5 Changes between 1.0.2 and 1.1.0 [xx XXX xxxx]
7 *) Experimental support for a new, fast, unbiased prime candidate generator,
8 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
9 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
11 *) New output format NSS in the sess_id command line tool. This allows
12 exporting the session id and the master key in NSS keylog format.
13 [Martin Kaiser <martin@kaiser.cx>]
15 *) Harmonize version and its documentation. -f flag is used to display
17 [mancha <mancha1@zoho.com>]
19 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
21 [mancha <mancha1@zoho.com>]
23 *) Fix some double frees. These are not thought to be exploitable.
24 [mancha <mancha1@zoho.com>]
26 *) A missing bounds check in the handling of the TLS heartbeat extension
27 can be used to reveal up to 64k of memory to a connected client or
30 Thanks for Neel Mehta of Google Security for discovering this bug and to
31 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
32 preparing the fix (CVE-2014-0160)
33 [Adam Langley, Bodo Moeller]
35 *) Fix for the attack described in the paper "Recovering OpenSSL
36 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
37 by Yuval Yarom and Naomi Benger. Details can be obtained from:
38 http://eprint.iacr.org/2014/140
40 Thanks to Yuval Yarom and Naomi Benger for discovering this
41 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
42 [Yuval Yarom and Naomi Benger]
44 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
45 this fixes a limiation in previous versions of OpenSSL.
48 *) Experimental encrypt-then-mac support.
50 Experimental support for encrypt then mac from
51 draft-gutmann-tls-encrypt-then-mac-02.txt
53 To enable it set the appropriate extension number (0x42 for the test
54 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
56 For non-compliant peers (i.e. just about everything) this should have no
59 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
63 *) Add callbacks supporting generation and retrieval of supplemental
65 [Scott Deboy <sdeboy@apache.org>, Trevor Perrin and Ben Laurie]
67 *) Add EVP support for key wrapping algorithms, to avoid problems with
68 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
69 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
70 algorithms and include tests cases.
73 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
77 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
78 MGF1 digest and OAEP label.
81 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
82 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
83 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
86 *) Make openssl verify return errors.
87 [Chris Palmer <palmer@google.com> and Ben Laurie]
89 *) New function ASN1_TIME_diff to calculate the difference between two
90 ASN1_TIME structures or one structure and the current time.
93 *) Update fips_test_suite to support multiple command line options. New
94 test to induce all self test errors in sequence and check expected
98 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
99 sign or verify all in one operation.
102 *) Add fips_algvs: a multicall fips utility incorporaing all the algorithm
103 test programs and fips_test_suite. Includes functionality to parse
104 the minimal script output of fipsalgest.pl directly.
107 *) Add authorisation parameter to FIPS_module_mode_set().
110 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
113 *) Use separate DRBG fields for internal and external flags. New function
114 FIPS_drbg_health_check() to perform on demand health checking. Add
115 generation tests to fips_test_suite with reduced health check interval to
116 demonstrate periodic health checking. Add "nodh" option to
117 fips_test_suite to skip very slow DH test.
120 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
124 *) More extensive health check for DRBG checking many more failure modes.
125 New function FIPS_selftest_drbg_all() to handle every possible DRBG
126 combination: call this in fips_test_suite.
129 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
130 and POST to handle Dual EC cases.
133 *) Add support for canonical generation of DSA parameter 'g'. See
136 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
137 POST to handle HMAC cases.
140 *) Add functions FIPS_module_version() and FIPS_module_version_text()
141 to return numerical and string versions of the FIPS module number.
144 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
145 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
146 outside the validated module in the FIPS capable OpenSSL.
149 *) Minor change to DRBG entropy callback semantics. In some cases
150 there is no multiple of the block length between min_len and
151 max_len. Allow the callback to return more than max_len bytes
152 of entropy but discard any extra: it is the callback's responsibility
153 to ensure that the extra data discarded does not impact the
154 requested amount of entropy.
157 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
158 information in FIPS186-3, SP800-57 and SP800-131A.
161 *) CCM support via EVP. Interface is very similar to GCM case except we
162 must supply all data in one chunk (i.e. no update, final) and the
163 message length must be supplied if AAD is used. Add algorithm test
167 *) Initial version of POST overhaul. Add POST callback to allow the status
168 of POST to be monitored and/or failures induced. Modify fips_test_suite
169 to use callback. Always run all selftests even if one fails.
172 *) XTS support including algorithm test driver in the fips_gcmtest program.
173 Note: this does increase the maximum key length from 32 to 64 bytes but
174 there should be no binary compatibility issues as existing applications
175 will never use XTS mode.
178 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
179 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
180 performs algorithm blocking for unapproved PRNG types. Also do not
181 set PRNG type in FIPS_mode_set(): leave this to the application.
182 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
183 the standard OpenSSL PRNG: set additional data to a date time vector.
186 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
187 This shouldn't present any incompatibility problems because applications
188 shouldn't be using these directly and any that are will need to rethink
189 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
192 *) Extensive self tests and health checking required by SP800-90 DRBG.
193 Remove strength parameter from FIPS_drbg_instantiate and always
194 instantiate at maximum supported strength.
197 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
200 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
203 *) New function DH_compute_key_padded() to compute a DH key and pad with
204 leading zeroes if needed: this complies with SP800-56A et al.
207 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
208 anything, incomplete, subject to change and largely untested at present.
211 *) Modify fipscanisteronly build option to only build the necessary object
212 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
215 *) Add experimental option FIPSSYMS to give all symbols in
216 fipscanister.o and FIPS or fips prefix. This will avoid
217 conflicts with future versions of OpenSSL. Add perl script
218 util/fipsas.pl to preprocess assembly language source files
219 and rename any affected symbols.
222 *) Add selftest checks and algorithm block of non-fips algorithms in
223 FIPS mode. Remove DES2 from selftests.
226 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
227 return internal method without any ENGINE dependencies. Add new
228 tiny fips sign and verify functions.
231 *) New build option no-ec2m to disable characteristic 2 code.
234 *) New build option "fipscanisteronly". This only builds fipscanister.o
235 and (currently) associated fips utilities. Uses the file Makefile.fips
236 instead of Makefile.org as the prototype.
239 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
240 Update fips_gcmtest to use IV generator.
243 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
244 setting output buffer to NULL. The *Final function must be
245 called although it will not retrieve any additional data. The tag
246 can be set or retrieved with a ctrl. The IV length is by default 12
247 bytes (96 bits) but can be set to an alternative value. If the IV
248 length exceeds the maximum IV length (currently 16 bytes) it cannot be
252 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
253 underlying do_cipher function handles all cipher semantics itself
254 including padding and finalisation. This is useful if (for example)
255 an ENGINE cipher handles block padding itself. The behaviour of
256 do_cipher is subtly changed if this flag is set: the return value
257 is the number of characters written to the output buffer (zero is
258 no longer an error code) or a negative error code. Also if the
259 input buffer is NULL and length 0 finalisation should be performed.
262 *) If a candidate issuer certificate is already part of the constructed
263 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
266 *) Improve forward-security support: add functions
268 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
269 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
271 for use by SSL/TLS servers; the callback function will be called whenever a
272 new session is created, and gets to decide whether the session may be
273 cached to make it resumable (return 0) or not (return 1). (As by the
274 SSL/TLS protocol specifications, the session_id sent by the server will be
275 empty to indicate that the session is not resumable; also, the server will
276 not generate RFC 4507 (RFC 5077) session tickets.)
278 A simple reasonable callback implementation is to return is_forward_secure.
279 This parameter will be set to 1 or 0 depending on the ciphersuite selected
280 by the SSL/TLS server library, indicating whether it can provide forward
282 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
284 *) New -verify_name option in command line utilities to set verification
288 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
289 Add CMAC pkey methods.
292 *) Experimental regnegotiation in s_server -www mode. If the client
293 browses /reneg connection is renegotiated. If /renegcert it is
294 renegotiated requesting a certificate.
297 *) Add an "external" session cache for debugging purposes to s_server. This
298 should help trace issues which normally are only apparent in deployed
299 multi-process servers.
302 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
303 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
304 BIO_set_cipher() and some obscure PEM functions were changed so they
305 can now return an error. The RAND changes required a change to the
306 RAND_METHOD structure.
309 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
310 a gcc attribute to warn if the result of a function is ignored. This
311 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
312 whose return value is often ignored.
315 Changes between 1.0.1e and 1.0.2 [xx XXX xxxx]
317 *) Keep original DTLS digest and encryption contexts in retransmission
318 structures so we can use the previous session parameters if they need
319 to be resent. (CVE-2013-6450)
322 *) TLS pad extension: draft-agl-tls-padding-03
324 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
325 TLS client Hello record length value would otherwise be > 255 and
326 less that 512 pad with a dummy extension containing zeroes so it
327 is at least 512 bytes long.
329 [Adam Langley, Steve Henson]
331 *) Add functions to allocate and set the fields of an ECDSA_METHOD
333 [Douglas E. Engert, Steve Henson]
335 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
336 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
337 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
338 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
339 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
340 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
341 [Rob Stradling, Adam Langley]
343 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
344 difference in days and seconds between two tm or ASN1_TIME structures.
347 *) Add -rev test option to s_server to just reverse order of characters
348 received by client and send back to server. Also prints an abbreviated
349 summary of the connection parameters.
352 *) New option -brief for s_client and s_server to print out a brief summary
353 of connection parameters.
356 *) Add callbacks for arbitrary TLS extensions.
357 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
359 *) New option -crl_download in several openssl utilities to download CRLs
360 from CRLDP extension in certificates.
363 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
366 *) New function X509_CRL_diff to generate a delta CRL from the difference
367 of two full CRLs. Add support to "crl" utility.
370 *) New functions to set lookup_crls function and to retrieve
371 X509_STORE from X509_STORE_CTX.
374 *) Print out deprecated issuer and subject unique ID fields in
378 *) Extend OCSP I/O functions so they can be used for simple general purpose
379 HTTP as well as OCSP. New wrapper function which can be used to download
380 CRLs using the OCSP API.
383 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
386 *) SSL_CONF* functions. These provide a common framework for application
387 configuration using configuration files or command lines.
390 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
391 message callback and prints the results. Needs compile time option
392 "enable-ssl-trace". New options to s_client and s_server to enable
396 *) New ctrl and macro to retrieve supported points extensions.
397 Print out extension in s_server and s_client.
400 *) New functions to retrieve certificate signature and signature
404 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
408 *) New Suite B modes for TLS code. These use and enforce the requirements
409 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
410 only use Suite B curves. The Suite B modes can be set by using the
411 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
414 *) New chain verification flags for Suite B levels of security. Check
415 algorithms are acceptable when flags are set in X509_verify_cert.
418 *) Make tls1_check_chain return a set of flags indicating checks passed
419 by a certificate chain. Add additional tests to handle client
420 certificates: checks for matching certificate type and issuer name
424 *) If an attempt is made to use a signature algorithm not in the peer
425 preference list abort the handshake. If client has no suitable
426 signature algorithms in response to a certificate request do not
430 *) If server EC tmp key is not in client preference list abort handshake.
