any. Also, make the perl
6958 scripts that use symlink() to test if it really exists and use "cp"
6959 if it doesn't. All this made OpenSSL compilable and installable in
6963 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6964 If SEQUENCE is length is indefinite just set c->slen to the total
6965 amount of data available.
6966 [Steve Henson, reported by shige@FreeBSD.org]
6967 [This change does not apply to 0.9.7.]
6969 *) Change bctest to avoid here-documents inside command substitution
6970 (workaround for FreeBSD /bin/sh bug).
6971 For compatibility with Ultrix, avoid shell functions (introduced
6972 in the bctest version that searches along $PATH).
6975 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6976 with des_encrypt() defined on some operating systems, like Solaris
6980 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6981 On the Importance of Eliminating Errors in Cryptographic
6982 Computations, J. Cryptology 14 (2001) 2, 101-119,
6983 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6986 *) MIPS assembler BIGNUM division bug fix.
6989 *) Disabled incorrect Alpha assembler code.
6992 *) Fix PKCS#7 decode routines so they correctly update the length
6993 after reading an EOC for the EXPLICIT tag.
6995 [This change does not apply to 0.9.7.]
6997 *) Fix bug in PKCS#12 key generation routines. This was triggered
6998 if a 3DES key was generated with a 0 initial byte. Include
6999 PKCS12_BROKEN_KEYGEN compilation option to retain the old
7000 (but broken) behaviour.
7003 *) Enhance bctest to search for a working bc along $PATH and print
7005 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7007 *) Fix memory leaks in err.c: free err_data string if necessary;
7008 don't write to the wrong index in ERR_set_error_data.
7011 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7015 *) Replace rdtsc with _emit statements for VC++ version 5.
7016 [Jeremy Cooper <jeremy@baymoo.org>]
7018 *) Make it possible to reuse SSLv2 sessions.
7021 *) In copy_email() check for >= 0 as a return value for
7022 X509_NAME_get_index_by_NID() since 0 is a valid index.
7023 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7025 *) Avoid coredump with unsupported or invalid public keys by checking if
7026 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7027 PKCS7_verify() fails with non detached data.
7030 *) Don't use getenv in library functions when run as setuid/setgid.
7031 New function OPENSSL_issetugid().
7034 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7035 due to incorrect handling of multi-threading:
7037 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7039 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7041 3. Count how many times MemCheck_off() has been called so that
7042 nested use can be treated correctly. This also avoids
7043 inband-signalling in the previous code (which relied on the
7044 assumption that thread ID 0 is impossible).
7047 *) Add "-rand" option also to s_client and s_server.
7050 *) Fix CPU detection on Irix 6.x.
7051 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7052 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7054 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7057 [This change does not apply to 0.9.7.]
7059 *) Use the cached encoding of an X509_NAME structure rather than
7060 copying it. This is apparently the reason for the libsafe "errors"
7061 but the code is actually correct.
7064 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7065 Bleichenbacher's DSA attack.
7066 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7067 to be set and top=0 forces the highest bit to be set; top=-1 is new
7068 and leaves the highest bit random.
7069 [Ulf Moeller, Bodo Moeller]
7071 *) In the NCONF_...-based implementations for CONF_... queries
7072 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7073 a temporary CONF structure with the data component set to NULL
7074 (which gives segmentation faults in lh_retrieve).
7075 Instead, use NULL for the CONF pointer in CONF_get_string and
7076 CONF_get_number (which may use environment variables) and directly
7077 return NULL from CONF_get_section.
7080 *) Fix potential buffer overrun for EBCDIC.
7083 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7084 keyUsage if basicConstraints absent for a CA.
7087 *) Make SMIME_write_PKCS7() write mail header values with a format that
7088 is more generally accepted (no spaces before the semicolon), since
7089 some programs can't parse those values properly otherwise. Also make
7090 sure BIO's that break lines after each write do not create invalid
7094 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7095 macros previously used would not encode an empty SEQUENCE OF
7096 and break the signature.
7098 [This change does not apply to 0.9.7.]
7100 *) Zero the premaster secret after deriving the master secret in
7104 *) Add some EVP_add_digest_alias registrations (as found in
7105 OpenSSL_add_all_digests()) to SSL_library_init()
7106 aka OpenSSL_add_ssl_algorithms(). This provides improved
7107 compatibility with peers using X.509 certificates
7108 with unconventional AlgorithmIdentifier OIDs.
7111 *) Fix for Irix with NO_ASM.
7112 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7114 *) ./config script fixes.
7115 [Ulf Moeller, Richard Levitte]
7117 *) Fix 'openssl passwd -1'.
7120 *) Change PKCS12_key_gen_asc() so it can cope with non null
7121 terminated strings whose length is passed in the passlen
7122 parameter, for example from PEM callbacks. This was done
7123 by adding an extra length parameter to asc2uni().
7124 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7126 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7127 call failed, free the DSA structure.
7130 *) Fix to uni2asc() to cope with zero length Unicode strings.
7131 These are present in some PKCS#12 files.
7134 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7135 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7136 when writing a 32767 byte record.
7137 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7139 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7140 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7142 (RSA objects have a reference count access to which is protected
7143 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7144 so they are meant to be shared between threads.)
7145 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7146 "Reddie, Steven" <Steven.Reddie@ca.com>]
7148 *) Fix a deadlock in CRYPTO_mem_leaks().
7151 *) Use better test patterns in bntest.
7154 *) rand_win.c fix for Borland C.
7157 *) BN_rshift bugfix for n == 0.
7160 *) Add a 'bctest' script that checks for some known 'bc' bugs
7161 so that 'make test' does not abort just because 'bc' is broken.
7164 *) Store verify_result within SSL_SESSION also for client side to
7165 avoid potential security hole. (Re-used sessions on the client side
7166 always resulted in verify_result==X509_V_OK, not using the original
7167 result of the server certificate verification.)
7170 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
7171 SSL3_RT_APPLICATION_DATA, return 0.
7172 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
7176 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
7177 releases, have been re-implemented by renaming the previous
7178 implementations of ssl2_read and ssl3_read to ssl2_read_internal
7179 and ssl3_read_internal, respectively, and adding 'peek' parameters
7180 to them. The new ssl[23]_{read,peek} functions are calls to
7181 ssl[23]_read_internal with the 'peek' flag set appropriately.
7182 A 'peek' parameter has also been added to ssl3_read_bytes, which
7183 does the actual work for ssl3_read_internal.
7186 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
7187 the method-specific "init()" handler. Also clean up ex_data after
7188 calling the method-specific "finish()" handler. Previously, this was
7189 happening the other way round.
7192 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
7193 The previous value, 12, was not always sufficient for BN_mod_exp().
7196 *) Make sure that shared libraries get the internal name engine with
7197 the full version number and not just 0. This should mark the
7198 shared libraries as not backward compatible. Of course, this should
7199 be changed again when we can guarantee backward binary compatibility.
7202 *) Fix typo in get_cert_by_subject() in by_dir.c
7203 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
7205 *) Rework the system to generate shared libraries:
7207 - Make note of the expected extension for the shared libraries and
7208 if there is a need for symbolic links from for example libcrypto.so.0
7209 to libcrypto.so.0.9.7. There is extended info in Configure for
7212 - Make as few rebuilds of the shared libraries as possible.
7214 - Still avoid linking the OpenSSL programs with the shared libraries.
7216 - When installing, install the shared libraries separately from the
7220 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
7222 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
7223 and not in SSL_clear because the latter is also used by the
7224 accept/connect functions; previously, the settings made by
7225 SSL_set_read_ahead would be lost during the handshake.
7226 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
7228 *) Correct util/mkdef.pl to be selective about disabled algorithms.
7229 Previously, it would create entries for disableed algorithms no
7233 *) Added several new manual pages for SSL_* function.
7236 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
7238 *) In ssl23_get_client_hello, generate an error message when faced
7239 with an initial SSL 3.0/TLS record that is too small to contain the
7240 first two bytes of the ClientHello message, i.e. client_version.
7241 (Note that this is a pathologic case that probably has never happened
7242 in real life.) The previous approach was to use the version number
7243 from the record header as a substitute; but our protocol choice
7244 should not depend on that one because it is not authenticated
7245 by the Finished messages.
7248 *) More robust randomness gathering functions for Windows.
7249 [Jeffrey Altman <jaltman@columbia.edu>]
7251 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
7252 not set then we don't setup the error code for issuer check errors
7253 to avoid possibly overwriting other errors which the callback does
7254 handle. If an application does set the flag then we assume it knows
7255 what it is doing and can handle the new informational codes
7259 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
7260 a general "ANY" type, as such it should be able to decode anything
7261 including tagged types. However it didn't check the class so it would
7262 wrongly interpret tagged types in the same way as their universal
7263 counterpart and unknown types were just rejected. Changed so that the
7264 tagged and unknown types are handled in the same way as a SEQUENCE:
7265 that is the encoding is stored intact. There is also a new type
7266 "V_ASN1_OTHER" which is used when the class is not universal, in this
7267 case we have no idea what the actual type is so we just lump them all
7271 *) On VMS, stdout may very well lead to a file that is written to
7272 in a record-oriented fashion. That means that every write() will
7273 write a separate record, which will be read separately by the
7274 programs trying to read from it. This can be very confusing.
7276 The solution is to put a BIO filter in the way that will buffer
7277 text until a linefeed is reached, and then write everything a
7278 line at a time, so every record written will be an actual line,
7279 not chunks of lines and not (usually doesn't happen, but I've
7280 seen it once) several lines in one record. BIO_f_linebuffer() is
7283 Currently, it's a VMS-only method, because that's where it has
7284 been tested well enough.
7287 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
7288 it can return incorrect results.
7289 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
7290 but it was in 0.9.6-beta[12].)
7293 *) Disable the check for content being present when verifying detached
7294 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
7295 include zero length content when signing messages.
7298 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
7299 BIO_ctrl (for BIO pairs).
7302 *) Add DSO method for VMS.
7305 *) Bug fix: Montgomery multiplication could produce results with the
7309 *) Add RPM specification openssl.spec and modify it to build three
7310 packages. The default package contains applications, application
7311 documentation and run-time libraries. The devel package contains
7312 include files, static libraries and function documentation. The
7313 doc package contains the contents of the doc directory. The original
7314 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
7317 *) Add a large number of documentation files for many SSL routines.
7318 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7320 *) Add a configuration entry for Sony News 4.
7321 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
7323 *) Don't set the two most significant bits to one when generating a
7324 random number < q in the DSA library.
7327 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
7328 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
7329 the underlying transport is blocking) if a handshake took place.
7330 (The default behaviour is needed by applications such as s_client
7331 and s_server that use select() to determine when to use SSL_read;
7332 but for applications that know in advance when to expect data, it
7333 just makes things more complicated.)
7336 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
7340 *) Add a few more EBCDIC conditionals that make `req' and `x509'
7341 work better on such systems.
7342 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7344 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
7345 Update PKCS12_parse() so it copies the friendlyName and the
7346 keyid to the certificates aux info.
7349 *) Fix bug in PKCS7_verify() which caused an infinite loop
7350 if there was more than one signature.
7351 [Sven Uszpelkat <su@celocom.de>]
7353 *) Major change in util/mkdef.pl to include extra information
7354 about each symbol, as well as presentig variables as well
7355 as functions. This change means that there's n more need
7356 to rebuild the .num files when some algorithms are excluded.
7359 *) Allow the verify time to be set by an application,
7360 rather than always using the current time.
7363 *) Phase 2 verify code reorganisation. The certificate
7364 verify code now looks up an issuer certificate by a
7365 number of criteria: subject name, authority key id
7366 and key usage. It also verifies self signed certificates
7367 by the same criteria. The main comparison function is
7368 X509_check_issued() which performs these checks.
7370 Lot of changes were necessary in order to support this
7371 without completely rewriting the lookup code.
7373 Authority and subject key identifier are now cached.
7375 The LHASH 'certs' is X509_STORE has now been replaced
7376 by a STACK_OF(X509_OBJECT). This is mainly because an
7377 LHASH can't store or retrieve multiple objects with
7378 the same hash value.
7380 As a result various functions (which were all internal
7381 use only) have changed to handle the new X509_STORE
7382 structure. This will break anything that messed round
7383 with X509_STORE internally.
7385 The functions X509_STORE_add_cert() now checks for an
7386 exact match, rather than just subject name.
7388 The X509_STORE API doesn't directly support the retrieval
7389 of multiple certificates matching a given criteria, however
7390 this can be worked round by performing a lookup first
7391 (which will fill the cache with candidate certificates)
7392 and then examining the cache for matches. This is probably
7393 the best we can do without throwing out X509_LOOKUP
7394 entirely (maybe later...).
7396 The X509_VERIFY_CTX structure has been enhanced considerably.
7398 All certificate lookup operations now go via a get_issuer()
7399 callback. Although this currently uses an X509_STORE it
7400 can be replaced by custom lookups. This is a simple way
7401 to bypass the X509_STORE hackery necessary to make this
7402 work and makes it possible to use more efficient techniques
7403 in future. A very simple version which uses a simple
7404 STACK for its trusted certificate store is also provided
7405 using X509_STORE_CTX_trusted_stack().
7407 The verify_cb() and verify() callbacks now have equivalents
7408 in the X509_STORE_CTX structure.
7410 X509_STORE_CTX also has a 'flags' field which can be used
7411 to customise the verify behaviour.
7414 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
7415 excludes S/MIME capabilities.
7418 *) When a certificate request is read in keep a copy of the
7419 original encoding of the signed data and use it when outputing
7420 again. Signatures then use the original encoding rather than
7421 a decoded, encoded version which may cause problems if the
7422 request is improperly encoded.