433 *) Add support for certificate stores in CERT structure. This makes it
434 possible to have different stores per SSL structure or one store in
435 the parent SSL_CTX. Include distint stores for certificate chain
436 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
437 to build and store a certificate chain in CERT structure: returing
438 an error if the chain cannot be built: this will allow applications
439 to test if a chain is correctly configured.
441 Note: if the CERT based stores are not set then the parent SSL_CTX
442 store is used to retain compatibility with existing behaviour.
446 *) New function ssl_set_client_disabled to set a ciphersuite disabled
447 mask based on the current session, check mask when sending client
448 hello and checking the requested ciphersuite.
451 *) New ctrls to retrieve and set certificate types in a certificate
452 request message. Print out received values in s_client. If certificate
453 types is not set with custom values set sensible values based on
454 supported signature algorithms.
457 *) Support for distinct client and server supported signature algorithms.
460 *) Add certificate callback. If set this is called whenever a certificate
461 is required by client or server. An application can decide which
462 certificate chain to present based on arbitrary criteria: for example
463 supported signature algorithms. Add very simple example to s_server.
464 This fixes many of the problems and restrictions of the existing client
465 certificate callback: for example you can now clear an existing
466 certificate and specify the whole chain.
469 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
470 the certificate can be used for (if anything). Set valid_flags field
471 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
472 to have similar checks in it.
474 Add new "cert_flags" field to CERT structure and include a "strict mode".
475 This enforces some TLS certificate requirements (such as only permitting
476 certificate signature algorithms contained in the supported algorithms
477 extension) which some implementations ignore: this option should be used
478 with caution as it could cause interoperability issues.
481 *) Update and tidy signature algorithm extension processing. Work out
482 shared signature algorithms based on preferences and peer algorithms
483 and print them out in s_client and s_server. Abort handshake if no
484 shared signature algorithms.
487 *) Add new functions to allow customised supported signature algorithms
488 for SSL and SSL_CTX structures. Add options to s_client and s_server
492 *) New function SSL_certs_clear() to delete all references to certificates
493 from an SSL structure. Before this once a certificate had been added
494 it couldn't be removed.
497 *) Integrate hostname, email address and IP address checking with certificate
498 verification. New verify options supporting checking in opensl utility.
501 *) Fixes and wildcard matching support to hostname and email checking
502 functions. Add manual page.
503 [Florian Weimer (Red Hat Product Security Team)]
505 *) New functions to check a hostname email or IP address against a
506 certificate. Add options x509 utility to print results of checks against
510 *) Fix OCSP checking.
511 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
513 *) Initial experimental support for explicitly trusted non-root CAs.
514 OpenSSL still tries to build a complete chain to a root but if an
515 intermediate CA has a trust setting included that is used. The first
516 setting is used: whether to trust (e.g., -addtrust option to the x509
520 *) Add -trusted_first option which attempts to find certificates in the
521 trusted store even if an untrusted chain is also supplied.
524 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
525 platform support for Linux and Android.
528 *) Support for linux-x32, ILP32 environment in x86_64 framework.
531 *) RFC 5878 (TLS Authorization Extensions) support.
532 [Emilia Kasper, Adam Langley, Ben Laurie (Google)]
534 *) Experimental multi-implementation support for FIPS capable OpenSSL.
535 When in FIPS mode the approved implementations are used as normal,
536 when not in FIPS mode the internal unapproved versions are used instead.
537 This means that the FIPS capable OpenSSL isn't forced to use the
538 (often lower perfomance) FIPS implementations outside FIPS mode.
541 *) Transparently support X9.42 DH parameters when calling
542 PEM_read_bio_DHparameters. This means existing applications can handle
543 the new parameter format automatically.
546 *) Initial experimental support for X9.42 DH parameter format: mainly
547 to support use of 'q' parameter for RFC5114 parameters.
550 *) Add DH parameters from RFC5114 including test data to dhtest.
553 *) Support for automatic EC temporary key parameter selection. If enabled
554 the most preferred EC parameters are automatically used instead of
555 hardcoded fixed parameters. Now a server just has to call:
556 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
557 support ECDH and use the most appropriate parameters.
560 *) Enhance and tidy EC curve and point format TLS extension code. Use
561 static structures instead of allocation if default values are used.
562 New ctrls to set curves we wish to support and to retrieve shared curves.
563 Print out shared curves in s_server. New options to s_server and s_client
564 to set list of supported curves.
567 *) New ctrls to retrieve supported signature algorithms and
568 supported curve values as an array of NIDs. Extend openssl utility
569 to print out received values.
572 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
573 between NIDs and the more common NIST names such as "P-256". Enhance
574 ecparam utility and ECC method to recognise the NIST names for curves.
577 *) Enhance SSL/TLS certificate chain handling to support different
578 chains for each certificate instead of one chain in the parent SSL_CTX.
581 *) Support for fixed DH ciphersuite client authentication: where both
582 server and client use DH certificates with common parameters.
585 *) Support for fixed DH ciphersuites: those requiring DH server
589 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
591 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
592 supporting platforms or when small records were transferred.
593 [Andy Polyakov, Steve Henson]
595 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
597 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
599 This addresses the flaw in CBC record processing discovered by
600 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
601 at: http://www.isg.rhul.ac.uk/tls/
603 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
604 Security Group at Royal Holloway, University of London
605 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
606 Emilia Käsper for the initial patch.
608 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
610 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
611 ciphersuites which can be exploited in a denial of service attack.
612 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
613 and detecting this bug and to Wolfgang Ettlinger
614 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
618 *) Return an error when checking OCSP signatures when key is NULL.
619 This fixes a DoS attack. (CVE-2013-0166)
622 *) Make openssl verify return errors.
623 [Chris Palmer <palmer@google.com> and Ben Laurie]
625 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
626 the right response is stapled. Also change SSL_get_certificate()
627 so it returns the certificate actually sent.
628 See http://rt.openssl.org/Ticket/Display.html?id=2836.
629 [Rob Stradling <rob.stradling@comodo.com>]
631 *) Fix possible deadlock when decoding public keys.
634 *) Don't use TLS 1.0 record version number in initial client hello
638 Changes between 1.0.1b and 1.0.1c [10 May 2012]
640 *) Sanity check record length before skipping explicit IV in TLS
641 1.2, 1.1 and DTLS to fix DoS attack.
643 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
644 fuzzing as a service testing platform.
648 *) Initialise tkeylen properly when encrypting CMS messages.
649 Thanks to Solar Designer of Openwall for reporting this issue.
652 *) In FIPS mode don't try to use composite ciphers as they are not
656 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
658 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
659 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
660 mean any application compiled against OpenSSL 1.0.0 headers setting
661 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
662 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
663 0x10000000L Any application which was previously compiled against
664 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
665 will need to be recompiled as a result. Letting be results in
666 inability to disable specifically TLS 1.1 and in client context,
667 in unlike event, limit maximum offered version to TLS 1.0 [see below].
670 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
671 disable just protocol X, but all protocols above X *if* there are
672 protocols *below* X still enabled. In more practical terms it means
673 that if application wants to disable TLS1.0 in favor of TLS1.1 and
674 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
675 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
679 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
681 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
682 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
683 in CRYPTO_realloc_clean.
685 Thanks to Tavis Ormandy, Google Security Team, for discovering this
686 issue and to Adam Langley <agl@chromium.org> for fixing it.
688 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
690 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
693 *) Workarounds for some broken servers that "hang" if a client hello
694 record length exceeds 255 bytes.
696 1. Do not use record version number > TLS 1.0 in initial client
697 hello: some (but not all) hanging servers will now work.
698 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
699 the number of ciphers sent in the client hello. This should be
700 set to an even number, such as 50, for example by passing:
701 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
702 Most broken servers should now work.
703 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
704 TLS 1.2 client support entirely.
707 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
710 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
712 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
713 STRING form instead of a DigestInfo.
716 *) The format used for MDC2 RSA signatures is inconsistent between EVP
717 and the RSA_sign/RSA_verify functions. This was made more apparent when
718 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
719 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
720 the correct format in RSA_verify so both forms transparently work.
723 *) Some servers which support TLS 1.0 can choke if we initially indicate
724 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
725 encrypted premaster secret. As a workaround use the maximum pemitted
726 client version in client hello, this should keep such servers happy
727 and still work with previous versions of OpenSSL.
730 *) Add support for TLS/DTLS heartbeats.
731 [Robin Seggelmann <seggelmann@fh-muenster.de>]
733 *) Add support for SCTP.
734 [Robin Seggelmann <seggelmann@fh-muenster.de>]
736 *) Improved PRNG seeding for VOS.
737 [Paul Green <Paul.Green@stratus.com>]
739 *) Extensive assembler packs updates, most notably:
741 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
742 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
743 - x86_64: bit-sliced AES implementation;
744 - ARM: NEON support, contemporary platforms optimizations;
745 - s390x: z196 support;
746 - *: GHASH and GF(2^m) multiplication implementations;
750 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
751 (removal of unnecessary code)
752 [Peter Sylvester <peter.sylvester@edelweb.fr>]
754 *) Add TLS key material exporter from RFC 5705.
757 *) Add DTLS-SRTP negotiation from RFC 5764.
760 *) Add Next Protocol Negotiation,
761 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
762 disabled with a no-npn flag to config or Configure. Code donated
764 [Adam Langley <agl@google.com> and Ben Laurie]
766 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
767 NIST-P256, NIST-P521, with constant-time single point multiplication on
768 typical inputs. Compiler support for the nonstandard type __uint128_t is
769 required to use this (present in gcc 4.4 and later, for 64-bit builds).
770 Code made available under Apache License version 2.0.
772 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
773 line to include this in your build of OpenSSL, and run "make depend" (or
774 "make update"). This enables the following EC_METHODs:
776 EC_GFp_nistp224_method()
777 EC_GFp_nistp256_method()
778 EC_GFp_nistp521_method()
780 EC_GROUP_new_by_curve_name() will automatically use these (while
781 EC_GROUP_new_curve_GFp() currently prefers the more flexible
783 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
785 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
786 all platforms. Move ssize_t definition from e_os.h to the public
787 header file e_os2.h as it now appears in public header file cms.h
790 *) New -sigopt option to the ca, req and x509 utilities. Additional
791 signature parameters can be passed using this option and in
795 *) Add RSA PSS signing function. This will generate and set the
796 appropriate AlgorithmIdentifiers for PSS based on those in the
797 corresponding EVP_MD_CTX structure. No application support yet.
800 *) Support for companion algorithm specific ASN1 signing routines.
801 New function ASN1_item_sign_ctx() signs a pre-initialised
802 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
803 the appropriate parameters.
806 *) Add new algorithm specific ASN1 verification initialisation function
807 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
808 handling will be the same no matter what EVP_PKEY_METHOD is used.