7425 *) For consistency with other BIO_puts implementations, call
7426 buffer_write(b, ...) directly in buffer_puts instead of calling
7429 In BIO_puts, increment b->num_write as in BIO_write.
7430 [Peter.Sylvester@EdelWeb.fr]
7432 *) Fix BN_mul_word for the case where the word is 0. (We have to use
7433 BN_zero, we may not return a BIGNUM with an array consisting of
7437 *) Avoid calling abort() from within the library when problems are
7438 detected, except if preprocessor symbols have been defined
7439 (such as REF_CHECK, BN_DEBUG etc.).
7442 *) New openssl application 'rsautl'. This utility can be
7443 used for low level RSA operations. DER public key
7444 BIO/fp routines also added.
7447 *) New Configure entry and patches for compiling on QNX 4.
7448 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
7450 *) A demo state-machine implementation was sponsored by
7451 Nuron (http://www.nuron.com/) and is now available in
7452 demos/state_machine.
7455 *) New options added to the 'dgst' utility for signature
7456 generation and verification.
7459 *) Unrecognized PKCS#7 content types are now handled via a
7460 catch all ASN1_TYPE structure. This allows unsupported
7461 types to be stored as a "blob" and an application can
7462 encode and decode it manually.
7465 *) Fix various signed/unsigned issues to make a_strex.c
7467 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7469 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7470 length if passed a buffer. ASN1_INTEGER_to_BN failed
7471 if passed a NULL BN and its argument was negative.
7472 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7474 *) Modification to PKCS#7 encoding routines to output definite
7475 length encoding. Since currently the whole structures are in
7476 memory there's not real point in using indefinite length
7477 constructed encoding. However if OpenSSL is compiled with
7478 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7481 *) Added BIO_vprintf() and BIO_vsnprintf().
7484 *) Added more prefixes to parse for in the the strings written
7485 through a logging bio, to cover all the levels that are available
7486 through syslog. The prefixes are now:
7488 PANIC, EMERG, EMR => LOG_EMERG
7489 ALERT, ALR => LOG_ALERT
7490 CRIT, CRI => LOG_CRIT
7491 ERROR, ERR => LOG_ERR
7492 WARNING, WARN, WAR => LOG_WARNING
7493 NOTICE, NOTE, NOT => LOG_NOTICE
7494 INFO, INF => LOG_INFO
7495 DEBUG, DBG => LOG_DEBUG
7497 and as before, if none of those prefixes are present at the
7498 beginning of the string, LOG_ERR is chosen.
7500 On Win32, the LOG_* levels are mapped according to this:
7502 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7503 LOG_WARNING => EVENTLOG_WARNING_TYPE
7504 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7508 *) Made it possible to reconfigure with just the configuration
7509 argument "reconf" or "reconfigure". The command line arguments
7510 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7511 and are retrieved from there when reconfiguring.
7515 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7517 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7520 *) The obj_dat.pl script was messing up the sorting of object
7521 names. The reason was that it compared the quoted version
7522 of strings as a result "OCSP" > "OCSP Signing" because
7523 " > SPACE. Changed script to store unquoted versions of
7524 names and add quotes on output. It was also omitting some
7525 names from the lookup table if they were given a default
7526 value (that is if SN is missing it is given the same
7527 value as LN and vice versa), these are now added on the
7528 grounds that if an object has a name we should be able to
7529 look it up. Finally added warning output when duplicate
7530 short or long names are found.
7533 *) Changes needed for Tandem NSK.
7534 [Scott Uroff <scott@xypro.com>]
7536 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7537 RSA_padding_check_SSLv23(), special padding was never detected
7538 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7539 version rollback attacks was not effective.
7541 In s23_clnt.c, don't use special rollback-attack detection padding
7542 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7543 client; similarly, in s23_srvr.c, don't do the rollback check if
7544 SSL 2.0 is the only protocol enabled in the server.
7547 *) Make it possible to get hexdumps of unprintable data with 'openssl
7548 asn1parse'. By implication, the functions ASN1_parse_dump() and
7549 BIO_dump_indent() are added.
7552 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7553 these print out strings and name structures based on various
7554 flags including RFC2253 support and proper handling of
7555 multibyte characters. Added options to the 'x509' utility
7556 to allow the various flags to be set.
7559 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7560 Also change the functions X509_cmp_current_time() and
7561 X509_gmtime_adj() work with an ASN1_TIME structure,
7562 this will enable certificates using GeneralizedTime in validity
7563 dates to be checked.
7566 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7567 negative public key encodings) on by default,
7568 NO_NEG_PUBKEY_BUG can be set to disable it.
7571 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7572 content octets. An i2c_ASN1_OBJECT is unnecessary because
7573 the encoding can be trivially obtained from the structure.
7576 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7577 not read locks (CRYPTO_r_[un]lock).
7580 *) A first attempt at creating official support for shared
7581 libraries through configuration. I've kept it so the
7582 default is static libraries only, and the OpenSSL programs
7583 are always statically linked for now, but there are
7584 preparations for dynamic linking in place.
7585 This has been tested on Linux and Tru64.
7588 *) Randomness polling function for Win9x, as described in:
7589 Peter Gutmann, Software Generation of Practically Strong
7593 *) Fix so PRNG is seeded in req if using an already existing
7597 *) New options to smime application. -inform and -outform
7598 allow alternative formats for the S/MIME message including
7599 PEM and DER. The -content option allows the content to be
7600 specified separately. This should allow things like Netscape
7601 form signing output easier to verify.
7604 *) Fix the ASN1 encoding of tags using the 'long form'.
7607 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
7608 STRING types. These convert content octets to and from the
7609 underlying type. The actual tag and length octets are
7610 already assumed to have been read in and checked. These
7611 are needed because all other string types have virtually
7612 identical handling apart from the tag. By having versions
7613 of the ASN1 functions that just operate on content octets
7614 IMPLICIT tagging can be handled properly. It also allows
7615 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
7616 and ASN1_INTEGER are identical apart from the tag.
7619 *) Change the handling of OID objects as follows:
7621 - New object identifiers are inserted in objects.txt, following
7622 the syntax given in objects.README.
7623 - objects.pl is used to process obj_mac.num and create a new
7625 - obj_dat.pl is used to create a new obj_dat.h, using the data in
7628 This is currently kind of a hack, and the perl code in objects.pl
7629 isn't very elegant, but it works as I intended. The simplest way
7630 to check that it worked correctly is to look in obj_dat.h and
7631 check the array nid_objs and make sure the objects haven't moved
7632 around (this is important!). Additions are OK, as well as
7633 consistent name changes.
7636 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
7639 *) Addition of the command line parameter '-rand file' to 'openssl req'.
7640 The given file adds to whatever has already been seeded into the
7641 random pool through the RANDFILE configuration file option or
7642 environment variable, or the default random state file.
7645 *) mkstack.pl now sorts each macro group into lexical order.
7646 Previously the output order depended on the order the files
7647 appeared in the directory, resulting in needless rewriting
7651 *) Patches to make OpenSSL compile under Win32 again. Mostly
7652 work arounds for the VC++ problem that it treats func() as
7653 func(void). Also stripped out the parts of mkdef.pl that
7654 added extra typesafe functions: these no longer exist.
7657 *) Reorganisation of the stack code. The macros are now all
7658 collected in safestack.h . Each macro is defined in terms of
7659 a "stack macro" of the form SKM_<name>(type, a, b). The
7660 DEBUG_SAFESTACK is now handled in terms of function casts,
7661 this has the advantage of retaining type safety without the
7662 use of additional functions. If DEBUG_SAFESTACK is not defined
7663 then the non typesafe macros are used instead. Also modified the
7664 mkstack.pl script to handle the new form. Needs testing to see
7665 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
7666 the default if no major problems. Similar behaviour for ASN1_SET_OF
7667 and PKCS12_STACK_OF.
7670 *) When some versions of IIS use the 'NET' form of private key the
7671 key derivation algorithm is different. Normally MD5(password) is
7672 used as a 128 bit RC4 key. In the modified case
7673 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
7674 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
7675 as the old Netscape_RSA functions except they have an additional
7676 'sgckey' parameter which uses the modified algorithm. Also added
7677 an -sgckey command line option to the rsa utility. Thanks to
7678 Adrian Peck <bertie@ncipher.com> for posting details of the modified
7679 algorithm to openssl-dev.
7682 *) The evp_local.h macros were using 'c.##kname' which resulted in
7683 invalid expansion on some systems (SCO 5.0.5 for example).
7684 Corrected to 'c.kname'.
7685 [Phillip Porch <root@theporch.com>]
7687 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
7688 a STACK of email addresses from a certificate or request, these look
7689 in the subject name and the subject alternative name extensions and
7690 omit any duplicate addresses.
7693 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
7694 This makes DSA verification about 2 % faster.
7697 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
7698 (meaning that now 2^5 values will be precomputed, which is only 4 KB
7699 plus overhead for 1024 bit moduli).
7700 This makes exponentiations about 0.5 % faster for 1024 bit
7701 exponents (as measured by "openssl speed rsa2048").
7704 *) Rename memory handling macros to avoid conflicts with other
7706 Malloc => OPENSSL_malloc
7707 Malloc_locked => OPENSSL_malloc_locked
7708 Realloc => OPENSSL_realloc
7709 Free => OPENSSL_free
7712 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
7713 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
7716 *) CygWin32 support.
7717 [John Jarvie <jjarvie@newsguy.com>]
7719 *) The type-safe stack code has been rejigged. It is now only compiled
7720 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
7721 by default all type-specific stack functions are "#define"d back to
7722 standard stack functions. This results in more streamlined output
7723 but retains the type-safety checking possibilities of the original
7727 *) The STACK code has been cleaned up, and certain type declarations
7728 that didn't make a lot of sense have been brought in line. This has
7729 also involved a cleanup of sorts in safestack.h to more correctly
7730 map type-safe stack functions onto their plain stack counterparts.
7731 This work has also resulted in a variety of "const"ifications of
7732 lots of the code, especially "_cmp" operations which should normally
7733 be prototyped with "const" parameters anyway.
7736 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
7737 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
7738 (The PRNG state consists of two parts, the large pool 'state' and 'md',
7739 where all of 'md' is used each time the PRNG is used, but 'state'
7740 is used only indexed by a cyclic counter. As entropy may not be
7741 well distributed from the beginning, 'md' is important as a
7742 chaining variable. However, the output function chains only half
7743 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
7744 all of 'md', and seeding with STATE_SIZE dummy bytes will result
7745 in all of 'state' being rewritten, with the new values depending
7746 on virtually all of 'md'. This overcomes the 80 bit limitation.)
7749 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
7750 the handshake is continued after ssl_verify_cert_chain();
7751 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
7752 can lead to 'unexplainable' connection aborts later.
7753 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
7755 *) Major EVP API cipher revision.
7756 Add hooks for extra EVP features. This allows various cipher
7757 parameters to be set in the EVP interface. Support added for variable
7758 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
7759 setting of RC2 and RC5 parameters.
7761 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
7764 Remove lots of duplicated code from the EVP library. For example *every*
7765 cipher init() function handles the 'iv' in the same way according to the
7766 cipher mode. They also all do nothing if the 'key' parameter is NULL and
7767 for CFB and OFB modes they zero ctx->num.
7769 New functionality allows removal of S/MIME code RC2 hack.
7771 Most of the routines have the same form and so can be declared in terms
7774 By shifting this to the top level EVP_CipherInit() it can be removed from
7775 all individual ciphers. If the cipher wants to handle IVs or keys
7776 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7779 Change lots of functions like EVP_EncryptUpdate() to now return a
7780 value: although software versions of the algorithms cannot fail
7781 any installed hardware versions can.
7784 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7785 this option is set, tolerate broken clients that send the negotiated
7786 protocol version number instead of the requested protocol version
7790 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7791 i.e. non-zero for export ciphersuites, zero otherwise.
7792 Previous versions had this flag inverted, inconsistent with
7793 rsa_tmp_cb (..._TMP_RSA_CB).
7794 [Bodo Moeller; problem reported by Amit Chopra]
7796 *) Add missing DSA library text string. Work around for some IIS
7797 key files with invalid SEQUENCE encoding.
7800 *) Add a document (doc/standards.txt) that list all kinds of standards
7801 and so on that are implemented in OpenSSL.
7804 *) Enhance c_rehash script. Old version would mishandle certificates
7805 with the same subject name hash and wouldn't handle CRLs at all.
7806 Added -fingerprint option to crl utility, to support new c_rehash
7810 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7813 *) Fix for SSL server purpose checking. Server checking was
7814 rejecting certificates which had extended key usage present
7815 but no ssl client purpose.
7816 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7818 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7819 is a little unclear about how a blank password is handled.
7820 Since the password in encoded as a BMPString with terminating
7821 double NULL a zero length password would end up as just the
7822 double NULL. However no password at all is different and is
7823 handled differently in the PKCS#12 key generation code. NS
7824 treats a blank password as zero length. MSIE treats it as no
7825 password on export: but it will try both on import. We now do
7826 the same: PKCS12_parse() tries zero length and no password if
7827 the password is set to "" or NULL (NULL is now a valid password:
7828 it wasn't before) as does the pkcs12 application.
7831 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7832 perror when PEM_read_bio_X509_REQ fails, the error message must
7833 be obtained from the error queue.
7836 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7837 it in ERR_remove_state if appropriate, and change ERR_get_state
7838 accordingly to avoid race conditions (this is necessary because
7839 thread_hash is no longer constant once set).
7842 *) Bugfix for linux-elf makefile.one.
7845 *) RSA_get_default_method() will now cause a default
7846 RSA_METHOD to be chosen if one doesn't exist already.
7847 Previously this was only set during a call to RSA_new()
7848 or RSA_new_method(NULL) meaning it was possible for
7849 RSA_get_default_method() to return NULL.
7852 *) Added native name translation to the existing DSO code
7853 that will convert (if the flag to do so is set) filenames
7854 that are sufficiently small and have no path information
7855 into a canonical native form. Eg. "blah" converted to
7856 "libblah.so" or "blah.dll" etc.