809 Add a PSS handler to support verification of PSS signatures: checked
810 against a number of sample certificates.
813 *) Add signature printing for PSS. Add PSS OIDs.
814 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
816 *) Add algorithm specific signature printing. An individual ASN1 method
817 can now print out signatures instead of the standard hex dump.
819 More complex signatures (e.g. PSS) can print out more meaningful
820 information. Include DSA version that prints out the signature
824 *) Password based recipient info support for CMS library: implementing
828 *) Split password based encryption into PBES2 and PBKDF2 functions. This
829 neatly separates the code into cipher and PBE sections and is required
830 for some algorithms that split PBES2 into separate pieces (such as
834 *) Session-handling fixes:
835 - Fix handling of connections that are resuming with a session ID,
836 but also support Session Tickets.
837 - Fix a bug that suppressed issuing of a new ticket if the client
838 presented a ticket with an expired session.
839 - Try to set the ticket lifetime hint to something reasonable.
840 - Make tickets shorter by excluding irrelevant information.
841 - On the client side, don't ignore renewed tickets.
842 [Adam Langley, Bodo Moeller (Google)]
844 *) Fix PSK session representation.
847 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
849 This work was sponsored by Intel.
852 *) Add GCM support to TLS library. Some custom code is needed to split
853 the IV between the fixed (from PRF) and explicit (from TLS record)
854 portions. This adds all GCM ciphersuites supported by RFC5288 and
855 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
856 add a special AESGCM string for GCM only.
859 *) Expand range of ctrls for AES GCM. Permit setting invocation
860 field on decrypt and retrieval of invocation field only on encrypt.
863 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
864 As required by RFC5289 these ciphersuites cannot be used if for
865 versions of TLS earlier than 1.2.
868 *) For FIPS capable OpenSSL interpret a NULL default public key method
869 as unset and return the appopriate default but do *not* set the default.
870 This means we can return the appopriate method in applications that
871 swicth between FIPS and non-FIPS modes.
874 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
875 ENGINE is used then we cannot handle that in the FIPS module so we
876 keep original code iff non-FIPS operations are allowed.
879 *) Add -attime option to openssl utilities.
880 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
882 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
885 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
886 FIPS EC methods unconditionally for now.
889 *) New build option no-ec2m to disable characteristic 2 code.
892 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
893 all cases can be covered as some introduce binary incompatibilities.
896 *) Redirect RSA operations to FIPS module including keygen,
897 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
900 *) Add similar low level API blocking to ciphers.
903 *) Low level digest APIs are not approved in FIPS mode: any attempt
904 to use these will cause a fatal error. Applications that *really* want
905 to use them can use the private_* version instead.
908 *) Redirect cipher operations to FIPS module for FIPS builds.
911 *) Redirect digest operations to FIPS module for FIPS builds.
914 *) Update build system to add "fips" flag which will link in fipscanister.o
915 for static and shared library builds embedding a signature if needed.
918 *) Output TLS supported curves in preference order instead of numerical
919 order. This is currently hardcoded for the highest order curves first.
920 This should be configurable so applications can judge speed vs strength.
923 *) Add TLS v1.2 server support for client authentication.
926 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
930 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
931 FIPS modules versions.
934 *) Add TLS v1.2 client side support for client authentication. Keep cache
935 of handshake records longer as we don't know the hash algorithm to use
936 until after the certificate request message is received.
939 *) Initial TLS v1.2 client support. Add a default signature algorithms
940 extension including all the algorithms we support. Parse new signature
941 format in client key exchange. Relax some ECC signing restrictions for
942 TLS v1.2 as indicated in RFC5246.
945 *) Add server support for TLS v1.2 signature algorithms extension. Switch
946 to new signature format when needed using client digest preference.
947 All server ciphersuites should now work correctly in TLS v1.2. No client
948 support yet and no support for client certificates.
951 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
952 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
953 ciphersuites. At present only RSA key exchange ciphersuites work with
954 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
955 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
956 and version checking.
959 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
960 with this defined it will not be affected by any changes to ssl internal
961 structures. Add several utility functions to allow openssl application
962 to work with OPENSSL_NO_SSL_INTERN defined.
966 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
968 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
971 *) Permit abbreviated handshakes when renegotiating using the function
972 SSL_renegotiate_abbreviated().
973 [Robin Seggelmann <seggelmann@fh-muenster.de>]
975 *) Add call to ENGINE_register_all_complete() to
976 ENGINE_load_builtin_engines(), so some implementations get used
977 automatically instead of needing explicit application support.
980 *) Add support for TLS key exporter as described in RFC5705.
981 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
983 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
984 a few changes are required:
986 Add SSL_OP_NO_TLSv1_1 flag.
988 Update version checking logic to handle version 1.1.
989 Add explicit IV handling (ported from DTLS code).
990 Add command line options to s_client/s_server.
993 Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
995 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
997 This addresses the flaw in CBC record processing discovered by
998 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
999 at: http://www.isg.rhul.ac.uk/tls/
1001 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1002 Security Group at Royal Holloway, University of London
1003 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1004 Emilia Käsper for the initial patch.
1006 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1008 *) Return an error when checking OCSP signatures when key is NULL.
1009 This fixes a DoS attack. (CVE-2013-0166)
1012 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1013 the right response is stapled. Also change SSL_get_certificate()
1014 so it returns the certificate actually sent.
1015 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1016 (This is a backport)
1017 [Rob Stradling <rob.stradling@comodo.com>]
1019 *) Fix possible deadlock when decoding public keys.
1022 Changes between 1.0.0i and 1.0.0j [10 May 2012]
1024 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
1027 *) Sanity check record length before skipping explicit IV in DTLS
1030 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1031 fuzzing as a service testing platform.
1035 *) Initialise tkeylen properly when encrypting CMS messages.
1036 Thanks to Solar Designer of Openwall for reporting this issue.
1039 Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
1041 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1042 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1043 in CRYPTO_realloc_clean.
1045 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1046 issue and to Adam Langley <agl@chromium.org> for fixing it.
1048 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1050 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
1052 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1053 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1054 content decryption and always return the same error. Note: this attack
1055 needs on average 2^20 messages so it only affects automated senders. The
1056 old behaviour can be reenabled in the CMS code by setting the
1057 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1058 an MMA defence is not necessary.
1059 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1060 this issue. (CVE-2012-0884)
1063 *) Fix CVE-2011-4619: make sure we really are receiving a
1064 client hello before rejecting multiple SGC restarts. Thanks to
1065 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1068 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
1070 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1071 Thanks to Antonio Martin, Enterprise Secure Access Research and
1072 Development, Cisco Systems, Inc. for discovering this bug and
1073 preparing a fix. (CVE-2012-0050)
1076 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
1078 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1079 of the Vaudenay padding oracle attack on CBC mode encryption
1080 which enables an efficient plaintext recovery attack against
1081 the OpenSSL implementation of DTLS. Their attack exploits timing
1082 differences arising during decryption processing. A research
1083 paper describing this attack can be found at:
1084 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1085 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1086 Security Group at Royal Holloway, University of London
1087 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1088 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1089 for preparing the fix. (CVE-2011-4108)
1090 [Robin Seggelmann, Michael Tuexen]
1092 *) Clear bytes used for block padding of SSL 3.0 records.
1094 [Adam Langley (Google)]
1096 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1097 Kadianakis <desnacked@gmail.com> for discovering this issue and
1098 Adam Langley for preparing the fix. (CVE-2011-4619)
1099 [Adam Langley (Google)]
1101 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
1102 [Andrey Kulikov <amdeich@gmail.com>]
1104 *) Prevent malformed RFC3779 data triggering an assertion failure.
1105 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1106 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1107 [Rob Austein <sra@hactrn.net>]
1109 *) Improved PRNG seeding for VOS.
1110 [Paul Green <Paul.Green@stratus.com>]
1112 *) Fix ssl_ciph.c set-up race.
1113 [Adam Langley (Google)]
1115 *) Fix spurious failures in ecdsatest.c.
1116 [Emilia Käsper (Google)]
1118 *) Fix the BIO_f_buffer() implementation (which was mixing different
1119 interpretations of the '..._len' fields).
1120 [Adam Langley (Google)]
1122 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1123 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1124 threads won't reuse the same blinding coefficients.
1126 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1127 lock to call BN_BLINDING_invert_ex, and avoids one use of
1128 BN_BLINDING_update for each BN_BLINDING structure (previously,
1129 the last update always remained unused).
1130 [Emilia Käsper (Google)]
1132 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
1133 [Bob Buckholz (Google)]
1135 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
1137 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
1138 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
1139 [Kaspar Brand <ossl@velox.ch>]
1141 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1142 for multi-threaded use of ECDH. (CVE-2011-3210)
1143 [Adam Langley (Google)]
1145 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1148 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
1149 signature public key algorithm by using OID xref utilities instead.
1150 Before this you could only use some ECC ciphersuites with SHA1 only.
1153 *) Add protection against ECDSA timing attacks as mentioned in the paper
1154 by Billy Bob Brumley and Nicola Tuveri, see:
1156 http://eprint.iacr.org/2011/232.pdf
1158 [Billy Bob Brumley and Nicola Tuveri]
1160 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
1162 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1163 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1165 *) Fix bug in string printing code: if *any* escaping is enabled we must
1166 escape the escape character (backslash) or the resulting string is
1170 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
1172 *) Disable code workaround for ancient and obsolete Netscape browsers
1173 and servers: an attacker can use it in a ciphersuite downgrade attack.
1174 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1177 *) Fixed J-PAKE implementation error, originally discovered by
1178 Sebastien Martini, further info and confirmation from Stefan
1179 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1182 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
1184 *) Fix extension code to avoid race conditions which can result in a buffer
1185 overrun vulnerability: resumed sessions must not be modified as they can
1186 be shared by multiple threads. CVE-2010-3864
1189 *) Fix WIN32 build system to correctly link an ENGINE directory into
1193 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
1195 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
1197 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
1199 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
1201 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
1202 context. The operation can be customised via the ctrl mechanism in
1203 case ENGINEs want to include additional functionality.
1206 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
1209 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
1210 output hashes compatible with older versions of OpenSSL.
1211 [Willy Weisz <weisz@vcpc.univie.ac.at>]
1213 *) Fix compression algorithm handling: if resuming a session use the
1214 compression algorithm of the resumed session instead of determining
1215 it from client hello again. Don't allow server to change algorithm.
1218 *) Add load_crls() function to apps tidying load_certs() too. Add option
1219 to verify utility to allow additional CRLs to be included.
1222 *) Update OCSP request code to permit adding custom headers to the request:
1223 some responders need this.
1226 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
1228 [Julia Lawall <julia@diku.dk>]
1230 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
1231 needlessly dereferenced structures, used obsolete functions and
1232 didn't handle all updated verify codes correctly.
1235 *) Disable MD2 in the default configuration.
1238 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
1239 indicate the initial BIO being pushed or popped. This makes it possible
1240 to determine whether the BIO is the one explicitly called or as a result
1241 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
1242 it handles reference counts correctly and doesn't zero out the I/O bio
1243 when it is not being explicitly popped. WARNING: applications which
1244 included workarounds for the old buggy behaviour will need to be modified
1245 or they could free up already freed BIOs.
1248 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
1249 renaming to all platforms (within the 0.9.8 branch, this was
1250 done conditionally on Netware platforms to avoid a name clash).
1251 [Guenter <lists@gknw.net>]
1253 *) Add ECDHE and PSK support to DTLS.
1254 [Michael Tuexen <tuexen@fh-muenster.de>]
1256 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
1260 *) Add "missing" function EVP_MD_flags() (without this the only way to
1261 retrieve a digest flags is by accessing the structure directly. Update
1262 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
1263 or cipher is registered as in the "from" argument. Print out all
1264 registered digests in the dgst usage message instead of manually
1265 attempting to work them out.
1268 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
1269 this allows the use of compression and extensions. Change default cipher
1270 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
1271 by default unless an application cipher string requests it.
1274 *) Alter match criteria in PKCS12_parse(). It used to try to use local
1275 key ids to find matching certificates and keys but some PKCS#12 files
1276 don't follow the (somewhat unwritten) rules and this strategy fails.
1277 Now just gather all certificates together and the first private key
1278 then look for the first certificate that matches the key.
1281 *) Support use of registered digest and cipher names for dgst and cipher
1282 commands instead of having to add each one as a special case. So now
1289 openssl dgst -sha256 foo
1291 and this works for ENGINE based algorithms too.
1295 *) Update Gost ENGINE to support parameter files.
1296 [Victor B. Wagner <vitus@cryptocom.ru>]
1298 *) Support GeneralizedTime in ca utility.
1299 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
1301 *) Enhance the hash format used for certificate directory links. The new
1302 form uses the canonical encoding (meaning equivalent names will work
1303 even if they aren't identical) and uses SHA1 instead of MD5. This form
1304 is incompatible with the older format and as a result c_rehash should
1305 be used to rebuild symbolic links.
1308 *) Make PKCS#8 the default write format for private keys, replacing the
1309 traditional format. This form is standardised, more secure and doesn't
1310 include an implicit MD5 dependency.
1313 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
1314 committed to OpenSSL should pass this lot as a minimum.
1317 *) Add session ticket override functionality for use by EAP-FAST.
1318 [Jouni Malinen <j@w1.fi>]
1320 *) Modify HMAC functions to return a value. Since these can be implemented
1321 in an ENGINE errors can occur.
1324 *) Type-checked OBJ_bsearch_ex.
1327 *) Type-checked OBJ_bsearch. Also some constification necessitated
1328 by type-checking. Still to come: TXT_DB, bsearch(?),
1329 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
1333 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
1334 seconds to a tm structure directly, instead of going through OS
1335 specific date routines. This avoids any issues with OS routines such
1336 as the year 2038 bug. New *_adj() functions for ASN1 time structures
1337 and X509_time_adj_ex() to cover the extended range. The existing
1338 X509_time_adj() is still usable and will no longer have any date issues.
1341 *) Delta CRL support. New use deltas option which will attempt to locate
1342 and search any appropriate delta CRLs available.
1344 This work was sponsored by Google.
1347 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
1348 code and add additional score elements. Validate alternate CRL paths
1349 as part of the CRL checking and indicate a new error "CRL path validation
1350 error" in this case. Applications wanting additional details can use
1351 the verify callback and check the new "parent" field. If this is not
1352 NULL CRL path validation is taking place. Existing applications wont
1353 see this because it requires extended CRL support which is off by
1356 This work was sponsored by Google.
1359 *) Support for freshest CRL extension.
1361 This work was sponsored by Google.
1364 *) Initial indirect CRL support. Currently only supported in the CRLs
1365 passed directly and not via lookup. Process certificate issuer
1366 CRL entry extension and lookup CRL entries by bother issuer name
1367 and serial number. Check and process CRL issuer entry in IDP extension.
1369 This work was sponsored by Google.
1372 *) Add support for distinct certificate and CRL paths. The CRL issuer
1373 certificate is validated separately in this case. Only enabled if
1374 an extended CRL support flag is set: this flag will enable additional
1375 CRL functionality in future.
1377 This work was sponsored by Google.
1380 *) Add support for policy mappings extension.
1382 This work was sponsored by Google.
1385 *) Fixes to pathlength constraint, self issued certificate handling,
1386 policy processing to align with RFC3280 and PKITS tests.
1388 This work was sponsored by Google.
1391 *) Support for name constraints certificate extension. DN, email, DNS
1392 and URI types are currently supported.
1394 This work was sponsored by Google.
1397 *) To cater for systems that provide a pointer-based thread ID rather
1398 than numeric, deprecate the current numeric thread ID mechanism and
1399 replace it with a structure and associated callback type. This
1400 mechanism allows a numeric "hash" to be extracted from a thread ID in
1401 either case, and on platforms where pointers are larger than 'long',
1402 mixing is done to help ensure the numeric 'hash' is usable even if it
1403 can't be guaranteed unique. The default mechanism is to use "&errno"
1404 as a pointer-based thread ID to distinguish between threads.
1406 Applications that want to provide their own thread IDs should now use
1407 CRYPTO_THREADID_set_callback() to register a callback that will call
1408 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
1410 Note that ERR_remove_state() is now deprecated, because it is tied
1411 to the assumption that thread IDs are numeric. ERR_remove_state(0)
1412 to free the current thread's error state should be replaced by
1413 ERR_remove_thread_state(NULL).
1415 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
1416 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
1417 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
1418 application was previously providing a numeric thread callback that
1419 was inappropriate for distinguishing threads, then uniqueness might
1420 have been obtained with &errno that happened immediately in the
1421 intermediate development versions of OpenSSL; this is no longer the
1422 case, the numeric thread callback will now override the automatic use
1424 [Geoff Thorpe, with help from Bodo Moeller]
1426 *) Initial support for different CRL issuing certificates. This covers a
1427 simple case where the self issued certificates in the chain exist and
1428 the real CRL issuer is higher in the existing chain.
1430 This work was sponsored by Google.
1433 *) Removed effectively defunct crypto/store from the build.
1436 *) Revamp of STACK to provide stronger type-checking. Still to come:
1437 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
1438 ASN1_STRING, CONF_VALUE.
1441 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
1442 RAM on SSL connections. This option can save about 34k per idle SSL.
1445 *) Revamp of LHASH to provide stronger type-checking. Still to come:
1446 STACK, TXT_DB, bsearch, qsort.
1449 *) Initial support for Cryptographic Message Syntax (aka CMS) based
1450 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
1451 support for data, signedData, compressedData, digestedData and
1452 encryptedData, envelopedData types included. Scripts to check against
1453 RFC4134 examples draft and interop and consistency checks of many
1454 content types and variants.
1457 *) Add options to enc utility to support use of zlib compression BIO.
1460 *) Extend mk1mf to support importing of options and assembly language
1461 files from Configure script, currently only included in VC-WIN32.
1462 The assembly language rules can now optionally generate the source
1463 files from the associated perl scripts.
1466 *) Implement remaining functionality needed to support GOST ciphersuites.
1467 Interop testing has been performed using CryptoPro implementations.
1468 [Victor B. Wagner <vitus@cryptocom.ru>]
1470 *) s390x assembler pack.
1473 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
1477 *) Implement Opaque PRF Input TLS extension as specified in
1478 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
1479 official specification yet and no extension type assignment by
1480 IANA exists, this extension (for now) will have to be explicitly
1481 enabled when building OpenSSL by providing the extension number
1482 to use. For example, specify an option
1484 -DTLSEXT_TYPE_opaque_prf_input=0x9527
1486 to the "config" or "Configure" script to enable the extension,
1487 assuming extension number 0x9527 (which is a completely arbitrary
1488 and unofficial assignment based on the MD5 hash of the Internet
1489 Draft). Note that by doing so, you potentially lose
1490 interoperability with other TLS implementations since these might
1491 be using the same extension number for other purposes.
1493 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
1494 opaque PRF input value to use in the handshake. This will create
1495 an interal copy of the length-'len' string at 'src', and will
1496 return non-zero for success.
1498 To get more control and flexibility, provide a callback function
1501 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
1502 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
1506 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
1509 Callback function 'cb' will be called in handshakes, and is
1510 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
1511 Argument 'arg' is for application purposes (the value as given to
1512 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
1513 be provided to the callback function). The callback function
1514 has to return non-zero to report success: usually 1 to use opaque
1515 PRF input just if possible, or 2 to enforce use of the opaque PRF
1516 input. In the latter case, the library will abort the handshake
1517 if opaque PRF input is not successfully negotiated.
1519 Arguments 'peerinput' and 'len' given to the callback function
1520 will always be NULL and 0 in the case of a client. A server will
1521 see the client's opaque PRF input through these variables if
1522 available (NULL and 0 otherwise). Note that if the server
1523 provides an opaque PRF input, the length must be the same as the
1524 length of the client's opaque PRF input.
1526 Note that the callback function will only be called when creating
1527 a new session (session resumption can resume whatever was
1528 previously negotiated), and will not be called in SSL 2.0
1529 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
1530 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
1531 for applications that need to enforce opaque PRF input.
1535 *) Update ssl code to support digests other than SHA1+MD5 for handshake
1538 [Victor B. Wagner <vitus@cryptocom.ru>]
1540 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1541 RFC4507bis. The encrypted ticket format is an encrypted encoded
1542 SSL_SESSION structure, that way new session features are automatically
1545 If a client application caches session in an SSL_SESSION structure
1546 support is transparent because tickets are now stored in the encoded
1549 The SSL_CTX structure automatically generates keys for ticket
1550 protection in servers so again support should be possible
1551 with no application modification.
1553 If a client or server wishes to disable RFC4507 support then the option
1554 SSL_OP_NO_TICKET can be set.
1556 Add a TLS extension debugging callback to allow the contents of any client
1557 or server extensions to be examined.
1559 This work was sponsored by Google.
1562 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1563 OpenSSL should now compile cleanly on gcc 4.2
1564 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1566 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1567 support including streaming MAC support: this is required for GOST
1568 ciphersuite support.
1569 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1571 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1572 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1573 to output in BER and PEM format.
1576 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1577 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1578 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1579 ENGINE support for HMAC keys which are unextractable. New -mac and
1580 -macopt options to dgst utility.
1583 *) New option -sigopt to dgst utility. Update dgst to use
1584 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1585 alternative signing paramaters such as X9.31 or PSS in the dgst
1589 *) Change ssl_cipher_apply_rule(), the internal function that does
1590 the work each time a ciphersuite string requests enabling
1591 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1592 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1593 the order of disabled ciphersuites such that those ciphersuites
1594 that most recently went from enabled to disabled not only stay
1595 in order with respect to each other, but also have higher priority
1596 than other disabled ciphersuites the next time ciphersuites are
1599 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1600 the same ciphersuites as with "HIGH" alone, but in a specific
1601 order where the PSK ciphersuites come first (since they are the
1602 most recently disabled ciphersuites when "HIGH" is parsed).