7859 *) New function ERR_error_string_n(e, buf, len) which is like
7860 ERR_error_string(e, buf), but writes at most 'len' bytes
7861 including the 0 terminator. For ERR_error_string_n, 'buf'
7863 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7865 *) CONF library reworked to become more general. A new CONF
7866 configuration file reader "class" is implemented as well as a
7867 new functions (NCONF_*, for "New CONF") to handle it. The now
7868 old CONF_* functions are still there, but are reimplemented to
7869 work in terms of the new functions. Also, a set of functions
7870 to handle the internal storage of the configuration data is
7871 provided to make it easier to write new configuration file
7872 reader "classes" (I can definitely see something reading a
7873 configuration file in XML format, for example), called _CONF_*,
7874 or "the configuration storage API"...
7876 The new configuration file reading functions are:
7878 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7879 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7881 NCONF_default, NCONF_WIN32
7883 NCONF_dump_fp, NCONF_dump_bio
7885 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7886 NCONF_new creates a new CONF object. This works in the same way
7887 as other interfaces in OpenSSL, like the BIO interface.
7888 NCONF_dump_* dump the internal storage of the configuration file,
7889 which is useful for debugging. All other functions take the same
7890 arguments as the old CONF_* functions wth the exception of the
7891 first that must be a `CONF *' instead of a `LHASH *'.
7893 To make it easer to use the new classes with the old CONF_* functions,
7894 the function CONF_set_default_method is provided.
7897 *) Add '-tls1' option to 'openssl ciphers', which was already
7898 mentioned in the documentation but had not been implemented.
7899 (This option is not yet really useful because even the additional
7900 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7903 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7904 OpenSSL-based applications) load shared libraries and bind to
7905 them in a portable way.
7906 [Geoff Thorpe, with contributions from Richard Levitte]
7908 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7910 *) Make sure _lrotl and _lrotr are only used with MSVC.
7912 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7913 (the default implementation of RAND_status).
7915 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7916 to '-clrext' (= clear extensions), as intended and documented.
7917 [Bodo Moeller; inconsistency pointed out by Michael Attili
7920 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7921 was larger than the MD block size.
7922 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7924 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7925 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7926 using the passed key: if the passed key was a private key the result
7927 of X509_print(), for example, would be to print out all the private key
7931 *) des_quad_cksum() byte order bug fix.
7932 [Ulf Möller, using the problem description in krb4-0.9.7, where
7933 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7935 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7937 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7939 *) For easily testing in shell scripts whether some command
7940 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7941 returns with exit code 0 iff no command of the given name is available.
7942 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7943 the output goes to stdout and nothing is printed to stderr.
7944 Additional arguments are always ignored.
7946 Since for each cipher there is a command of the same name,
7947 the 'no-cipher' compilation switches can be tested this way.
7949 ('openssl no-XXX' is not able to detect pseudo-commands such
7950 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7953 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7956 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7957 is set; it will be thrown away anyway because each handshake creates
7959 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7960 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7961 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7962 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7965 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7966 'Q' and 'R' lose their special meanings (quit/renegotiate).
7967 This is part of what -quiet does; unlike -quiet, -ign_eof
7968 does not suppress any output.
7971 *) Add compatibility options to the purpose and trust code. The
7972 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7973 accepts a certificate or CA, this was the previous behaviour,
7974 with all the associated security issues.
7976 X509_TRUST_COMPAT is the old trust behaviour: only and
7977 automatically trust self signed roots in certificate store. A
7978 new trust setting X509_TRUST_DEFAULT is used to specify that
7979 a purpose has no associated trust setting and it should instead
7980 use the value in the default purpose.
7983 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7984 and fix a memory leak.
7987 *) In util/mkerr.pl (which implements 'make errors'), preserve
7988 reason strings from the previous version of the .c file, as
7989 the default to have only downcase letters (and digits) in
7990 automatically generated reasons codes is not always appropriate.
7993 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7994 using strerror. Previously, ERR_reason_error_string() returned
7995 library names as reason strings for SYSerr; but SYSerr is a special
7996 case where small numbers are errno values, not library numbers.
7999 *) Add '-dsaparam' option to 'openssl dhparam' application. This
8000 converts DSA parameters into DH parameters. (When creating parameters,
8001 DSA_generate_parameters is used.)
8004 *) Include 'length' (recommended exponent length) in C code generated
8005 by 'openssl dhparam -C'.
8008 *) The second argument to set_label in perlasm was already being used
8009 so couldn't be used as a "file scope" flag. Moved to third argument
8013 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8014 instead of RAND_bytes for encryption IVs and salts.
8017 *) Include RAND_status() into RAND_METHOD instead of implementing
8018 it only for md_rand.c Otherwise replacing the PRNG by calling
8019 RAND_set_rand_method would be impossible.
8022 *) Don't let DSA_generate_key() enter an infinite loop if the random
8023 number generation fails.
8026 *) New 'rand' application for creating pseudo-random output.
8029 *) Added configuration support for Linux/IA64
8030 [Rolf Haberrecker <rolf@suse.de>]
8032 *) Assembler module support for Mingw32.
8035 *) Shared library support for HPUX (in shlib/).
8036 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8038 *) Shared library support for Solaris gcc.
8039 [Lutz Behnke <behnke@trustcenter.de>]
8041 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8043 *) PKCS7_encrypt() was adding text MIME headers twice because they
8044 were added manually and by SMIME_crlf_copy().
8047 *) In bntest.c don't call BN_rand with zero bits argument.
8048 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8050 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8051 case was implemented. This caused BN_div_recp() to fail occasionally.
8054 *) Add an optional second argument to the set_label() in the perl
8055 assembly language builder. If this argument exists and is set
8056 to 1 it signals that the assembler should use a symbol whose
8057 scope is the entire file, not just the current function. This
8058 is needed with MASM which uses the format label:: for this scope.
8059 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8061 *) Change the ASN1 types so they are typedefs by default. Before
8062 almost all types were #define'd to ASN1_STRING which was causing
8063 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8067 *) Change names of new functions to the new get1/get0 naming
8068 convention: After 'get1', the caller owns a reference count
8069 and has to call ..._free; 'get0' returns a pointer to some
8070 data structure without incrementing reference counters.
8071 (Some of the existing 'get' functions increment a reference
8072 counter, some don't.)
8073 Similarly, 'set1' and 'add1' functions increase reference
8074 counters or duplicate objects.
8077 *) Allow for the possibility of temp RSA key generation failure:
8078 the code used to assume it always worked and crashed on failure.
8081 *) Fix potential buffer overrun problem in BIO_printf().
8082 [Ulf Möller, using public domain code by Patrick Powell; problem
8083 pointed out by David Sacerdote <das33@cornell.edu>]
8085 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8086 RAND_egd() and RAND_status(). In the command line application,
8087 the EGD socket can be specified like a seed file using RANDFILE
8091 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8092 Some CAs (e.g. Verisign) distribute certificates in this form.
8095 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8096 list to exclude them. This means that no special compilation option
8097 is needed to use anonymous DH: it just needs to be included in the
8101 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8102 EVP_MD_type. The old functionality is available in a new macro called
8103 EVP_MD_md(). Change code that uses it and update docs.
8106 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8107 where the 'void *' argument is replaced by a function pointer argument.
8108 Previously 'void *' was abused to point to functions, which works on
8109 many platforms, but is not correct. As these functions are usually
8110 called by macros defined in OpenSSL header files, most source code
8111 should work without changes.
8114 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8115 sections with information on -D... compiler switches used for
8116 compiling the library so that applications can see them. To enable
8117 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8118 must be defined. E.g.,
8119 #define OPENSSL_ALGORITHM_DEFINES
8120 #include <openssl/opensslconf.h>
8121 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8122 [Richard Levitte, Ulf and Bodo Möller]
8124 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8128 *) Change the 'other' type in certificate aux info to a STACK_OF
8129 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8130 the required ASN1 format: arbitrary types determined by an OID.
8133 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8134 argument to 'req'. This is not because the function is newer or
8135 better than others it just uses the work 'NEW' in the certificate
8136 request header lines. Some software needs this.
8139 *) Reorganise password command line arguments: now passwords can be
8140 obtained from various sources. Delete the PEM_cb function and make
8141 it the default behaviour: i.e. if the callback is NULL and the
8142 usrdata argument is not NULL interpret it as a null terminated pass
8143 phrase. If usrdata and the callback are NULL then the pass phrase
8144 is prompted for as usual.
8147 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8148 the support is automatically enabled. The resulting binaries will
8149 autodetect the card and use it if present.
8150 [Ben Laurie and Compaq Inc.]
8152 *) Work around for Netscape hang bug. This sends certificate request
8153 and server done in one record. Since this is perfectly legal in the
8154 SSL/TLS protocol it isn't a "bug" option and is on by default. See
8155 the bugs/SSLv3 entry for more info.
8158 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
8161 *) Add -rand argument to smime and pkcs12 applications and read/write
8165 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
8168 *) Add command line password options to the remaining applications.
8171 *) Bug fix for BN_div_recp() for numerators with an even number of
8175 *) More tests in bntest.c, and changed test_bn output.
8178 *) ./config recognizes MacOS X now.
8181 *) Bug fix for BN_div() when the first words of num and divsor are
8182 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
8185 *) Add support for various broken PKCS#8 formats, and command line
8186 options to produce them.
8189 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
8190 get temporary BIGNUMs from a BN_CTX.
8193 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
8197 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
8198 include a #define from the old name to the new. The original intent
8199 was that statically linked binaries could for example just call
8200 SSLeay_add_all_ciphers() to just add ciphers to the table and not
8201 link with digests. This never worked becayse SSLeay_add_all_digests()
8202 and SSLeay_add_all_ciphers() were in the same source file so calling
8203 one would link with the other. They are now in separate source files.
8206 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
8209 *) Use a less unusual form of the Miller-Rabin primality test (it used
8210 a binary algorithm for exponentiation integrated into the Miller-Rabin
8211 loop, our standard modexp algorithms are faster).
8214 *) Support for the EBCDIC character set completed.
8215 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
8217 *) Source code cleanups: use const where appropriate, eliminate casts,
8218 use void * instead of char * in lhash.
8221 *) Bugfix: ssl3_send_server_key_exchange was not restartable
8222 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
8223 this the server could overwrite ephemeral keys that the client
8227 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
8228 using 50 iterations of the Rabin-Miller test.
8230 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
8231 iterations of the Rabin-Miller test as required by the appendix
8232 to FIPS PUB 186[-1]) instead of DSA_is_prime.
8233 As BN_is_prime_fasttest includes trial division, DSA parameter
8234 generation becomes much faster.
8236 This implies a change for the callback functions in DSA_is_prime
8237 and DSA_generate_parameters: The callback function is called once
8238 for each positive witness in the Rabin-Miller test, not just
8239 occasionally in the inner loop; and the parameters to the
8240 callback function now provide an iteration count for the outer
8241 loop rather than for the current invocation of the inner loop.
8242 DSA_generate_parameters additionally can call the callback
8243 function with an 'iteration count' of -1, meaning that a
8244 candidate has passed the trial division test (when q is generated
8245 from an application-provided seed, trial division is skipped).
8248 *) New function BN_is_prime_fasttest that optionally does trial
8249 division before starting the Rabin-Miller test and has
8250 an additional BN_CTX * argument (whereas BN_is_prime always
8251 has to allocate at least one BN_CTX).
8252 'callback(1, -1, cb_arg)' is called when a number has passed the
8253 trial division stage.
8256 *) Fix for bug in CRL encoding. The validity dates weren't being handled
8260 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
8263 *) New function BN_pseudo_rand().
8266 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
8267 bignum version of BN_from_montgomery() with the working code from
8268 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
8272 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
8273 made it impossible to use the same SSL_SESSION data structure in
8274 SSL2 clients in multiple threads.
8277 *) The return value of RAND_load_file() no longer counts bytes obtained
8278 by stat(). RAND_load_file(..., -1) is new and uses the complete file
8279 to seed the PRNG (previously an explicit byte count was required).
8280 [Ulf Möller, Bodo Möller]
8282 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
8283 used (char *) instead of (void *) and had casts all over the place.
8286 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
8289 *) Retain source code compatibility for BN_prime_checks macro:
8290 BN_is_prime(..., BN_prime_checks, ...) now uses
8291 BN_prime_checks_for_size to determine the appropriate number of
8292 Rabin-Miller iterations.
8295 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
8296 DH_CHECK_P_NOT_SAFE_PRIME.
8297 (Check if this is true? OpenPGP calls them "strong".)
8300 *) Merge the functionality of "dh" and "gendh" programs into a new program
8301 "dhparam". The old programs are retained for now but will handle DH keys
8302 (instead of parameters) in future.
8305 *) Make the ciphers, s_server and s_client programs check the return values
8306 when a new cipher list is set.
8309 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
8310 ciphers. Before when the 56bit ciphers were enabled the sorting was
8313 The syntax for the cipher sorting has been extended to support sorting by
8314 cipher-strength (using the strength_bits hard coded in the tables).
8315 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
8317 Fix a bug in the cipher-command parser: when supplying a cipher command
8318 string with an "undefined" symbol (neither command nor alphanumeric
8319 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
8320 an error is flagged.
8322 Due to the strength-sorting extension, the code of the
8323 ssl_create_cipher_list() function was completely rearranged. I hope that
8324 the readability was also increased :-)
8325 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8327 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
8328 for the first serial number and places 2 in the serial number file. This
8329 avoids problems when the root CA is created with serial number zero and
8330 the first user certificate has the same issuer name and serial number
8334 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
8335 the new code. Add documentation for this stuff.
8338 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
8339 X509_*() to X509at_*() on the grounds that they don't handle X509
8340 structures and behave in an analagous way to the X509v3 functions:
8341 they shouldn't be called directly but wrapper functions should be used
8344 So we also now have some wrapper functions that call the X509at functions
8345 when passed certificate requests. (TO DO: similar things can be done with
8346 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
8347 things. Some of these need some d2i or i2d and print functionality
8348 because they handle more complex structures.)
8351 *) Add missing #ifndefs that caused missing symbols when building libssl
8352 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
8353 NO_RSA in ssl/s2*.c.
8354 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
8356 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
8357 has a return value which indicates the quality of the random data
8358 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
8359 error queue. New function RAND_pseudo_bytes() generates output that is
8360 guaranteed to be unique but not unpredictable. RAND_add is like
8361 RAND_seed, but takes an extra argument for an entropy estimate
8362 (RAND_seed always assumes full entropy).
8365 *) Do more iterations of Rabin-Miller probable prime test (specifically,
8366 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
8367 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
8368 in crypto/bn/bn_prime.c for the complete table). This guarantees a
8369 false-positive rate of at most 2^-80 for random input.
8372 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
8375 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
8376 in the 0.9.5 release), this returns the chain
8377 from an X509_CTX structure with a dup of the stack and all
8378 the X509 reference counts upped: so the stack will exist
8379 after X509_CTX_cleanup() has been called. Modify pkcs12.c
8382 Also make SSL_SESSION_print() print out the verify return
8386 *) Add manpage for the pkcs12 command. Also change the default
8387 behaviour so MAC iteration counts are used unless the new
8388 -nomaciter option is used. This improves file security and
8389 only older versions of MSIE (4.0 for example) need it.
8392 *) Honor the no-xxx Configure options when creating .DEF files.
8395 *) Add PKCS#10 attributes to field table: challengePassword,
8396 unstructuredName and unstructuredAddress. These are taken from
8397 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
8398 international characters are used.
8400 More changes to X509_ATTRIBUTE code: allow the setting of types
8401 based on strings. Remove the 'loc' parameter when adding
8402 attributes because these will be a SET OF encoding which is sorted
8406 *) Initial changes to the 'req' utility to allow request generation
8407 automation. This will allow an application to just generate a template
8408 file containing all the field values and have req construct the
8411 Initial support for X509_ATTRIBUTE handling. Stacks of these are
8412 used all over the place including certificate requests and PKCS#7
8413 structures. They are currently handled manually where necessary with
8414 some primitive wrappers for PKCS#7. The new functions behave in a
8415 manner analogous to the X509 extension functions: they allow
8416 attributes to be looked up by NID and added.
8418 Later something similar to the X509V3 code would be desirable to
8419 automatically handle the encoding, decoding and printing of the
8420 more complex types. The string types like challengePassword can
8421 be handled by the string table functions.
8423 Also modified the multi byte string table handling. Now there is
8424 a 'global mask' which masks out certain types. The table itself
8425 can use the flag STABLE_NO_MASK to ignore the mask setting: this
8426 is useful when for example there is only one permissible type
8427 (as in countryName) and using the mask might result in no valid
8431 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
8432 SSL_get_peer_finished to allow applications to obtain the latest
8433 Finished messages sent to the peer or expected from the peer,
8434 respectively. (SSL_get_peer_finished is usually the Finished message
8435 actually received from the peer, otherwise the protocol will be aborted.)
8437 As the Finished message are message digests of the complete handshake
8438 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
8439 be used for external authentication procedures when the authentication
8440 provided by SSL/TLS is not desired or is not enough.
8443 *) Enhanced support for Alpha Linux is added. Now ./config checks if
8444 the host supports BWX extension and if Compaq C is present on the
8445 $PATH. Just exploiting of the BWX extension results in 20-30%
8446 performance kick for some algorithms, e.g. DES and RC4 to mention
8447 a couple. Compaq C in turn generates ~20% faster code for MD5 and
8451 *) Add support for MS "fast SGC". This is arguably a violation of the
8452 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
8453 weak crypto and after checking the certificate is SGC a second one
8454 with strong crypto. MS SGC stops the first handshake after receiving
8455 the server certificate message and sends a second client hello. Since
8456 a server will typically do all the time consuming operations before
8457 expecting any further messages from the client (server key exchange
8458 is the most expensive) there is little difference between the two.
8460 To get OpenSSL to support MS SGC we have to permit a second client
8461 hello message after we have sent server done. In addition we have to
8462 reset the MAC if we do get this second client hello.
8465 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
8466 if a DER encoded private key is RSA or DSA traditional format. Changed
8467 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
8468 format DER encoded private key. Newer code should use PKCS#8 format which
8469 has the key type encoded in the ASN1 structure. Added DER private key
8470 support to pkcs8 application.
8473 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8474 ciphersuites has been selected (as required by the SSL 3/TLS 1
8475 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8476 is set, we interpret this as a request to violate the specification
8477 (the worst that can happen is a handshake failure, and 'correct'
8478 behaviour would result in a handshake failure anyway).
8481 *) In SSL_CTX_add_session, take into account that there might be multiple
8482 SSL_SESSION structures with the same session ID (e.g. when two threads
8483 concurrently obtain them from an external cache).
8484 The internal cache can handle only one SSL_SESSION with a given ID,
8485 so if there's a conflict, we now throw out the old one to achieve
8489 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8490 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8491 some routines that use cipher OIDs: some ciphers do not have OIDs
8492 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8496 *) Simplify the trust setting structure and code. Now we just have
8497 two sequences of OIDs for trusted and rejected settings. These will
8498 typically have values the same as the extended key usage extension
8499 and any application specific purposes.
8501 The trust checking code now has a default behaviour: it will just
8502 check for an object with the same NID as the passed id. Functions can
8503 be provided to override either the default behaviour or the behaviour
8504 for a given id. SSL client, server and email already have functions
8505 in place for compatibility: they check the NID and also return "trusted"
8506 if the certificate is self signed.
8509 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8510 traditional format into an EVP_PKEY structure.
8513 *) Add a password callback function PEM_cb() which either prompts for
8514 a password if usr_data is NULL or otherwise assumes it is a null
8515 terminated password. Allow passwords to be passed on command line
8516 environment or config files in a few more utilities.
8519 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8520 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8521 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8522 Update documentation.
8525 *) Support for ASN1 "NULL" type. This could be handled before by using
8526 ASN1_TYPE but there wasn't any function that would try to read a NULL
8527 and produce an error if it couldn't. For compatibility we also have
8528 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8529 don't allocate anything because they don't need to.
8532 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8534 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8536 *) Rebuild of the memory allocation routines used by OpenSSL code and
8537 possibly others as well. The purpose is to make an interface that
8538 provide hooks so anyone can build a separate set of allocation and
8539 deallocation routines to be used by OpenSSL, for example memory
8540 pool implementations, or something else, which was previously hard
8541 since Malloc(), Realloc() and Free() were defined as macros having
8542 the values malloc, realloc and free, respectively (except for Win32
8543 compilations). The same is provided for memory debugging code.
8544 OpenSSL already comes with functionality to find memory leaks, but
8545 this gives people a chance to debug other memory problems.
8547 With these changes, a new set of functions and macros have appeared:
8549 CRYPTO_set_mem_debug_functions() [F]
8550 CRYPTO_get_mem_debug_functions() [F]
8551 CRYPTO_dbg_set_options() [F]
8552 CRYPTO_dbg_get_options() [F]
8553 CRYPTO_malloc_debug_init() [M]
8555 The memory debug functions are NULL by default, unless the library
8556 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8557 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8558 gives the standard debugging functions that come with OpenSSL) or
8559 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8560 provided by the library user) must be used. When the standard
8561 debugging functions are used, CRYPTO_dbg_set_options can be used to
8562 request additional information:
8563 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8564 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8566 Also, things like CRYPTO_set_mem_functions will always give the
8567 expected result (the new set of functions is used for allocation
8568 and deallocation) at all times, regardless of platform and compiler
8571 To finish it up, some functions that were never use in any other
8572 way than through macros have a new API and new semantic:
8575 CRYPTO_dbg_realloc()
8578 All macros of value have retained their old syntax.
8579 [Richard Levitte and Bodo Moeller]
8581 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8582 ordering of SMIMECapabilities wasn't in "strength order" and there
8583 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8587 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8588 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8589 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
8591 *) Merge in my S/MIME library for OpenSSL. This provides a simple
8592 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
8593 functionality to handle multipart/signed properly) and a utility
8594 called 'smime' to call all this stuff. This is based on code I
8595 originally wrote for Celo who have kindly allowed it to be
8596 included in OpenSSL.
8599 *) Add variants des_set_key_checked and des_set_key_unchecked of
8600 des_set_key (aka des_key_sched). Global variable des_check_key
8601 decides which of these is called by des_set_key; this way
8602 des_check_key behaves as it always did, but applications and
8603 the library itself, which was buggy for des_check_key == 1,
8604 have a cleaner way to pick the version they need.
8607 *) New function PKCS12_newpass() which changes the password of a
8611 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
8612 dynamic mix. In both cases the ids can be used as an index into the
8613 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
8614 functions so they accept a list of the field values and the
8615 application doesn't need to directly manipulate the X509_TRUST
8619 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
8623 *) Modify the way the V3 extension code looks up extensions. This now
8624 works in a similar way to the object code: we have some "standard"
8625 extensions in a static table which is searched with OBJ_bsearch()
8626 and the application can add dynamic ones if needed. The file
8627 crypto/x509v3/ext_dat.h now has the info: this file needs to be
8628 updated whenever a new extension is added to the core code and kept
8629 in ext_nid order. There is a simple program 'tabtest.c' which checks
8630 this. New extensions are not added too often so this file can readily
8631 be maintained manually.
8633 There are two big advantages in doing things this way. The extensions
8634 can be looked up immediately and no longer need to be "added" using
8635 X509V3_add_standard_extensions(): this function now does nothing.
8636 [Side note: I get *lots* of email saying the extension code doesn't
8637 work because people forget to call this function]
8638 Also no dynamic allocation is done unless new extensions are added:
8639 so if we don't add custom extensions there is no need to call
8640 X509V3_EXT_cleanup().
8643 *) Modify enc utility's salting as follows: make salting the default. Add a
8644 magic header, so unsalted files fail gracefully instead of just decrypting
8645 to garbage. This is because not salting is a big security hole, so people
8646 should be discouraged from doing it.
8649 *) Fixes and enhancements to the 'x509' utility. It allowed a message
8650 digest to be passed on the command line but it only used this
8651 parameter when signing a certificate. Modified so all relevant
8652 operations are affected by the digest parameter including the
8653 -fingerprint and -x509toreq options. Also -x509toreq choked if a
8654 DSA key was used because it didn't fix the digest.
8657 *) Initial certificate chain verify code. Currently tests the untrusted
8658 certificates for consistency with the verify purpose (which is set
8659 when the X509_STORE_CTX structure is set up) and checks the pathlength.
8661 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
8662 this is because it will reject chains with invalid extensions whereas
8663 every previous version of OpenSSL and SSLeay made no checks at all.
8665 Trust code: checks the root CA for the relevant trust settings. Trust
8666 settings have an initial value consistent with the verify purpose: e.g.
8667 if the verify purpose is for SSL client use it expects the CA to be
8668 trusted for SSL client use. However the default value can be changed to
8669 permit custom trust settings: one example of this would be to only trust
8670 certificates from a specific "secure" set of CAs.
8672 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
8673 which should be used for version portability: especially since the
8674 verify structure is likely to change more often now.
8676 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
8677 to set them. If not set then assume SSL clients will verify SSL servers
8680 Two new options to the verify program: -untrusted allows a set of
8681 untrusted certificates to be passed in and -purpose which sets the
8682 intended purpose of the certificate. If a purpose is set then the
8683 new chain verify code is used to check extension consistency.
8686 *) Support for the authority information access extension.
8689 *) Modify RSA and DSA PEM read routines to transparently handle
8690 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
8691 public keys in a format compatible with certificate
8692 SubjectPublicKeyInfo structures. Unfortunately there were already
8693 functions called *_PublicKey_* which used various odd formats so
8694 these are retained for compatibility: however the DSA variants were
8695 never in a public release so they have been deleted. Changed dsa/rsa
8696 utilities to handle the new format: note no releases ever handled public
8697 keys so we should be OK.
8699 The primary motivation for this change is to avoid the same fiasco
8700 that dogs private keys: there are several incompatible private key
8701 formats some of which are standard and some OpenSSL specific and
8702 require various evil hacks to allow partial transparent handling and
8703 even then it doesn't work with DER formats. Given the option anything
8704 other than PKCS#8 should be dumped: but the other formats have to
8705 stay in the name of compatibility.
8707 With public keys and the benefit of hindsight one standard format
8708 is used which works with EVP_PKEY, RSA or DSA structures: though
8709 it clearly returns an error if you try to read the wrong kind of key.
8711 Added a -pubkey option to the 'x509' utility to output the public key.
8712 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
8713 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
8714 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
8715 that do the same as the EVP_PKEY_assign_*() except they up the
8716 reference count of the added key (they don't "swallow" the
8720 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
8721 CRLs would fail if the file contained no certificates or no CRLs:
8722 added a new function to read in both types and return the number
8723 read: this means that if none are read it will be an error. The
8724 DER versions of the certificate and CRL reader would always fail
8725 because it isn't possible to mix certificates and CRLs in DER format
8726 without choking one or the other routine. Changed this to just read
8727 a certificate: this is the best we can do. Also modified the code
8728 in apps/verify.c to take notice of return codes: it was previously
8729 attempting to read in certificates from NULL pointers and ignoring
8730 any errors: this is one reason why the cert and CRL reader seemed
8731 to work. It doesn't check return codes from the default certificate
8732 routines: these may well fail if the certificates aren't installed.
8735 *) Code to support otherName option in GeneralName.