1604 Also, change ssl_create_cipher_list() (using this new
1605 funcionality) such that between otherwise identical
1606 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
1610 *) Change ssl_create_cipher_list() so that it automatically
1611 arranges the ciphersuites in reasonable order before starting
1612 to process the rule string. Thus, the definition for "DEFAULT"
1613 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
1614 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
1615 This makes it much easier to arrive at a reasonable default order
1616 in applications for which anonymous ciphers are OK (meaning
1617 that you can't actually use DEFAULT).
1618 [Bodo Moeller; suggested by Victor Duchovni]
1620 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
1621 processing) into multiple integers instead of setting
1622 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
1623 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
1624 (These masks as well as the individual bit definitions are hidden
1625 away into the non-exported interface ssl/ssl_locl.h, so this
1626 change to the definition of the SSL_CIPHER structure shouldn't
1627 affect applications.) This give us more bits for each of these
1628 categories, so there is no longer a need to coagulate AES128 and
1629 AES256 into a single algorithm bit, and to coagulate Camellia128
1630 and Camellia256 into a single algorithm bit, which has led to all
1633 Thus, among other things, the kludge introduced in 0.9.7m and
1634 0.9.8e for masking out AES256 independently of AES128 or masking
1635 out Camellia256 independently of AES256 is not needed here in 0.9.9.
1637 With the change, we also introduce new ciphersuite aliases that
1638 so far were missing: "AES128", "AES256", "CAMELLIA128", and
1642 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
1643 Use the leftmost N bytes of the signature input if the input is
1644 larger than the prime q (with N being the size in bytes of q).
1647 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1648 it yet and it is largely untested.
1651 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1654 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1655 some compilers (gcc 4.2 and later) reject their use. Safestack is
1656 reimplemented. Update ASN1 to avoid use of legacy functions.
1659 *) Win32/64 targets are linked with Winsock2.
1662 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1663 to external functions. This can be used to increase CRL handling
1664 efficiency especially when CRLs are very large by (for example) storing
1665 the CRL revoked certificates in a database.
1668 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1669 new CRLs added to a directory can be used. New command line option
1670 -verify_return_error to s_client and s_server. This causes real errors
1671 to be returned by the verify callback instead of carrying on no matter
1672 what. This reflects the way a "real world" verify callback would behave.
1675 *) GOST engine, supporting several GOST algorithms and public key formats.
1676 Kindly donated by Cryptocom.
1679 *) Partial support for Issuing Distribution Point CRL extension. CRLs
1680 partitioned by DP are handled but no indirect CRL or reason partitioning
1681 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
1682 selected via a scoring technique which handles IDP and AKID in CRLs.
1685 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1686 will ultimately be used for all verify operations: this will remove the
1687 X509_STORE dependency on certificate verification and allow alternative
1688 lookup methods. X509_STORE based implementations of these two callbacks.
1691 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1692 Modify get_crl() to find a valid (unexpired) CRL if possible.
1695 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1696 this would be called X509_CRL_cmp() but that name is already used by
1697 a function that just compares CRL issuer names. Cache several CRL
1698 extensions in X509_CRL structure and cache CRLDP in X509.
1701 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1702 this maps equivalent X509_NAME structures into a consistent structure.
1703 Name comparison can then be performed rapidly using memcmp().
1706 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1710 *) Allow digests to supply their own micalg string for S/MIME type using
1711 the ctrl EVP_MD_CTRL_MICALG.
1714 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1715 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1716 ctrl. It can then customise the structure before and/or after signing
1720 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1721 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1722 to free up any added signature OIDs.
1725 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1726 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1727 digest and cipher tables. New options added to openssl utility:
1728 list-message-digest-algorithms and list-cipher-algorithms.
1731 *) Change the array representation of binary polynomials: the list
1732 of degrees of non-zero coefficients is now terminated with -1.
1733 Previously it was terminated with 0, which was also part of the
1734 value; thus, the array representation was not applicable to
1735 polynomials where t^0 has coefficient zero. This change makes
1736 the array representation useful in a more general context.
1739 *) Various modifications and fixes to SSL/TLS cipher string
1740 handling. For ECC, the code now distinguishes between fixed ECDH
1741 with RSA certificates on the one hand and with ECDSA certificates
1742 on the other hand, since these are separate ciphersuites. The
1743 unused code for Fortezza ciphersuites has been removed.
1745 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1746 (not "ECDHE"). For consistency with the code for DH
1747 certificates, use of ECDH certificates is now considered ECDH
1748 authentication, not RSA or ECDSA authentication (the latter is
1749 merely the CA's signing algorithm and not actively used in the
1752 The temporary ciphersuite alias "ECCdraft" is no longer
1753 available, and ECC ciphersuites are no longer excluded from "ALL"
1754 and "DEFAULT". The following aliases now exist for RFC 4492
1755 ciphersuites, most of these by analogy with the DH case:
1757 kECDHr - ECDH cert, signed with RSA
1758 kECDHe - ECDH cert, signed with ECDSA
1759 kECDH - ECDH cert (signed with either RSA or ECDSA)
1760 kEECDH - ephemeral ECDH
1761 ECDH - ECDH cert or ephemeral ECDH
1767 AECDH - anonymous ECDH
1768 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1772 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1773 Use correct micalg parameters depending on digest(s) in signed message.
1776 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1777 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1780 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1781 an engine to register a method. Add ENGINE lookups for methods and
1782 functional reference processing.
1785 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1786 EVP_{Sign,Verify}* which allow an application to customise the signature
1790 *) New -resign option to smime utility. This adds one or more signers
1791 to an existing PKCS#7 signedData structure. Also -md option to use an
1792 alternative message digest algorithm for signing.
1795 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1796 create PKCS7 structures containing multiple signers. Update smime
1797 application to support multiple signers.
1800 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1804 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1805 Reorganize PBE internals to lookup from a static table using NIDs,
1806 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1807 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1808 PRF which will be automatically used with PBES2.
1811 *) Replace the algorithm specific calls to generate keys in "req" with the
1815 *) Update PKCS#7 enveloped data routines to use new API. This is now
1816 supported by any public key method supporting the encrypt operation. A
1817 ctrl is added to allow the public key algorithm to examine or modify
1818 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1822 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1823 a default digest type to use. In most cases this will be SHA1 but some
1824 algorithms (such as GOST) need to specify an alternative digest. The
1825 return value indicates how strong the prefernce is 1 means optional and
1826 2 is mandatory (that is it is the only supported type). Modify
1827 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1828 use the default md. Update openssl utilities to use the default digest
1829 type for signing if it is not explicitly indicated.
1832 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1833 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1834 signing method from the key type. This effectively removes the link
1835 between digests and public key types.
1838 *) Add an OID cross reference table and utility functions. Its purpose is to
1839 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1840 rsaEncryption. This will allow some of the algorithm specific hackery
1841 needed to use the correct OID to be removed.
1844 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1845 structures for PKCS7_sign(). They are now set up by the relevant public
1849 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1852 *) Add support for key derivation (agreement) in the API, DH method and
1856 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1857 public and private key formats. As a side effect these add additional
1858 command line functionality not previously available: DSA signatures can be
1859 generated and verified using pkeyutl and DH key support and generation in
1864 [Oliver Tappe <zooey@hirschkaefer.de>]
1866 *) New make target "install_html_docs" installs HTML renditions of the
1868 [Oliver Tappe <zooey@hirschkaefer.de>]
1870 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1871 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1872 support key and parameter generation and add initial key generation
1873 functionality for RSA.
1876 *) Add functions for main EVP_PKEY_method operations. The undocumented
1877 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1878 EVP_PKEY_{encrypt,decrypt}_old.
1881 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1882 key API, doesn't do much yet.
1885 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1886 public key algorithms. New option to openssl utility:
1887 "list-public-key-algorithms" to print out info.
1890 *) Implement the Supported Elliptic Curves Extension for
1891 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1894 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1895 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1898 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1899 utilities such as rsa, dsa, dsaparam etc except they process any key
1903 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1904 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1905 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1909 *) Initial support for pluggable public key ASN1.
1910 De-spaghettify the public key ASN1 handling. Move public and private
1911 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1912 algorithm specific handling to a single module within the relevant
1913 algorithm directory. Add functions to allow (near) opaque processing
1914 of public and private key structures.
1917 *) Implement the Supported Point Formats Extension for
1918 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1921 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1922 for the psk identity [hint] and the psk callback functions to the
1923 SSL_SESSION, SSL and SSL_CTX structure.
1926 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1930 SSL_CTX_use_psk_identity_hint
1931 SSL_get_psk_identity_hint
1932 SSL_get_psk_identity
1933 SSL_use_psk_identity_hint
1935 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1937 *) Add RFC 3161 compliant time stamp request creation, response generation
1938 and response verification functionality.
1939 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1941 *) Add initial support for TLS extensions, specifically for the server_name
1942 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1943 have new members for a host name. The SSL data structure has an
1944 additional member SSL_CTX *initial_ctx so that new sessions can be
1945 stored in that context to allow for session resumption, even after the
1946 SSL has been switched to a new SSL_CTX in reaction to a client's
1947 server_name extension.
1949 New functions (subject to change):
1951 SSL_get_servername()
1952 SSL_get_servername_type()
1955 New CTRL codes and macros (subject to change):
1957 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1958 - SSL_CTX_set_tlsext_servername_callback()
1959 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1960 - SSL_CTX_set_tlsext_servername_arg()
1961 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1963 openssl s_client has a new '-servername ...' option.
1965 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1966 '-key2 ...', '-servername_fatal' (subject to change). This allows
1967 testing the HostName extension for a specific single host name ('-cert'
1968 and '-key' remain fallbacks for handshakes without HostName
1969 negotiation). If the unrecogninzed_name alert has to be sent, this by
1970 default is a warning; it becomes fatal with the '-servername_fatal'
1973 [Peter Sylvester, Remy Allais, Christophe Renou]
1975 *) Whirlpool hash implementation is added.
1978 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1979 bn(64,32). Because of instruction set limitations it doesn't have
1980 any negative impact on performance. This was done mostly in order
1981 to make it possible to share assembler modules, such as bn_mul_mont
1982 implementations, between 32- and 64-bit builds without hassle.
1985 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1986 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1990 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1991 dedicated Montgomery multiplication procedure, is introduced.
1992 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1993 "64-bit" performance on certain 32-bit targets.
1996 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1997 in SSL structures. New SSL ctrl to set maximum send fragment size.
1998 Save memory by seeting the I/O buffer sizes dynamically instead of
1999 using the maximum available value.
2002 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
2003 in addition to the text details.
2006 *) Very, very preliminary EXPERIMENTAL support for printing of general
2007 ASN1 structures. This currently produces rather ugly output and doesn't
2008 handle several customised structures at all.