8738 *) First update to verify code. Change the verify utility
8739 so it warns if it is passed a self signed certificate:
8740 for consistency with the normal behaviour. X509_verify
8741 has been modified to it will now verify a self signed
8742 certificate if *exactly* the same certificate appears
8743 in the store: it was previously impossible to trust a
8744 single self signed certificate. This means that:
8745 openssl verify ss.pem
8746 now gives a warning about a self signed certificate but
8747 openssl verify -CAfile ss.pem ss.pem
8751 *) For servers, store verify_result in SSL_SESSION data structure
8752 (and add it to external session representation).
8753 This is needed when client certificate verifications fails,
8754 but an application-provided verification callback (set by
8755 SSL_CTX_set_cert_verify_callback) allows accepting the session
8756 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
8757 but returns 1): When the session is reused, we have to set
8758 ssl->verify_result to the appropriate error code to avoid
8760 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
8762 *) Fix a bug in the new PKCS#7 code: it didn't consider the
8763 case in PKCS7_dataInit() where the signed PKCS7 structure
8764 didn't contain any existing data because it was being created.
8765 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
8767 *) Add a salt to the key derivation routines in enc.c. This
8768 forms the first 8 bytes of the encrypted file. Also add a
8769 -S option to allow a salt to be input on the command line.
8772 *) New function X509_cmp(). Oddly enough there wasn't a function
8773 to compare two certificates. We do this by working out the SHA1
8774 hash and comparing that. X509_cmp() will be needed by the trust
8778 *) SSL_get1_session() is like SSL_get_session(), but increments
8779 the reference count in the SSL_SESSION returned.
8780 [Geoff Thorpe <geoff@eu.c2.net>]
8782 *) Fix for 'req': it was adding a null to request attributes.
8783 Also change the X509_LOOKUP and X509_INFO code to handle
8784 certificate auxiliary information.
8787 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8791 *) Add the possibility to add extra information to the memory leak
8792 detecting output, to form tracebacks, showing from where each
8793 allocation was originated: CRYPTO_push_info("constant string") adds
8794 the string plus current file name and line number to a per-thread
8795 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8796 is like calling CYRPTO_pop_info() until the stack is empty.
8797 Also updated memory leak detection code to be multi-thread-safe.
8800 *) Add options -text and -noout to pkcs7 utility and delete the
8801 encryption options which never did anything. Update docs.
8804 *) Add options to some of the utilities to allow the pass phrase
8805 to be included on either the command line (not recommended on
8806 OSes like Unix) or read from the environment. Update the
8807 manpages and fix a few bugs.
8810 *) Add a few manpages for some of the openssl commands.
8813 *) Fix the -revoke option in ca. It was freeing up memory twice,
8814 leaking and not finding already revoked certificates.
8817 *) Extensive changes to support certificate auxiliary information.
8818 This involves the use of X509_CERT_AUX structure and X509_AUX
8819 functions. An X509_AUX function such as PEM_read_X509_AUX()
8820 can still read in a certificate file in the usual way but it
8821 will also read in any additional "auxiliary information". By
8822 doing things this way a fair degree of compatibility can be
8823 retained: existing certificates can have this information added
8824 using the new 'x509' options.
8826 Current auxiliary information includes an "alias" and some trust
8827 settings. The trust settings will ultimately be used in enhanced
8828 certificate chain verification routines: currently a certificate
8829 can only be trusted if it is self signed and then it is trusted
8833 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8834 The problem was that one of the replacement routines had not been working
8835 since SSLeay releases. For now the offending routine has been replaced
8836 with non-optimised assembler. Even so, this now gives around 95%
8837 performance improvement for 1024 bit RSA signs.
8840 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8841 handling. Most clients have the effective key size in bits equal to
8842 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8843 A few however don't do this and instead use the size of the decrypted key
8844 to determine the RC2 key length and the AlgorithmIdentifier to determine
8845 the effective key length. In this case the effective key length can still
8846 be 40 bits but the key length can be 168 bits for example. This is fixed
8847 by manually forcing an RC2 key into the EVP_PKEY structure because the
8848 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8849 the key length and effective key length are equal.
8852 *) Add a bunch of functions that should simplify the creation of
8853 X509_NAME structures. Now you should be able to do:
8854 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8855 and have it automatically work out the correct field type and fill in
8856 the structures. The more adventurous can try:
8857 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8858 and it will (hopefully) work out the correct multibyte encoding.
8861 *) Change the 'req' utility to use the new field handling and multibyte
8862 copy routines. Before the DN field creation was handled in an ad hoc
8863 way in req, ca, and x509 which was rather broken and didn't support
8864 BMPStrings or UTF8Strings. Since some software doesn't implement
8865 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8866 using the dirstring_type option. See the new comment in the default
8867 openssl.cnf for more info.
8870 *) Make crypto/rand/md_rand.c more robust:
8871 - Assure unique random numbers after fork().
8872 - Make sure that concurrent threads access the global counter and
8873 md serializably so that we never lose entropy in them
8874 or use exactly the same state in multiple threads.
8875 Access to the large state is not always serializable because
8876 the additional locking could be a performance killer, and
8877 md should be large enough anyway.
8880 *) New file apps/app_rand.c with commonly needed functionality
8881 for handling the random seed file.
8883 Use the random seed file in some applications that previously did not:
8885 dsaparam -genkey (which also ignored its '-rand' option),
8888 x509 (when signing).
8889 Except on systems with /dev/urandom, it is crucial to have a random
8890 seed file at least for key creation, DSA signing, and for DH exchanges;
8891 for RSA signatures we could do without one.
8893 gendh and gendsa (unlike genrsa) used to read only the first byte
8894 of each file listed in the '-rand' option. The function as previously
8895 found in genrsa is now in app_rand.c and is used by all programs
8896 that support '-rand'.
8899 *) In RAND_write_file, use mode 0600 for creating files;
8900 don't just chmod when it may be too late.
8903 *) Report an error from X509_STORE_load_locations
8904 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8907 *) New function ASN1_mbstring_copy() this copies a string in either
8908 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8909 into an ASN1_STRING type. A mask of permissible types is passed
8910 and it chooses the "minimal" type to use or an error if not type
8914 *) Add function equivalents to the various macros in asn1.h. The old
8915 macros are retained with an M_ prefix. Code inside the library can
8916 use the M_ macros. External code (including the openssl utility)
8917 should *NOT* in order to be "shared library friendly".
8920 *) Add various functions that can check a certificate's extensions
8921 to see if it usable for various purposes such as SSL client,
8922 server or S/MIME and CAs of these types. This is currently
8923 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8924 verification. Also added a -purpose flag to x509 utility to
8925 print out all the purposes.
8928 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8932 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8933 for, obtain and decode and extension and obtain its critical flag.
8934 This allows all the necessary extension code to be handled in a
8935 single function call.
8938 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8939 platforms. See crypto/rc4/rc4_enc.c for further details.
8942 *) New -noout option to asn1parse. This causes no output to be produced
8943 its main use is when combined with -strparse and -out to extract data
8944 from a file (which may not be in ASN.1 format).
8947 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8948 when producing the local key id.
8949 [Richard Levitte <levitte@stacken.kth.se>]
8951 *) New option -dhparam in s_server. This allows a DH parameter file to be
8952 stated explicitly. If it is not stated then it tries the first server
8953 certificate file. The previous behaviour hard coded the filename
8957 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8958 a public key to be input or output. For example:
8959 openssl rsa -in key.pem -pubout -out pubkey.pem
8960 Also added necessary DSA public key functions to handle this.
8963 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8964 in the message. This was handled by allowing
8965 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8966 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8968 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8969 to the end of the strings whereas this didn't. This would cause problems
8970 if strings read with d2i_ASN1_bytes() were later modified.
8971 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8973 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8974 data and it contains EOF it will end up returning an error. This is
8975 caused by input 46 bytes long. The cause is due to the way base64
8976 BIOs find the start of base64 encoded data. They do this by trying a
8977 trial decode on each line until they find one that works. When they
8978 do a flag is set and it starts again knowing it can pass all the
8979 data directly through the decoder. Unfortunately it doesn't reset
8980 the context it uses. This means that if EOF is reached an attempt
8981 is made to pass two EOFs through the context and this causes the
8982 resulting error. This can also cause other problems as well. As is
8983 usual with these problems it takes *ages* to find and the fix is
8984 trivial: move one line.
8985 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8987 *) Ugly workaround to get s_client and s_server working under Windows. The
8988 old code wouldn't work because it needed to select() on sockets and the
8989 tty (for keypresses and to see if data could be written). Win32 only
8990 supports select() on sockets so we select() with a 1s timeout on the
8991 sockets and then see if any characters are waiting to be read, if none
8992 are present then we retry, we also assume we can always write data to
8993 the tty. This isn't nice because the code then blocks until we've
8994 received a complete line of data and it is effectively polling the
8995 keyboard at 1s intervals: however it's quite a bit better than not
8996 working at all :-) A dedicated Windows application might handle this
8997 with an event loop for example.
9000 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9001 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9002 will be called when RSA_sign() and RSA_verify() are used. This is useful
9003 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9004 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9005 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9006 This necessitated the support of an extra signature type NID_md5_sha1
9007 for SSL signatures and modifications to the SSL library to use it instead
9008 of calling RSA_public_decrypt() and RSA_private_encrypt().
9011 *) Add new -verify -CAfile and -CApath options to the crl program, these
9012 will lookup a CRL issuers certificate and verify the signature in a
9013 similar way to the verify program. Tidy up the crl program so it
9014 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9015 less strict. It will now permit CRL extensions even if it is not
9016 a V2 CRL: this will allow it to tolerate some broken CRLs.
9019 *) Initialize all non-automatic variables each time one of the openssl
9020 sub-programs is started (this is necessary as they may be started
9021 multiple times from the "OpenSSL>" prompt).
9022 [Lennart Bang, Bodo Moeller]
9024 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9025 removing all other RSA functionality (this is what NO_RSA does). This
9026 is so (for example) those in the US can disable those operations covered
9027 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9031 *) Non-copying interface to BIO pairs.
9032 (still largely untested)
9035 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9036 ASCII string. This was handled independently in various places before.
9039 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9040 UTF8 strings a character at a time.
9043 *) Use client_version from client hello to select the protocol
9044 (s23_srvr.c) and for RSA client key exchange verification
9045 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9048 *) Add various utility functions to handle SPKACs, these were previously
9049 handled by poking round in the structure internals. Added new function
9050 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9051 print, verify and generate SPKACs. Based on an original idea from
9052 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9055 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9058 *) Allow the config file extension section to be overwritten on the
9059 command line. Based on an original idea from Massimiliano Pala
9060 <madwolf@comune.modena.it>. The new option is called -extensions
9061 and can be applied to ca, req and x509. Also -reqexts to override
9062 the request extensions in req and -crlexts to override the crl extensions
9066 *) Add new feature to the SPKAC handling in ca. Now you can include
9067 the same field multiple times by preceding it by "XXXX." for example:
9070 this is the same syntax as used in the req config file.
9073 *) Allow certificate extensions to be added to certificate requests. These
9074 are specified in a 'req_extensions' option of the req section of the
9075 config file. They can be printed out with the -text option to req but
9076 are otherwise ignored at present.
9079 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9080 data read consists of only the final block it would not decrypted because
9081 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9082 A misplaced 'break' also meant the decrypted final block might not be
9083 copied until the next read.
9086 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9087 a few extra parameters to the DH structure: these will be useful if
9088 for example we want the value of 'q' or implement X9.42 DH.
9091 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9092 provides hooks that allow the default DSA functions or functions on a
9093 "per key" basis to be replaced. This allows hardware acceleration and
9094 hardware key storage to be handled without major modification to the
9095 library. Also added low level modexp hooks and CRYPTO_EX structure and
9096 associated functions.
9099 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9100 as "read only": it can't be written to and the buffer it points to will
9101 not be freed. Reading from a read only BIO is much more efficient than
9102 a normal memory BIO. This was added because there are several times when
9103 an area of memory needs to be read from a BIO. The previous method was
9104 to create a memory BIO and write the data to it, this results in two
9105 copies of the data and an O(n^2) reading algorithm. There is a new
9106 function BIO_new_mem_buf() which creates a read only memory BIO from
9107 an area of memory. Also modified the PKCS#7 routines to use read only
9111 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9112 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9113 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9114 but a retry condition occured while trying to read the rest.
9117 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9118 NID_pkcs7_encrypted by default: this was wrong since this should almost
9119 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9120 the encrypted data type: this is a more sensible place to put it and it
9121 allows the PKCS#12 code to be tidied up that duplicated this
9125 *) Changed obj_dat.pl script so it takes its input and output files on
9126 the command line. This should avoid shell escape redirection problems
9130 *) Initial support for certificate extension requests, these are included
9131 in things like Xenroll certificate requests. Included functions to allow
9132 extensions to be obtained and added.
9135 *) -crlf option to s_client and s_server for sending newlines as
9136 CRLF (as required by many protocols).
9139 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9141 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9142 [Ralf S. Engelschall]
9144 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9145 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9147 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
9151 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
9152 DH parameters/keys (q is lost during that conversion, but the resulting
9153 DH parameters contain its length).
9155 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
9156 much faster than DH_generate_parameters (which creates parameters
9157 where p = 2*q + 1), and also the smaller q makes DH computations
9158 much more efficient (160-bit exponentiation instead of 1024-bit
9159 exponentiation); so this provides a convenient way to support DHE
9160 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
9161 utter importance to use
9162 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9164 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9165 when such DH parameters are used, because otherwise small subgroup
9166 attacks may become possible!
9169 *) Avoid memory leak in i2d_DHparams.
9172 *) Allow the -k option to be used more than once in the enc program:
9173 this allows the same encrypted message to be read by multiple recipients.
9176 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
9177 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
9178 it will always use the numerical form of the OID, even if it has a short
9182 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
9183 method only got called if p,q,dmp1,dmq1,iqmp components were present,
9184 otherwise bn_mod_exp was called. In the case of hardware keys for example
9185 no private key components need be present and it might store extra data
9186 in the RSA structure, which cannot be accessed from bn_mod_exp.
9187 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
9188 private key operations.
9191 *) Added support for SPARC Linux.
9194 *) pem_password_cb function type incompatibly changed from
9195 typedef int pem_password_cb(char *buf, int size, int rwflag);
9197 ....(char *buf, int size, int rwflag, void *userdata);
9198 so that applications can pass data to their callbacks:
9199 The PEM[_ASN1]_{read,write}... functions and macros now take an
9200 additional void * argument, which is just handed through whenever
9201 the password callback is called.
9202 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
9204 New function SSL_CTX_set_default_passwd_cb_userdata.
9206 Compatibility note: As many C implementations push function arguments
9207 onto the stack in reverse order, the new library version is likely to
9208 interoperate with programs that have been compiled with the old
9209 pem_password_cb definition (PEM_whatever takes some data that
9210 happens to be on the stack as its last argument, and the callback
9211 just ignores this garbage); but there is no guarantee whatsoever that
9214 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
9215 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
9216 problems not only on Windows, but also on some Unix platforms.
9217 To avoid problematic command lines, these definitions are now in an
9218 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
9219 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
9222 *) MIPS III/IV assembler module is reimplemented.
9225 *) More DES library cleanups: remove references to srand/rand and
9226 delete an unused file.
9229 *) Add support for the the free Netwide assembler (NASM) under Win32,
9230 since not many people have MASM (ml) and it can be hard to obtain.
9231 This is currently experimental but it seems to work OK and pass all
9232 the tests. Check out INSTALL.W32 for info.
9235 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
9236 without temporary keys kept an extra copy of the server key,
9237 and connections with temporary keys did not free everything in case
9241 *) New function RSA_check_key and new openssl rsa option -check
9242 for verifying the consistency of RSA keys.
9243 [Ulf Moeller, Bodo Moeller]
9245 *) Various changes to make Win32 compile work:
9246 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
9247 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
9248 comparison" warnings.
9249 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
9252 *) Add a debugging option to PKCS#5 v2 key generation function: when
9253 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
9254 derived keys are printed to stderr.
9257 *) Copy the flags in ASN1_STRING_dup().
9258 [Roman E. Pavlov <pre@mo.msk.ru>]
9260 *) The x509 application mishandled signing requests containing DSA
9261 keys when the signing key was also DSA and the parameters didn't match.
9263 It was supposed to omit the parameters when they matched the signing key:
9264 the verifying software was then supposed to automatically use the CA's
9265 parameters if they were absent from the end user certificate.
9267 Omitting parameters is no longer recommended. The test was also
9268 the wrong way round! This was probably due to unusual behaviour in
9269 EVP_cmp_parameters() which returns 1 if the parameters match.
9270 This meant that parameters were omitted when they *didn't* match and
9271 the certificate was useless. Certificates signed with 'ca' didn't have
9273 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
9275 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
9276 The interface is as follows:
9277 Applications can use
9278 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
9279 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
9280 "off" is now the default.
9281 The library internally uses
9282 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
9283 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
9284 to disable memory-checking temporarily.
9286 Some inconsistent states that previously were possible (and were
9287 even the default) are now avoided.
9289 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
9290 with each memory chunk allocated; this is occasionally more helpful
9291 than just having a counter.
9293 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
9295 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
9299 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
9300 which largely parallels "options", but is for changing API behaviour,
9301 whereas "options" are about protocol behaviour.
9302 Initial "mode" flags are:
9304 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
9305 a single record has been written.
9306 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
9307 retries use the same buffer location.
9308 (But all of the contents must be
9312 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
9315 *) Fix problems with no-hmac etc.
9316 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
9318 *) New functions RSA_get_default_method(), RSA_set_method() and
9319 RSA_get_method(). These allows replacement of RSA_METHODs without having
9320 to mess around with the internals of an RSA structure.
9323 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
9324 Also really enable memory leak checks in openssl.c and in some
9326 [Chad C. Mulligan, Bodo Moeller]
9328 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
9329 up the length of negative integers. This has now been simplified to just
9330 store the length when it is first determined and use it later, rather
9331 than trying to keep track of where data is copied and updating it to
9333 [Steve Henson, reported by Brien Wheeler
9334 <bwheeler@authentica-security.com>]
9336 *) Add a new function PKCS7_signatureVerify. This allows the verification
9337 of a PKCS#7 signature but with the signing certificate passed to the
9338 function itself. This contrasts with PKCS7_dataVerify which assumes the
9339 certificate is present in the PKCS#7 structure. This isn't always the
9340 case: certificates can be omitted from a PKCS#7 structure and be
9341 distributed by "out of band" means (such as a certificate database).
9344 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
9345 function prototypes in pem.h, also change util/mkdef.pl to add the
9346 necessary function names.
9349 *) mk1mf.pl (used by Windows builds) did not properly read the
9350 options set by Configure in the top level Makefile, and Configure
9351 was not even able to write more than one option correctly.
9352 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
9355 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
9356 file to be loaded from a BIO or FILE pointer. The BIO version will
9357 for example allow memory BIOs to contain config info.
9360 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
9361 Whoever hopes to achieve shared-library compatibility across versions
9362 must use this, not the compile-time macro.
9363 (Exercise 0.9.4: Which is the minimum library version required by
9365 Note: All this applies only to multi-threaded programs, others don't
9369 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
9370 through a BIO pair triggered the default case, i.e.
9371 SSLerr(...,SSL_R_UNKNOWN_STATE).
9374 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
9375 can use the SSL library even if none of the specific BIOs is
9379 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
9380 for the encoded length.
9381 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
9383 *) Add initial documentation of the X509V3 functions.
9386 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
9387 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
9388 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
9389 secure PKCS#8 private key format with a high iteration count.
9392 *) Fix determination of Perl interpreter: A perl or perl5
9393 _directory_ in $PATH was also accepted as the interpreter.
9394 [Ralf S. Engelschall]
9396 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
9397 wrong with it but it was very old and did things like calling
9398 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
9402 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
9403 to use the new extension code.
9406 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
9407 with macros. This should make it easier to change their form, add extra
9408 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
9412 *) Add to configuration table a new entry that can specify an alternative
9413 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
9414 according to Mark Crispin <MRC@Panda.COM>.
9418 *) DES CBC did not update the IV. Weird.
9421 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
9422 Changing the behaviour of the former might break existing programs --
9423 where IV updating is needed, des_ncbc_encrypt can be used.
9426 *) When bntest is run from "make test" it drives bc to check its
9427 calculations, as well as internally checking them. If an internal check
9428 fails, it needs to cause bc to give a non-zero result or make test carries
9429 on without noticing the failure. Fixed.
9432 *) DES library cleanups.
9435 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
9436 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
9437 ciphers. NOTE: although the key derivation function has been verified
9438 against some published test vectors it has not been extensively tested
9439 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
9443 *) Instead of "mkdir -p", which is not fully portable, use new
9444 Perl script "util/mkdir-p.pl".
9447 *) Rewrite the way password based encryption (PBE) is handled. It used to
9448 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
9449 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
9450 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
9451 the 'parameter' field of the AlgorithmIdentifier is passed to the
9452 underlying key generation function so it must do its own ASN1 parsing.
9453 This has also changed the EVP_PBE_CipherInit() function which now has a
9454 'parameter' argument instead of literal salt and iteration count values
9455 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
9458 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
9459 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
9460 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
9461 KEY" because this clashed with PKCS#8 unencrypted string. Since this
9462 value was just used as a "magic string" and not used directly its
9463 value doesn't matter.
9466 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9470 *) "linux-sparc64" configuration (ultrapenguin).
9471 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9472 "linux-sparc" configuration.
9473 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9475 *) config now generates no-xxx options for missing ciphers.
9478 *) Support the EBCDIC character set (work in progress).
9479 File ebcdic.c not yet included because it has a different license.
9480 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9482 *) Support BS2000/OSD-POSIX.
9483 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9485 *) Make callbacks for key generation use void * instead of char *.
9488 *) Make S/MIME samples compile (not yet tested).
9491 *) Additional typesafe stacks.
9494 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9498 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9500 *) New configuration variant "sco5-gcc".
9502 *) Updated some demos.
9503 [Sean O Riordain, Wade Scholine]
9505 *) Add missing BIO_free at exit of pkcs12 application.
9508 *) Fix memory leak in conf.c.
9511 *) Updates for Win32 to assembler version of MD5.
9514 *) Set #! path to perl in apps/der_chop to where we found it
9515 instead of using a fixed path.
9518 *) SHA library changes for irix64-mips4-cc.
9521 *) Improvements for VMS support.
9525 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9527 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9528 This also avoids the problems with SC4.2 and unpatched SC5.
9529 [Andy Polyakov <appro@fy.chalmers.se>]
9531 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9532 These are required because of the typesafe stack would otherwise break
9533 existing code. If old code used a structure member which used to be STACK
9534 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9535 sk_num or sk_value it would produce an error because the num, data members
9536 are not present in STACK_OF. Now it just produces a warning. sk_set
9537 replaces the old method of assigning a value to sk_value
9538 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9539 that does this will no longer work (and should use sk_set instead) but
9540 this could be regarded as a "questionable" behaviour anyway.
9543 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9544 correctly handle encrypted S/MIME data.
9547 *) Change type of various DES function arguments from des_cblock
9548 (which means, in function argument declarations, pointer to char)
9549 to des_cblock * (meaning pointer to array with 8 char elements),
9550 which allows the compiler to do more typechecking; it was like
9551 that back in SSLeay, but with lots of ugly casts.
9553 Introduce new type const_des_cblock.
9556 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9557 problems: find RecipientInfo structure that matches recipient certificate
9558 and initialise the ASN1 structures properly based on passed cipher.
9561 *) Belatedly make the BN tests actually check the results.
9564 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9565 to and from BNs: it was completely broken. New compilation option
9566 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9567 key elements as negative integers.
9570 *) Reorganize and speed up MD5.
9571 [Andy Polyakov <appro@fy.chalmers.se>]
9574 [Richard Levitte <richard@levitte.org>]
9576 *) New option -out to asn1parse to allow the parsed structure to be
9577 output to a file. This is most useful when combined with the -strparse
9578 option to examine the output of things like OCTET STRINGS.
9581 *) Make SSL library a little more fool-proof by not requiring any longer
9582 that SSL_set_{accept,connect}_state be called before
9583 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9584 in many applications because usually everything *appeared* to work as
9585 intended anyway -- now it really works as intended).
9588 *) Move openssl.cnf out of lib/.
9591 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
9592 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
9593 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
9594 [Ralf S. Engelschall]
9596 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
9597 handle PKCS#7 enveloped data properly.
9598 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
9600 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
9601 copying pointers. The cert_st handling is changed by this in
9602 various ways (and thus what used to be known as ctx->default_cert
9603 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
9604 any longer when s->cert does not give us what we need).
9605 ssl_cert_instantiate becomes obsolete by this change.
9606 As soon as we've got the new code right (possibly it already is?),
9607 we have solved a couple of bugs of the earlier code where s->cert
9608 was used as if it could not have been shared with other SSL structures.
9610 Note that using the SSL API in certain dirty ways now will result
9611 in different behaviour than observed with earlier library versions:
9612 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
9613 does not influence s as it used to.
9615 In order to clean up things more thoroughly, inside SSL_SESSION
9616 we don't use CERT any longer, but a new structure SESS_CERT
9617 that holds per-session data (if available); currently, this is
9618 the peer's certificate chain and, for clients, the server's certificate
9619 and temporary key. CERT holds only those values that can have
9620 meaningful defaults in an SSL_CTX.
9623 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
9624 from the internal representation. Various PKCS#7 fixes: remove some
9625 evil casts and set the enc_dig_alg field properly based on the signing
9629 *) Allow PKCS#12 password to be set from the command line or the
9630 environment. Let 'ca' get its config file name from the environment
9631 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
9635 *) Allow certificate policies extension to use an IA5STRING for the
9636 organization field. This is contrary to the PKIX definition but
9637 VeriSign uses it and IE5 only recognises this form. Document 'x509'
9641 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
9642 without disallowing inline assembler and the like for non-pedantic builds.
9645 *) Support Borland C++ builder.
9646 [Janez Jere <jj@void.si>, modified by Ulf Möller]
9651 *) SHA-1 cleanups and performance enhancements.
9652 [Andy Polyakov <appro@fy.chalmers.se>]
9654 *) Sparc v8plus assembler for the bignum library.
9655 [Andy Polyakov <appro@fy.chalmers.se>]
9657 *) Accept any -xxx and +xxx compiler options in Configure.
9660 *) Update HPUX configuration.
9663 *) Add missing sk_<type>_unshift() function to safestack.h
9664 [Ralf S. Engelschall]
9666 *) New function SSL_CTX_use_certificate_chain_file that sets the
9667 "extra_cert"s in addition to the certificate. (This makes sense
9668 only for "PEM" format files, as chains as a whole are not
9672 *) Support verify_depth from the SSL API.
9673 x509_vfy.c had what can be considered an off-by-one-error:
9674 Its depth (which was not part of the external interface)
9675 was actually counting the number of certificates in a chain;
9676 now it really counts the depth.
9679 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
9680 instead of X509err, which often resulted in confusing error
9681 messages since the error codes are not globally unique
9682 (e.g. an alleged error in ssl3_accept when a certificate
9683 didn't match the private key).
9685 *) New function SSL_CTX_set_session_id_context that allows to set a default
9686 value (so that you don't need SSL_set_session_id_context for each
9687 connection using the SSL_CTX).
9690 *) OAEP decoding bug fix.
9693 *) Support INSTALL_PREFIX for package builders, as proposed by
9697 *) New Configure options "threads" and "no-threads". For systems
9698 where the proper compiler options are known (currently Solaris
9699 and Linux), "threads" is the default.