2011 *) Integrated support for PVK file format and some related formats such
2012 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
2013 these in the 'rsa' and 'dsa' utilities.
2016 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
2019 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
2020 place for the (very old) "NETSCAPE" format certificates which are now
2021 handled using new ASN1 code equivalents.
2024 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
2025 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
2026 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
2029 *) Modify CRL distribution points extension code to print out previously
2030 unsupported fields. Enhance extension setting code to allow setting of
2034 *) Add print and set support for Issuing Distribution Point CRL extension.
2037 *) Change 'Configure' script to enable Camellia by default.
2040 Changes between 0.9.8x and 0.9.8y [5 Feb 2013]
2042 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2044 This addresses the flaw in CBC record processing discovered by
2045 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2046 at: http://www.isg.rhul.ac.uk/tls/
2048 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2049 Security Group at Royal Holloway, University of London
2050 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2051 Emilia Käsper for the initial patch.
2053 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2055 *) Return an error when checking OCSP signatures when key is NULL.
2056 This fixes a DoS attack. (CVE-2013-0166)
2059 *) Return an error when checking OCSP signatures when key is NULL.
2060 This fixes a DoS attack. (CVE-2013-0166)
2063 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2064 the right response is stapled. Also change SSL_get_certificate()
2065 so it returns the certificate actually sent.
2066 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2067 (This is a backport)
2068 [Rob Stradling <rob.stradling@comodo.com>]
2070 *) Fix possible deadlock when decoding public keys.
2073 Changes between 0.9.8w and 0.9.8x [10 May 2012]
2075 *) Sanity check record length before skipping explicit IV in DTLS
2078 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2079 fuzzing as a service testing platform.
2083 *) Initialise tkeylen properly when encrypting CMS messages.
2084 Thanks to Solar Designer of Openwall for reporting this issue.
2087 Changes between 0.9.8v and 0.9.8w [23 Apr 2012]
2089 *) The fix for CVE-2012-2110 did not take into account that the
2090 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an
2091 int in OpenSSL 0.9.8, making it still vulnerable. Fix by
2092 rejecting negative len parameter. (CVE-2012-2131)
2093 [Tomas Hoger <thoger@redhat.com>]
2095 Changes between 0.9.8u and 0.9.8v [19 Apr 2012]
2097 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2098 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2099 in CRYPTO_realloc_clean.
2101 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2102 issue and to Adam Langley <agl@chromium.org> for fixing it.
2104 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2106 Changes between 0.9.8t and 0.9.8u [12 Mar 2012]
2108 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2109 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2110 content decryption and always return the same error. Note: this attack
2111 needs on average 2^20 messages so it only affects automated senders. The
2112 old behaviour can be reenabled in the CMS code by setting the
2113 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2114 an MMA defence is not necessary.
2115 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2116 this issue. (CVE-2012-0884)
2119 *) Fix CVE-2011-4619: make sure we really are receiving a
2120 client hello before rejecting multiple SGC restarts. Thanks to
2121 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2124 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
2126 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2127 Thanks to Antonio Martin, Enterprise Secure Access Research and
2128 Development, Cisco Systems, Inc. for discovering this bug and
2129 preparing a fix. (CVE-2012-0050)
2132 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
2134 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2135 of the Vaudenay padding oracle attack on CBC mode encryption
2136 which enables an efficient plaintext recovery attack against
2137 the OpenSSL implementation of DTLS. Their attack exploits timing
2138 differences arising during decryption processing. A research
2139 paper describing this attack can be found at:
2140 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2141 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2142 Security Group at Royal Holloway, University of London
2143 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2144 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2145 for preparing the fix. (CVE-2011-4108)
2146 [Robin Seggelmann, Michael Tuexen]
2148 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
2149 [Ben Laurie, Kasper <ekasper@google.com>]
2151 *) Clear bytes used for block padding of SSL 3.0 records.
2153 [Adam Langley (Google)]
2155 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2156 Kadianakis <desnacked@gmail.com> for discovering this issue and
2157 Adam Langley for preparing the fix. (CVE-2011-4619)
2158 [Adam Langley (Google)]
2160 *) Prevent malformed RFC3779 data triggering an assertion failure.
2161 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2162 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2163 [Rob Austein <sra@hactrn.net>]
2165 *) Fix ssl_ciph.c set-up race.
2166 [Adam Langley (Google)]
2168 *) Fix spurious failures in ecdsatest.c.
2169 [Emilia Käsper (Google)]
2171 *) Fix the BIO_f_buffer() implementation (which was mixing different
2172 interpretations of the '..._len' fields).
2173 [Adam Langley (Google)]
2175 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2176 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2177 threads won't reuse the same blinding coefficients.
2179 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2180 lock to call BN_BLINDING_invert_ex, and avoids one use of
2181 BN_BLINDING_update for each BN_BLINDING structure (previously,
2182 the last update always remained unused).
2183 [Emilia Käsper (Google)]
2185 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2186 for multi-threaded use of ECDH.
2187 [Adam Langley (Google)]
2189 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2192 *) Add protection against ECDSA timing attacks as mentioned in the paper
2193 by Billy Bob Brumley and Nicola Tuveri, see:
2195 http://eprint.iacr.org/2011/232.pdf
2197 [Billy Bob Brumley and Nicola Tuveri]
2199 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
2201 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2202 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2204 *) Fix bug in string printing code: if *any* escaping is enabled we must
2205 escape the escape character (backslash) or the resulting string is
2209 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
2211 *) Disable code workaround for ancient and obsolete Netscape browsers
2212 and servers: an attacker can use it in a ciphersuite downgrade attack.
2213 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2216 *) Fixed J-PAKE implementation error, originally discovered by
2217 Sebastien Martini, further info and confirmation from Stefan
2218 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2221 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
2223 *) Fix extension code to avoid race conditions which can result in a buffer
2224 overrun vulnerability: resumed sessions must not be modified as they can
2225 be shared by multiple threads. CVE-2010-3864
2228 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
2231 *) Don't reencode certificate when calculating signature: cache and use
2232 the original encoding instead. This makes signature verification of
2233 some broken encodings work correctly.
2236 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
2237 is also one of the inputs.
2238 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
2240 *) Don't repeatedly append PBE algorithms to table if they already exist.
2241 Sort table on each new add. This effectively makes the table read only
2242 after all algorithms are added and subsequent calls to PKCS12_pbe_add
2246 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
2248 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
2251 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
2252 access or freeing data twice (CVE-2010-0742)
2253 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
2255 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
2256 common in certificates and some applications which only call
2257 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
2261 Reduce copying into .apps and .test in makevms.com
2262 Don't try to use blank CA certificate in CA.com
2263 Allow use of C files from original directories in maketests.com
2264 [Steven M. Schweda" <sms@antinode.info>]
2266 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2268 *) When rejecting SSL/TLS records due to an incorrect version number, never
2269 update s->server with a new major version number. As of
2270 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2271 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2272 the previous behavior could result in a read attempt at NULL when
2273 receiving specific incorrect SSL/TLS records once record payload
2274 protection is active. (CVE-2010-0740)
2275 [Bodo Moeller, Adam Langley <agl@chromium.org>]
2277 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
2278 could be crashed if the relevant tables were not present (e.g. chrooted).
2279 [Tomas Hoger <thoger@redhat.com>]
2281 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
2283 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
2284 [Martin Olsson, Neel Mehta]
2286 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
2287 accommodate for stack sorting, always a write lock!).
2290 *) On some versions of WIN32 Heap32Next is very slow. This can cause
2291 excessive delays in the RAND_poll(): over a minute. As a workaround
2292 include a time check in the inner Heap32Next loop too.
2295 *) The code that handled flushing of data in SSL/TLS originally used the
2296 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
2297 the problem outlined in PR#1949. The fix suggested there however can
2298 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
2299 of Apache). So instead simplify the code to flush unconditionally.
2300 This should be fine since flushing with no data to flush is a no op.
2303 *) Handle TLS versions 2.0 and later properly and correctly use the
2304 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
2305 off ancient servers have a habit of sticking around for a while...
2308 *) Modify compression code so it frees up structures without using the
2309 ex_data callbacks. This works around a problem where some applications
2310 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
2311 restarting) then use compression (e.g. SSL with compression) later.
2312 This results in significant per-connection memory leaks and
2313 has caused some security issues including CVE-2008-1678 and
2317 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
2318 change when encrypting or decrypting.
2321 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
2322 connect and renegotiate with servers which do not support RI.
2323 Until RI is more widely deployed this option is enabled by default.
2326 *) Add "missing" ssl ctrls to clear options and mode.
2329 *) If client attempts to renegotiate and doesn't support RI respond with
2330 a no_renegotiation alert as required by RFC5746. Some renegotiating
2331 TLS clients will continue a connection gracefully when they receive
2332 the alert. Unfortunately OpenSSL mishandled this alert and would hang
2333 waiting for a server hello which it will never receive. Now we treat a
2334 received no_renegotiation alert as a fatal error. This is because
2335 applications requesting a renegotiation might well expect it to succeed
2336 and would have no code in place to handle the server denying it so the
2337 only safe thing to do is to terminate the connection.
2340 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
2341 peer supports secure renegotiation and 0 otherwise. Print out peer
2342 renegotiation support in s_client/s_server.
2345 *) Replace the highly broken and deprecated SPKAC certification method with
2346 the updated NID creation version. This should correctly handle UTF8.
2349 *) Implement RFC5746. Re-enable renegotiation but require the extension
2350 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2351 turns out to be a bad idea. It has been replaced by
2352 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2353 SSL_CTX_set_options(). This is really not recommended unless you
2354 know what you are doing.
2355 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2357 *) Fixes to stateless session resumption handling. Use initial_ctx when
2358 issuing and attempting to decrypt tickets in case it has changed during
2359 servername handling. Use a non-zero length session ID when attempting
2360 stateless session resumption: this makes it possible to determine if
2361 a resumption has occurred immediately after receiving server hello
2362 (several places in OpenSSL subtly assume this) instead of later in
2366 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2367 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
2368 fixes for a few places where the return code is not checked
2370 [Julia Lawall <julia@diku.dk>]
2372 *) Add --strict-warnings option to Configure script to include devteam
2373 warnings in other configurations.
2376 *) Add support for --libdir option and LIBDIR variable in makefiles. This
2377 makes it possible to install openssl libraries in locations which
2378 have names other than "lib", for example "/usr/lib64" which some
2380 [Steve Henson, based on patch from Jeremy Utley]
2382 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
2383 X690 8.9.12 and can produce some misleading textual output of OIDs.
2384 [Steve Henson, reported by Dan Kaminsky]
2386 *) Delete MD2 from algorithm tables. This follows the recommendation in
2387 several standards that it is not used in new applications due to
2388 several cryptographic weaknesses. For binary compatibility reasons
2389 the MD2 API is still compiled in by default.