9702 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
9705 *) Install various scripts to $(OPENSSLDIR)/misc, not to
9706 $(INSTALLTOP)/bin -- they shouldn't clutter directories
9707 such as /usr/local/bin.
9710 *) "make linux-shared" to build shared libraries.
9711 [Niels Poppe <niels@netbox.org>]
9713 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
9716 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
9717 extension adding in x509 utility.
9720 *) Remove NOPROTO sections and error code comments.
9723 *) Partial rewrite of the DEF file generator to now parse the ANSI
9727 *) New Configure options --prefix=DIR and --openssldir=DIR.
9730 *) Complete rewrite of the error code script(s). It is all now handled
9731 by one script at the top level which handles error code gathering,
9732 header rewriting and C source file generation. It should be much better
9733 than the old method: it now uses a modified version of Ulf's parser to
9734 read the ANSI prototypes in all header files (thus the old K&R definitions
9735 aren't needed for error creation any more) and do a better job of
9736 translating function codes into names. The old 'ASN1 error code imbedded
9737 in a comment' is no longer necessary and it doesn't use .err files which
9738 have now been deleted. Also the error code call doesn't have to appear all
9739 on one line (which resulted in some large lines...).
9742 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
9745 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
9746 0 (which usually indicates a closed connection), but continue reading.
9749 *) Fix some race conditions.
9752 *) Add support for CRL distribution points extension. Add Certificate
9753 Policies and CRL distribution points documentation.
9756 *) Move the autogenerated header file parts to crypto/opensslconf.h.
9759 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
9760 8 of keying material. Merlin has also confirmed interop with this fix
9761 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
9762 [Merlin Hughes <merlin@baltimore.ie>]
9764 *) Fix lots of warnings.
9765 [Richard Levitte <levitte@stacken.kth.se>]
9767 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
9768 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9769 [Richard Levitte <levitte@stacken.kth.se>]
9771 *) Fix problems with sizeof(long) == 8.
9772 [Andy Polyakov <appro@fy.chalmers.se>]
9774 *) Change functions to ANSI C.
9777 *) Fix typos in error codes.
9778 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9780 *) Remove defunct assembler files from Configure.
9783 *) SPARC v8 assembler BIGNUM implementation.
9784 [Andy Polyakov <appro@fy.chalmers.se>]
9786 *) Support for Certificate Policies extension: both print and set.
9787 Various additions to support the r2i method this uses.
9790 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9791 return a const string when you are expecting an allocated buffer.
9794 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9795 types DirectoryString and DisplayText.
9798 *) Add code to allow r2i extensions to access the configuration database,
9799 add an LHASH database driver and add several ctx helper functions.
9802 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9803 fail when they extended the size of a BIGNUM.
9806 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9807 support typesafe stack.
9810 *) Fix typo in SSL_[gs]et_options().
9811 [Nils Frostberg <nils@medcom.se>]
9813 *) Delete various functions and files that belonged to the (now obsolete)
9814 old X509V3 handling code.
9817 *) New Configure option "rsaref".
9820 *) Don't auto-generate pem.h.
9823 *) Introduce type-safe ASN.1 SETs.
9826 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9827 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9829 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9830 that links with OpenSSL (well at least cause lots of warnings), but fear
9831 not: the conversion is trivial, and it eliminates loads of evil casts. A
9832 few STACKed things have been converted already. Feel free to convert more.
9833 In the fullness of time, I'll do away with the STACK type altogether.
9836 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9837 specified in <certfile> by updating the entry in the index.txt file.
9838 This way one no longer has to edit the index.txt file manually for
9839 revoking a certificate. The -revoke option does the gory details now.
9840 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9842 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9843 `-text' option at all and this way the `-noout -text' combination was
9844 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9845 [Ralf S. Engelschall]
9847 *) Make sure a corresponding plain text error message exists for the
9848 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9849 verify callback function determined that a certificate was revoked.
9850 [Ralf S. Engelschall]
9852 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9853 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9854 all available cipers including rc5, which was forgotten until now.
9855 In order to let the testing shell script know which algorithms
9856 are available, a new (up to now undocumented) command
9857 "openssl list-cipher-commands" is used.
9860 *) Bugfix: s_client occasionally would sleep in select() when
9861 it should have checked SSL_pending() first.
9864 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9865 the raw DSA values prior to ASN.1 encoding.
9868 *) Tweaks to Configure
9869 [Niels Poppe <niels@netbox.org>]
9871 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9875 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9878 *) New config option to avoid instructions that are illegal on the 80386.
9879 The default code is faster, but requires at least a 486.
9882 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9883 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9884 same as SSL2_VERSION anyway.
9887 *) New "-showcerts" option for s_client.
9890 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9891 application. Various cleanups and fixes.
9894 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9895 modify error routines to work internally. Add error codes and PBE init
9896 to library startup routines.
9899 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9900 packing functions to asn1 and evp. Changed function names and error
9901 codes along the way.
9904 *) PKCS12 integration: and so it begins... First of several patches to
9905 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9906 objects to objects.h
9909 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9910 and display support for Thawte strong extranet extension.
9913 *) Add LinuxPPC support.
9914 [Jeff Dubrule <igor@pobox.org>]
9916 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9917 bn_div_words in alpha.s.
9918 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9920 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9921 OAEP isn't supported when OpenSSL is built with RSAref.
9922 [Ulf Moeller <ulf@fitug.de>]
9924 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9925 so they no longer are missing under -DNOPROTO.
9926 [Soren S. Jorvang <soren@t.dk>]
9929 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9931 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9932 doesn't work when the session is reused. Coming soon!
9935 *) Fix a security hole, that allows sessions to be reused in the wrong
9936 context thus bypassing client cert protection! All software that uses
9937 client certs and session caches in multiple contexts NEEDS PATCHING to
9938 allow session reuse! A fuller solution is in the works.
9939 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9941 *) Some more source tree cleanups (removed obsolete files
9942 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9943 permission on "config" script to be executable) and a fix for the INSTALL
9945 [Ulf Moeller <ulf@fitug.de>]
9947 *) Remove some legacy and erroneous uses of malloc, free instead of
9949 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9951 *) Make rsa_oaep_test return non-zero on error.
9952 [Ulf Moeller <ulf@fitug.de>]
9954 *) Add support for native Solaris shared libraries. Configure
9955 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9956 if someone would make that last step automatic.
9957 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9959 *) ctx_size was not built with the right compiler during "make links". Fixed.
9962 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9963 except NULL ciphers". This means the default cipher list will no longer
9964 enable NULL ciphers. They need to be specifically enabled e.g. with
9965 the string "DEFAULT:eNULL".
9968 *) Fix to RSA private encryption routines: if p < q then it would
9969 occasionally produce an invalid result. This will only happen with
9970 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9973 *) Be less restrictive and allow also `perl util/perlpath.pl
9974 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9975 because this way one can also use an interpreter named `perl5' (which is
9976 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9977 installed as `perl').
9978 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9980 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9981 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9983 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9984 advapi32.lib to Win32 build and change the pem test comparision
9985 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9986 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9987 and crypto/des/ede_cbcm_enc.c.
9990 *) DES quad checksum was broken on big-endian architectures. Fixed.
9993 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9994 Win32 test batch file so it (might) work again. The Win32 test batch file
9995 is horrible: I feel ill....
9998 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9999 in e_os.h. Audit of header files to check ANSI and non ANSI
10000 sections: 10 functions were absent from non ANSI section and not exported
10001 from Windows DLLs. Fixed up libeay.num for new functions.
10004 *) Make `openssl version' output lines consistent.
10005 [Ralf S. Engelschall]
10007 *) Fix Win32 symbol export lists for BIO functions: Added
10008 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10009 to ms/libeay{16,32}.def.
10010 [Ralf S. Engelschall]
10012 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10013 fine under Unix and passes some trivial tests I've now added. But the
10014 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10015 added to make sure no one expects that this stuff really works in the
10016 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10017 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10019 [Ralf S. Engelschall]
10021 *) Fix the generation of two part addresses in perl.
10022 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10024 *) Add config entry for Linux on MIPS.
10025 [John Tobey <jtobey@channel1.com>]
10027 *) Make links whenever Configure is run, unless we are on Windoze.
10030 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10031 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10035 *) Add a useful kludge to allow package maintainers to specify compiler and
10036 other platforms details on the command line without having to patch the
10037 Configure script everytime: One now can use ``perl Configure
10038 <id>:<details>'', i.e. platform ids are allowed to have details appended
10039 to them (seperated by colons). This is treated as there would be a static
10040 pre-configured entry in Configure's %table under key <id> with value
10041 <details> and ``perl Configure <id>'' is called. So, when you want to
10042 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10043 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10044 now, which overrides the FreeBSD-elf entry on-the-fly.
10045 [Ralf S. Engelschall]
10047 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10050 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10051 on the `perl Configure ...' command line. This way one can compile
10052 OpenSSL libraries with Position Independent Code (PIC) which is needed
10053 for linking it into DSOs.
10054 [Ralf S. Engelschall]
10056 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10060 *) Cleaned up the LICENSE document: The official contact for any license
10061 questions now is the OpenSSL core team under openssl-core@openssl.org.
10062 And add a paragraph about the dual-license situation to make sure people
10063 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10064 to the OpenSSL toolkit.
10065 [Ralf S. Engelschall]
10067 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10068 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10069 Additonally cleaned up the `make links' target: Remove unnecessary
10070 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10071 to speed processing and no longer clutter the display with confusing
10072 stuff. Instead only the actually done links are displayed.
10073 [Ralf S. Engelschall]
10075 *) Permit null encryption ciphersuites, used for authentication only. It used
10076 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10077 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10081 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10082 signed attributes when verifying signatures (this would break them),
10083 the detached data encoding was wrong and public keys obtained using
10084 X509_get_pubkey() weren't freed.
10087 *) Add text documentation for the BUFFER functions. Also added a work around
10088 to a Win95 console bug. This was triggered by the password read stuff: the
10089 last character typed gets carried over to the next fread(). If you were
10090 generating a new cert request using 'req' for example then the last
10091 character of the passphrase would be CR which would then enter the first
10095 *) Added the new `Includes OpenSSL Cryptography Software' button as
10096 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10097 button and can be used by applications based on OpenSSL to show the
10098 relationship to the OpenSSL project.
10099 [Ralf S. Engelschall]
10101 *) Remove confusing variables in function signatures in files
10102 ssl/ssl_lib.c and ssl/ssl.h.
10103 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10105 *) Don't install bss_file.c under PREFIX/include/
10106 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10108 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10109 functions that return function pointers and has support for NT specific
10110 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10111 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10112 unsigned to signed types: this was killing the Win32 compile.
10115 *) Add new certificate file to stack functions,
10116 SSL_add_dir_cert_subjects_to_stack() and
10117 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10118 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10119 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10120 This means that Apache-SSL and similar packages don't have to mess around
10121 to add as many CAs as they want to the preferred list.
10124 *) Experiment with doxygen documentation. Currently only partially applied to
10126 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10127 openssl.doxy as the configuration file.
10130 *) Get rid of remaining C++-style comments which strict C compilers hate.
10131 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10133 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10134 compiled in by default: it has problems with large keys.
10137 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10138 DH private keys and/or callback functions which directly correspond to
10139 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10140 is needed for applications which have to configure certificates on a
10141 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10143 For the RSA certificate situation is makes no difference, but
10144 for the DSA certificate situation this fixes the "no shared cipher"
10145 problem where the OpenSSL cipher selection procedure failed because the
10146 temporary keys were not overtaken from the context and the API provided
10147 no way to reconfigure them.
10148 The new functions now let applications reconfigure the stuff and they
10149 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
10150 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
10151 non-public-API function ssl_cert_instantiate() is used as a helper
10152 function and also to reduce code redundancy inside ssl_rsa.c.
10153 [Ralf S. Engelschall]
10155 *) Move s_server -dcert and -dkey options out of the undocumented feature
10156 area because they are useful for the DSA situation and should be
10157 recognized by the users.
10158 [Ralf S. Engelschall]
10160 *) Fix the cipher decision scheme for export ciphers: the export bits are
10161 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
10162 SSL_EXP_MASK. So, the original variable has to be used instead of the
10163 already masked variable.
10164 [Richard Levitte <levitte@stacken.kth.se>]
10166 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
10167 [Richard Levitte <levitte@stacken.kth.se>]
10169 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
10170 from `int' to `unsigned int' because it's a length and initialized by
10171 EVP_DigestFinal() which expects an `unsigned int *'.
10172 [Richard Levitte <levitte@stacken.kth.se>]
10174 *) Don't hard-code path to Perl interpreter on shebang line of Configure
10175 script. Instead use the usual Shell->Perl transition trick.
10176 [Ralf S. Engelschall]
10178 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
10179 (in addition to RSA certificates) to match the behaviour of `openssl dsa
10180 -noout -modulus' as it's already the case for `openssl rsa -noout
10181 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
10182 currently the public key is printed (a decision which was already done by
10183 `openssl dsa -modulus' in the past) which serves a similar purpose.
10184 Additionally the NO_RSA no longer completely removes the whole -modulus
10185 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
10187 [Ralf S. Engelschall]
10189 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
10190 BIO. See the source (crypto/evp/bio_ok.c) for more info.
10191 [Arne Ansper <arne@ats.cyber.ee>]
10193 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
10194 to be added. Now both 'req' and 'ca' can use new objects defined in the
10198 *) Add cool BIO that does syslog (or event log on NT).
10199 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
10201 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
10202 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
10203 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
10204 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
10207 *) Add preliminary config info for new extension code.
10210 *) Make RSA_NO_PADDING really use no padding.
10211 [Ulf Moeller <ulf@fitug.de>]
10213 *) Generate errors when private/public key check is done.
10216 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
10217 for some CRL extensions and new objects added.
10220 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
10221 key usage extension and fuller support for authority key id.
10224 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
10225 padding method for RSA, which is recommended for new applications in PKCS
10226 #1 v2.0 (RFC 2437, October 1998).
10227 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
10228 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
10229 against Bleichbacher's attack on RSA.
10230 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
10233 *) Updates to the new SSL compression code
10234 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10236 *) Fix so that the version number in the master secret, when passed
10237 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
10238 (because the server will not accept higher), that the version number
10239 is 0x03,0x01, not 0x03,0x00
10240 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10242 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
10243 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
10244 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
10247 *) Support for RAW extensions where an arbitrary extension can be
10248 created by including its DER encoding. See apps/openssl.cnf for
10252 *) Make sure latest Perl versions don't interpret some generated C array
10253 code as Perl array code in the crypto/err/err_genc.pl script.
10254 [Lars Weber <3weber@informatik.uni-hamburg.de>]
10256 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
10257 not many people have the assembler. Various Win32 compilation fixes and
10258 update to the INSTALL.W32 file with (hopefully) more accurate Win32
10259 build instructions.
10262 *) Modify configure script 'Configure' to automatically create crypto/date.h
10263 file under Win32 and also build pem.h from pem.org. New script
10264 util/mkfiles.pl to create the MINFO file on environments that can't do a
10265 'make files': perl util/mkfiles.pl >MINFO should work.
10268 *) Major rework of DES function declarations, in the pursuit of correctness
10269 and purity. As a result, many evil casts evaporated, and some weirdness,
10270 too. You may find this causes warnings in your code. Zapping your evil
10271 casts will probably fix them. Mostly.
10274 *) Fix for a typo in asn1.h. Bug fix to object creation script
10275 obj_dat.pl. It considered a zero in an object definition to mean
10276 "end of object": none of the objects in objects.h have any zeros
10277 so it wasn't spotted.
10278 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
10280 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
10281 Masking (CBCM). In the absence of test vectors, the best I have been able
10282 to do is check that the decrypt undoes the encrypt, so far. Send me test
10283 vectors if you have them.
10286 *) Correct calculation of key length for export ciphers (too much space was
10287 allocated for null ciphers). This has not been tested!
10290 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
10291 message is now correct (it understands "crypto" and "ssl" on its
10292 command line). There is also now an "update" option. This will update
10293 the util/ssleay.num and util/libeay.num files with any new functions.
10295 perl util/mkdef.pl crypto ssl update
10296 it will update them.
10299 *) Overhauled the Perl interface (perl/*):
10300 - ported BN stuff to OpenSSL's different BN library
10301 - made the perl/ source tree CVS-aware
10302 - renamed the package from SSLeay to OpenSSL (the files still contain
10303 their history because I've copied them in the repository)
10304 - removed obsolete files (the test scripts will be replaced
10305 by better Test::Harness variants in the future)
10306 [Ralf S. Engelschall]
10308 *) First cut for a very conservative source tree cleanup:
10309 1. merge various obsolete readme texts into doc/ssleay.txt
10310 where we collect the old documents and readme texts.
10311 2. remove the first part of files where I'm already sure that we no
10312 longer need them because of three reasons: either they are just temporary
10313 files which were left by Eric or they are preserved original files where
10314 I've verified that the diff is also available in the CVS via "cvs diff
10315 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
10316 the crypto/md/ stuff).
10317 [Ralf S. Engelschall]
10319 *) More extension code. Incomplete support for subject and issuer alt
10320 name, issuer and authority key id. Change the i2v function parameters
10321 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
10322 what that's for :-) Fix to ASN1 macro which messed up
10323 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
10326 *) Preliminary support for ENUMERATED type. This is largely copied from the
10330 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
10331 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10333 *) Make sure `make rehash' target really finds the `openssl' program.
10334 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10336 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
10337 like to hear about it if this slows down other processors.
10340 *) Add CygWin32 platform information to Configure script.
10341 [Alan Batie <batie@aahz.jf.intel.com>]
10343 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
10344 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
10346 *) New program nseq to manipulate netscape certificate sequences
10349 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
10353 *) Fixes to BN code. Previously the default was to define BN_RECURSION
10354 but the BN code had some problems that would cause failures when
10355 doing certificate verification and some other functions.
10356 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10358 *) Add ASN1 and PEM code to support netscape certificate sequences.
10361 *) Add ASN1 and PEM code to support netscape certificate sequences.
10364 *) Add several PKIX and private extended key usage OIDs.
10367 *) Modify the 'ca' program to handle the new extension code. Modify
10368 openssl.cnf for new extension format, add comments.
10371 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
10372 and add a sample to openssl.cnf so req -x509 now adds appropriate
10376 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
10377 error code, add initial support to X509_print() and x509 application.
10380 *) Takes a deep breath and start addding X509 V3 extension support code. Add
10381 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
10382 stuff is currently isolated and isn't even compiled yet.
10385 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
10386 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
10387 Removed the versions check from X509 routines when loading extensions:
10388 this allows certain broken certificates that don't set the version
10389 properly to be processed.
10392 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
10393 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
10394 can still be regenerated with "make depend".
10397 *) Spelling mistake in C version of CAST-128.
10398 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
10400 *) Changes to the error generation code. The perl script err-code.pl
10401 now reads in the old error codes and retains the old numbers, only
10402 adding new ones if necessary. It also only changes the .err files if new
10403 codes are added. The makefiles have been modified to only insert errors
10404 when needed (to avoid needlessly modifying header files). This is done
10405 by only inserting errors if the .err file is newer than the auto generated
10406 C file. To rebuild all the error codes from scratch (the old behaviour)
10407 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
10408 or delete all the .err files.
10411 *) CAST-128 was incorrectly implemented for short keys. The C version has
10412 been fixed, but is untested. The assembler versions are also fixed, but
10413 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
10414 to regenerate it if needed.
10415 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
10416 Hagino <itojun@kame.net>]
10418 *) File was opened incorrectly in randfile.c.
10419 [Ulf Möller <ulf@fitug.de>]
10421 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
10422 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
10423 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
10424 al: it's just almost always a UTCTime. Note this patch adds new error
10425 codes so do a "make errors" if there are problems.
10428 *) Correct Linux 1 recognition in config.
10429 [Ulf Möller <ulf@fitug.de>]
10431 *) Remove pointless MD5 hash when using DSA keys in ca.
10432 [Anonymous <nobody@replay.com>]
10434 *) Generate an error if given an empty string as a cert directory. Also
10435 generate an error if handed NULL (previously returned 0 to indicate an
10436 error, but didn't set one).
10437 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
10439 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
10442 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
10443 parameters. This was causing a warning which killed off the Win32 compile.
10446 *) Remove C++ style comments from crypto/bn/bn_local.h.
10447 [Neil Costigan <neil.costigan@celocom.com>]
10449 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
10450 based on a text string, looking up short and long names and finally
10451 "dot" format. The "dot" format stuff didn't work. Added new function
10452 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
10453 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
10454 OID is not part of the table.
10457 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
10458 X509_LOOKUP_by_alias().
10461 *) Sort openssl functions by name.
10464 *) Get the gendsa program working (hopefully) and add it to app list. Remove
10465 encryption from sample DSA keys (in case anyone is interested the password
10469 *) Make _all_ *_free functions accept a NULL pointer.
10470 [Frans Heymans <fheymans@isaserver.be>]
10472 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10474 [Anonymous <nobody@replay.com>]
10476 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10477 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10479 *) Don't blow it for numeric -newkey arguments to apps/req.
10480 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10482 *) Temp key "for export" tests were wrong in s3_srvr.c.
10483 [Anonymous <nobody@replay.com>]
10485 *) Add prototype for temp key callback functions
10486 SSL_CTX_set_tmp_{rsa,dh}_callback().
10489 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10490 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10493 *) X509_name_add_entry() freed the wrong thing after an error.
10494 [Arne Ansper <arne@ats.cyber.ee>]
10496 *) rsa_eay.c would attempt to free a NULL context.
10497 [Arne Ansper <arne@ats.cyber.ee>]
10499 *) BIO_s_socket() had a broken should_retry() on Windoze.
10500 [Arne Ansper <arne@ats.cyber.ee>]
10502 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10503 [Arne Ansper <arne@ats.cyber.ee>]
10505 *) Make sure the already existing X509_STORE->depth variable is initialized
10506 in X509_STORE_new(), but document the fact that this variable is still
10507 unused in the certificate verification process.
10508 [Ralf S. Engelschall]
10510 *) Fix the various library and apps files to free up pkeys obtained from
10511 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10514 *) Fix reference counting in X509_PUBKEY_get(). This makes
10515 demos/maurice/example2.c work, amongst others, probably.
10516 [Steve Henson and Ben Laurie]
10518 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10519 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10520 are no longer created. This way we have a single and consistent command
10521 line interface `openssl <command>', similar to `cvs <command>'.
10522 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10524 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10525 BIT STRING wrapper always have zero unused bits.
10528 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10531 *) Make the top-level INSTALL documentation easier to understand.
10534 *) Makefiles updated to exit if an error occurs in a sub-directory
10535 make (including if user presses ^C) [Paul Sutton]
10537 *) Make Montgomery context stuff explicit in RSA data structure.
10540 *) Fix build order of pem and err to allow for generated pem.h.
10543 *) Fix renumbering bug in X509_NAME_delete_entry().
10546 *) Enhanced the err-ins.pl script so it makes the error library number
10547 global and can add a library name. This is needed for external ASN1 and
10548 other error libraries.
10551 *) Fixed sk_insert which never worked properly.
10554 *) Fix ASN1 macros so they can handle indefinite length construted
10555 EXPLICIT tags. Some non standard certificates use these: they can now
10559 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10560 into a single doc/ssleay.txt bundle. This way the information is still
10561 preserved but no longer messes up this directory. Now it's new room for
10562 the new set of documenation files.
10563 [Ralf S. Engelschall]
10565 *) SETs were incorrectly DER encoded. This was a major pain, because they
10566 shared code with SEQUENCEs, which aren't coded the same. This means that
10567 almost everything to do with SETs or SEQUENCEs has either changed name or
10568 number of arguments.
10569 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10571 *) Fix test data to work with the above.
10574 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10575 was already fixed by Eric for 0.9.1 it seems.
10576 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10578 *) Autodetect FreeBSD3.
10581 *) Fix various bugs in Configure. This affects the following platforms:
10585 unixware-2.0-pentium
10589 *) Eliminate generated files from CVS. Reorder tests to regenerate files
10590 before they are needed.
10593 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
10597 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
10599 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
10600 changed SSLeay to OpenSSL in version strings.
10601 [Ralf S. Engelschall]
10603 *) Some fixups to the top-level documents.
10606 *) Fixed the nasty bug where rsaref.h was not found under compile-time
10607 because the symlink to include/ was missing.
10608 [Ralf S. Engelschall]
10610 *) Incorporated the popular no-RSA/DSA-only patches
10611 which allow to compile a RSA-free SSLeay.
10612 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
10614 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
10615 when "ssleay" is still not found.
10616 [Ralf S. Engelschall]
10618 *) Added more platforms to Configure: Cray T3E, HPUX 11,
10619 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
10621 *) Updated the README file.
10622 [Ralf S. Engelschall]
10624 *) Added various .cvsignore files in the CVS repository subdirs
10625 to make a "cvs update" really silent.
10626 [Ralf S. Engelschall]
10628 *) Recompiled the error-definition header files and added
10629 missing symbols to the Win32 linker tables.
10630 [Ralf S. Engelschall]
10632 *) Cleaned up the top-level documents;
10633 o new files: CHANGES and LICENSE
10634 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
10635 o merged COPYRIGHT into LICENSE
10636 o removed obsolete TODO file
10637 o renamed MICROSOFT to INSTALL.W32
10638 [Ralf S. Engelschall]
10640 *) Removed dummy files from the 0.9.1b source tree:
10641 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
10642 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
10643 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
10644 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
10645 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
10646 [Ralf S. Engelschall]
10648 *) Added various platform portability fixes.
10651 *) The Genesis of the OpenSSL rpject:
10652 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
10653 Young and Tim J. Hudson created while they were working for C2Net until
10655 [The OpenSSL Project]
10658 Changes between 0.9.0b and 0.9.1b [not released]
10660 *) Updated a few CA certificates under certs/
10663 *) Changed some BIGNUM api stuff.
10666 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
10667 DGUX x86, Linux Alpha, etc.
10670 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
10671 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
10675 *) Add -strparse option to asn1pars program which parses nested
10677 [Dr Stephen Henson <shenson@bigfoot.com>]
10679 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
10682 *) DSA fix for "ca" program.
10685 *) Added "-genkey" option to "dsaparam" program.
10688 *) Added RIPE MD160 (rmd160) message digest.
10691 *) Added -a (all) option to "ssleay version" command.
10694 *) Added PLATFORM define which is the id given to Configure.
10697 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
10700 *) Extended the ASN.1 parser routines.
10703 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
10706 *) Added a BN_CTX to the BN library.
10709 *) Fixed the weak key values in DES library
10712 *) Changed API in EVP library for cipher aliases.
10715 *) Added support for RC2/64bit cipher.
10718 *) Converted the lhash library to the crypto/mem.c functions.
10721 *) Added more recognized ASN.1 object ids.
10724 *) Added more RSA padding checks for SSL/TLS.
10727 *) Added BIO proxy/filter functionality.
10730 *) Added extra_certs to SSL_CTX which can be used
10731 send extra CA certificates to the client in the CA cert chain sending
10732 process. It can be configured with SSL_CTX_add_extra_chain_cert().
10735 *) Now Fortezza is denied in the authentication phase because
10736 this is key exchange mechanism is not supported by SSLeay at all.
10739 *) Additional PKCS1 checks.
10742 *) Support the string "TLSv1" for all TLS v1 ciphers.
10745 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
10746 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
10749 *) Fixed a few memory leaks.
10752 *) Fixed various code and comment typos.
10755 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
10756 bytes sent in the client random.
10757 [Edward Bishop <ebishop@spyglass.com>]