2392 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
2396 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
2397 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
2399 [Guenter <lists@gknw.net>]
2401 *) Fix the server certificate chain building code to use X509_verify_cert(),
2402 it used to have an ad-hoc builder which was unable to cope with anything
2403 other than a simple chain.
2404 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
2406 *) Don't check self signed certificate signatures in X509_verify_cert()
2407 by default (a flag can override this): it just wastes time without
2408 adding any security. As a useful side effect self signed root CAs
2409 with non-FIPS digests are now usable in FIPS mode.
2412 *) In dtls1_process_out_of_seq_message() the check if the current message
2413 is already buffered was missing. For every new message was memory
2414 allocated, allowing an attacker to perform an denial of service attack
2415 with sending out of seq handshake messages until there is no memory
2416 left. Additionally every future messege was buffered, even if the
2417 sequence number made no sense and would be part of another handshake.
2418 So only messages with sequence numbers less than 10 in advance will be
2419 buffered. (CVE-2009-1378)
2420 [Robin Seggelmann, discovered by Daniel Mentz]
2422 *) Records are buffered if they arrive with a future epoch to be
2423 processed after finishing the corresponding handshake. There is
2424 currently no limitation to this buffer allowing an attacker to perform
2425 a DOS attack with sending records with future epochs until there is no
2426 memory left. This patch adds the pqueue_size() function to detemine
2427 the size of a buffer and limits the record buffer to 100 entries.
2429 [Robin Seggelmann, discovered by Daniel Mentz]
2431 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2432 parent structure is freed. (CVE-2009-1379)
2435 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2436 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2439 [Ilya O. <vrghost@gmail.com>]
2441 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2443 *) Disable renegotiation completely - this fixes a severe security
2444 problem (CVE-2009-3555) at the cost of breaking all
2445 renegotiation. Renegotiation can be re-enabled by setting
2446 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2447 run-time. This is really not recommended unless you know what
2451 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2453 *) Don't set val to NULL when freeing up structures, it is freed up by
2454 underlying code. If sizeof(void *) > sizeof(long) this can result in
2455 zeroing past the valid field. (CVE-2009-0789)
2456 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2458 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2459 checked correctly. This would allow some invalid signed attributes to
2460 appear to verify correctly. (CVE-2009-0591)
2461 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2463 *) Reject UniversalString and BMPString types with invalid lengths. This
2464 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2465 a legal length. (CVE-2009-0590)
2468 *) Set S/MIME signing as the default purpose rather than setting it
2469 unconditionally. This allows applications to override it at the store
2473 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2474 to handle some structures.
2477 *) Improve efficiency of mem_gets: don't search whole buffer each time
2479 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2481 *) New -hex option for openssl rand.
2484 *) Print out UTF8String and NumericString when parsing ASN1.
2487 *) Support NumericString type for name components.
2490 *) Allow CC in the environment to override the automatically chosen
2491 compiler. Note that nothing is done to ensure flags work with the
2495 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2497 *) Properly check EVP_VerifyFinal() and similar return values
2499 [Ben Laurie, Bodo Moeller, Google Security Team]
2501 *) Enable TLS extensions by default.
2504 *) Allow the CHIL engine to be loaded, whether the application is
2505 multithreaded or not. (This does not release the developer from the
2506 obligation to set up the dynamic locking callbacks.)
2507 [Sander Temme <sander@temme.net>]
2509 *) Use correct exit code if there is an error in dgst command.
2510 [Steve Henson; problem pointed out by Roland Dirlewanger]
2512 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2513 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2516 *) Add experimental JPAKE support, including demo authentication in
2517 s_client and s_server.
2520 *) Set the comparison function in v3_addr_canonize().
2521 [Rob Austein <sra@hactrn.net>]
2523 *) Add support for XMPP STARTTLS in s_client.
2524 [Philip Paeps <philip@freebsd.org>]
2526 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2527 to ensure that even with this option, only ciphersuites in the
2528 server's preference list will be accepted. (Note that the option
2529 applies only when resuming a session, so the earlier behavior was
2530 just about the algorithm choice for symmetric cryptography.)
2533 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2535 *) Fix NULL pointer dereference if a DTLS server received
2536 ChangeCipherSpec as first record (CVE-2009-1386).
2539 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
2540 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2543 *) The fix in 0.9.8c that supposedly got rid of unsafe
2544 double-checked locking was incomplete for RSA blinding,
2545 addressing just one layer of what turns out to have been
2546 doubly unsafe triple-checked locking.
2548 So now fix this for real by retiring the MONT_HELPER macro
2549 in crypto/rsa/rsa_eay.c.
2551 [Bodo Moeller; problem pointed out by Marius Schilder]
2553 *) Various precautionary measures:
2555 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2557 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2558 (NB: This would require knowledge of the secret session ticket key
2559 to exploit, in which case you'd be SOL either way.)
2561 - Change bn_nist.c so that it will properly handle input BIGNUMs
2562 outside the expected range.
2564 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2567 [Neel Mehta, Bodo Moeller]
2569 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2570 the load fails. Useful for distros.
2571 [Ben Laurie and the FreeBSD team]
2573 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2576 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2579 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2581 This work was sponsored by Logica.
2584 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2585 keystores. Support for SSL/TLS client authentication too.
2586 Not compiled unless enable-capieng specified to Configure.
2588 This work was sponsored by Logica.
2591 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2592 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2593 attribute creation routines such as certifcate requests and PKCS#12
2597 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2599 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2600 handshake which could lead to a cilent crash as found using the
2601 Codenomicon TLS test suite (CVE-2008-1672)
2602 [Steve Henson, Mark Cox]
2604 *) Fix double free in TLS server name extensions which could lead to
2605 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
2608 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
2610 Clear the error queue to ensure that error entries left from
2611 older function calls do not interfere with the correct operation.
2612 [Lutz Jaenicke, Erik de Castro Lopo]
2614 *) Remove root CA certificates of commercial CAs:
2616 The OpenSSL project does not recommend any specific CA and does not
2617 have any policy with respect to including or excluding any CA.
2618 Therefore it does not make any sense to ship an arbitrary selection
2619 of root CA certificates with the OpenSSL software.
2622 *) RSA OAEP patches to fix two separate invalid memory reads.
2623 The first one involves inputs when 'lzero' is greater than
2624 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
2625 before the beginning of from). The second one involves inputs where
2626 the 'db' section contains nothing but zeroes (there is a one-byte
2627 invalid read after the end of 'db').
2628 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2630 *) Partial backport from 0.9.9-dev:
2632 Introduce bn_mul_mont (dedicated Montgomery multiplication
2633 procedure) as a candidate for BIGNUM assembler implementation.
2634 While 0.9.9-dev uses assembler for various architectures, only
2635 x86_64 is available by default here in the 0.9.8 branch, and
2636 32-bit x86 is available through a compile-time setting.
2638 To try the 32-bit x86 assembler implementation, use Configure
2639 option "enable-montasm" (which exists only for this backport).
2641 As "enable-montasm" for 32-bit x86 disclaims code stability
2642 anyway, in this constellation we activate additional code
2643 backported from 0.9.9-dev for further performance improvements,
2644 namely BN_from_montgomery_word. (To enable this otherwise,
2645 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
2647 [Andy Polyakov (backport partially by Bodo Moeller)]
2649 *) Add TLS session ticket callback. This allows an application to set
2650 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
2651 values. This is useful for key rollover for example where several key
2652 sets may exist with different names.
2655 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
2656 This was broken until now in 0.9.8 releases, such that the only way
2657 a registered ENGINE could be used (assuming it initialises
2658 successfully on the host) was to explicitly set it as the default
2659 for the relevant algorithms. This is in contradiction with 0.9.7
2660 behaviour and the documentation. With this fix, when an ENGINE is
2661 registered into a given algorithm's table of implementations, the
2662 'uptodate' flag is reset so that auto-discovery will be used next
2663 time a new context for that algorithm attempts to select an
2665 [Ian Lister (tweaked by Geoff Thorpe)]
2667 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
2668 implemention in the following ways:
2670 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
2673 Lack of BER streaming support means one pass streaming processing is
2674 only supported if data is detached: setting the streaming flag is
2675 ignored for embedded content.
2677 CMS support is disabled by default and must be explicitly enabled
2678 with the enable-cms configuration option.
2681 *) Update the GMP engine glue to do direct copies between BIGNUM and
2682 mpz_t when openssl and GMP use the same limb size. Otherwise the
2683 existing "conversion via a text string export" trick is still used.
2684 [Paul Sheer <paulsheer@gmail.com>]
2686 *) Zlib compression BIO. This is a filter BIO which compressed and
2687 uncompresses any data passed through it.
2690 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
2691 RFC3394 compatible AES key wrapping.
2694 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
2695 sets string data without copying. X509_ALGOR_set0() and
2696 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
2697 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
2698 from an X509_ATTRIBUTE structure optionally checking it occurs only
2699 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
2703 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
2704 to get the expected BN_FLG_CONSTTIME behavior.
2705 [Bodo Moeller (Google)]
2709 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
2710 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
2711 - added some more tests to do_tests.pl
2712 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
2713 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
2714 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
2715 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
2716 - various changes to netware.pl to enable gcc-cross builds on Win32
2718 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
2719 - various changes to fix missing prototype warnings
2720 - fixed x86nasm.pl to create correct asm files for NASM COFF output
2721 - added AES, WHIRLPOOL and CPUID assembler code to build files
2722 - added missing AES assembler make rules to mk1mf.pl
2723 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
2724 [Guenter Knauf <eflash@gmx.net>]
2726 *) Implement certificate status request TLS extension defined in RFC3546.
2727 A client can set the appropriate parameters and receive the encoded
2728 OCSP response via a callback. A server can query the supplied parameters
2729 and set the encoded OCSP response in the callback. Add simplified examples
2730 to s_client and s_server.
2733 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
2735 *) Fix various bugs:
2736 + Binary incompatibility of ssl_ctx_st structure
2737 + DTLS interoperation with non-compliant servers
2738 + Don't call get_session_cb() without proposed session
2739 + Fix ia64 assembler code
2740 [Andy Polyakov, Steve Henson]
2742 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
2744 *) DTLS Handshake overhaul. There were longstanding issues with
2745 OpenSSL DTLS implementation, which were making it impossible for
2746 RFC 4347 compliant client to communicate with OpenSSL server.
2747 Unfortunately just fixing these incompatibilities would "cut off"
2748 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
2749 server keeps tolerating non RFC compliant syntax. The opposite is
2750 not true, 0.9.8f client can not communicate with earlier server.
2751 This update even addresses CVE-2007-4995.
2754 *) Changes to avoid need for function casts in OpenSSL: some compilers
2755 (gcc 4.2 and later) reject their use.
2756 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
2759 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2760 RFC4507bis. The encrypted ticket format is an encrypted encoded
2761 SSL_SESSION structure, that way new session features are automatically
2764 If a client application caches session in an SSL_SESSION structure
2765 support is transparent because tickets are now stored in the encoded
2768 The SSL_CTX structure automatically generates keys for ticket
2769 protection in servers so again support should be possible
2770 with no application modification.
2772 If a client or server wishes to disable RFC4507 support then the option
2773 SSL_OP_NO_TICKET can be set.
2775 Add a TLS extension debugging callback to allow the contents of any client
2776 or server extensions to be examined.
2778 This work was sponsored by Google.
2781 *) Add initial support for TLS extensions, specifically for the server_name
2782 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2783 have new members for a host name. The SSL data structure has an
2784 additional member SSL_CTX *initial_ctx so that new sessions can be
2785 stored in that context to allow for session resumption, even after the
2786 SSL has been switched to a new SSL_CTX in reaction to a client's
2787 server_name extension.
2789 New functions (subject to change):
2791 SSL_get_servername()
2792 SSL_get_servername_type()
2795 New CTRL codes and macros (subject to change):
2797 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2798 - SSL_CTX_set_tlsext_servername_callback()
2799 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2800 - SSL_CTX_set_tlsext_servername_arg()
2801 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2803 openssl s_client has a new '-servername ...' option.
2805 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2806 '-key2 ...', '-servername_fatal' (subject to change). This allows
2807 testing the HostName extension for a specific single host name ('-cert'
2808 and '-key' remain fallbacks for handshakes without HostName
2809 negotiation). If the unrecogninzed_name alert has to be sent, this by
2810 default is a warning; it becomes fatal with the '-servername_fatal'
2813 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2815 *) Add AES and SSE2 assembly language support to VC++ build.
2818 *) Mitigate attack on final subtraction in Montgomery reduction.
2821 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2822 (which previously caused an internal error).
2825 *) Squeeze another 10% out of IGE mode when in != out.
2828 *) AES IGE mode speedup.
2829 [Dean Gaudet (Google)]
2831 *) Add the Korean symmetric 128-bit cipher SEED (see
2832 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2833 add SEED ciphersuites from RFC 4162:
2835 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2836 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2837 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2838 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2840 To minimize changes between patchlevels in the OpenSSL 0.9.8
2841 series, SEED remains excluded from compilation unless OpenSSL
2842 is configured with 'enable-seed'.
2843 [KISA, Bodo Moeller]
2845 *) Mitigate branch prediction attacks, which can be practical if a
2846 single processor is shared, allowing a spy process to extract
2847 information. For detailed background information, see
2848 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2849 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2850 and Necessary Software Countermeasures"). The core of the change
2851 are new versions BN_div_no_branch() and
2852 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2853 respectively, which are slower, but avoid the security-relevant
2854 conditional branches. These are automatically called by BN_div()
2855 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2856 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2857 remove a conditional branch.
2859 BN_FLG_CONSTTIME is the new name for the previous
2860 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2861 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2862 in the exponent causes BN_mod_exp_mont() to use the alternative
2863 implementation in BN_mod_exp_mont_consttime().) The old name
2864 remains as a deprecated alias.
2866 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2867 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2868 constant-time implementations for more than just exponentiation.
2869 Here too the old name is kept as a deprecated alias.
2871 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2872 the BN_BLINDING structure gets an independent copy of the
2873 modulus. This means that the previous "BIGNUM *m" argument to
2874 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2875 essentially becomes "const BIGNUM *m", although we can't actually
2876 change this in the header file before 0.9.9. It allows
2877 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2878 enable BN_FLG_CONSTTIME.
2880 [Matthew D Wood (Intel Corp)]
2882 *) In the SSL/TLS server implementation, be strict about session ID
2883 context matching (which matters if an application uses a single
2884 external cache for different purposes). Previously,
2885 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2886 set. This did ensure strict client verification, but meant that,
2887 with applications using a single external cache for quite
2888 different requirements, clients could circumvent ciphersuite
2889 restrictions for a given session ID context by starting a session
2890 in a different context.
2893 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2894 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2895 authentication-only ciphersuites.
2898 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2899 not complete and could lead to a possible single byte overflow
2900 (CVE-2007-5135) [Ben Laurie]
2902 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2904 *) Since AES128 and AES256 (and similarly Camellia128 and
2905 Camellia256) share a single mask bit in the logic of
2906 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2907 kludge to work properly if AES128 is available and AES256 isn't
2908 (or if Camellia128 is available and Camellia256 isn't).
2911 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2912 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2913 When a point or a seed is encoded in a BIT STRING, we need to
2914 prevent the removal of trailing zero bits to get the proper DER
2915 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2916 of a NamedBitList, for which trailing 0 bits need to be removed.)
2919 *) Have SSL/TLS server implementation tolerate "mismatched" record
2920 protocol version while receiving ClientHello even if the
2921 ClientHello is fragmented. (The server can't insist on the
2922 particular protocol version it has chosen before the ServerHello
2923 message has informed the client about his choice.)
2926 *) Add RFC 3779 support.
2927 [Rob Austein for ARIN, Ben Laurie]
2929 *) Load error codes if they are not already present instead of using a
2930 static variable. This allows them to be cleanly unloaded and reloaded.
2931 Improve header file function name parsing.
2934 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2935 or CAPABILITY handshake as required by RFCs.
2938 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2940 *) Introduce limits to prevent malicious keys being able to
2941 cause a denial of service. (CVE-2006-2940)
2942 [Steve Henson, Bodo Moeller]
2944 *) Fix ASN.1 parsing of certain invalid structures that can result
2945 in a denial of service. (CVE-2006-2937) [Steve Henson]
2947 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2948 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2950 *) Fix SSL client code which could crash if connecting to a
2951 malicious SSLv2 server. (CVE-2006-4343)
2952 [Tavis Ormandy and Will Drewry, Google Security Team]
2954 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2955 match only those. Before that, "AES256-SHA" would be interpreted
2956 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2957 the same strength classification in 0.9.7h) as we currently only
2958 have a single AES bit in the ciphersuite description bitmap.
2959 That change, however, also applied to ciphersuite strings such as
2960 "RC4-MD5" that intentionally matched multiple ciphersuites --
2961 namely, SSL 2.0 ciphersuites in addition to the more common ones
2962 from SSL 3.0/TLS 1.0.
2964 So we change the selection algorithm again: Naming an explicit
2965 ciphersuite selects this one ciphersuite, and any other similar
2966 ciphersuite (same bitmap) from *other* protocol versions.
2967 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2968 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2970 Since SSL 2.0 does not have any ciphersuites for which the
2971 128/256 bit distinction would be relevant, this works for now.
2972 The proper fix will be to use different bits for AES128 and
2973 AES256, which would have avoided the problems from the beginning;
2974 however, bits are scarce, so we can only do this in a new release
2975 (not just a patchlevel) when we can change the SSL_CIPHER
2976 definition to split the single 'unsigned long mask' bitmap into
2977 multiple values to extend the available space.
2981 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2983 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2984 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2986 *) Add AES IGE and biIGE modes.
2989 *) Change the Unix randomness entropy gathering to use poll() when
2990 possible instead of select(), since the latter has some
2991 undesirable limitations.
2992 [Darryl Miles via Richard Levitte and Bodo Moeller]
2994 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2995 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2996 cannot be implicitly activated as part of, e.g., the "AES" alias.
2997 However, please upgrade to OpenSSL 0.9.9[-dev] for
2998 non-experimental use of the ECC ciphersuites to get TLS extension
2999 support, which is required for curve and point format negotiation
3000 to avoid potential handshake problems.
3003 *) Disable rogue ciphersuites:
3005 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3006 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3007 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3009 The latter two were purportedly from
3010 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3013 Also deactivate the remaining ciphersuites from
3014 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3015 unofficial, and the ID has long expired.
3018 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3019 dual-core machines) and other potential thread-safety issues.
3022 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3023 versions), which is now available for royalty-free use
3024 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3025 Also, add Camellia TLS ciphersuites from RFC 4132.
3027 To minimize changes between patchlevels in the OpenSSL 0.9.8
3028 series, Camellia remains excluded from compilation unless OpenSSL
3029 is configured with 'enable-camellia'.
3032 *) Disable the padding bug check when compression is in use. The padding
3033 bug check assumes the first packet is of even length, this is not
3034 necessarily true if compresssion is enabled and can result in false
3035 positives causing handshake failure. The actual bug test is ancient
3036 code so it is hoped that implementations will either have fixed it by
3037 now or any which still have the bug do not support compression.
3040 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3042 *) When applying a cipher rule check to see if string match is an explicit
3043 cipher suite and only match that one cipher suite if it is.
3046 *) Link in manifests for VC++ if needed.
3047 [Austin Ziegler <halostatue@gmail.com>]
3049 *) Update support for ECC-based TLS ciphersuites according to
3050 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3051 TLS extensions, which are supported starting with the 0.9.9
3052 branch, not in the OpenSSL 0.9.8 branch).
3055 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3056 opaque EVP_CIPHER_CTX handling.
3059 *) Fixes and enhancements to zlib compression code. We now only use
3060 "zlib1.dll" and use the default __cdecl calling convention on Win32
3061 to conform with the standards mentioned here:
3062 http://www.zlib.net/DLL_FAQ.txt
3063 Static zlib linking now works on Windows and the new --with-zlib-include
3064 --with-zlib-lib options to Configure can be used to supply the location
3065 of the headers and library. Gracefully handle case where zlib library
3069 *) Several fixes and enhancements to the OID generation code. The old code
3070 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3071 handle numbers larger than ULONG_MAX, truncated printing and had a
3072 non standard OBJ_obj2txt() behaviour.
3075 *) Add support for building of engines under engine/ as shared libraries
3076 under VC++ build system.
3079 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3080 Hopefully, we will not see any false combination of paths any more.
3083 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3085 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3086 (part of SSL_OP_ALL). This option used to disable the
3087 countermeasure against man-in-the-middle protocol-version
3088 rollback in the SSL 2.0 server implementation, which is a bad
3089 idea. (CVE-2005-2969)
3091 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3092 for Information Security, National Institute of Advanced Industrial
3093 Science and Technology [AIST], Japan)]
3095 *) Add two function to clear and return the verify parameter flags.
3098 *) Keep cipherlists sorted in the source instead of sorting them at
3099 runtime, thus removing the need for a lock.
3102 *) Avoid some small subgroup attacks in Diffie-Hellman.
3103 [Nick Mathewson and Ben Laurie]
3105 *) Add functions for well-known primes.
3108 *) Extended Windows CE support.
3109 [Satoshi Nakamura and Andy Polyakov]
3111 *) Initialize SSL_METHOD structures at compile time instead of during
3112 runtime, thus removing the need for a lock.
3115 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3116 attempting to decrypt each encrypted key in turn. Add support to
3120 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3122 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3125 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3128 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3129 key into the same file any more.
3132 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3135 *) Add -utf8 command line and config file option to 'ca'.
3136 [Stefan <stf@udoma.org]
3138 *) Removed the macro des_crypt(), as it seems to conflict with some
3139 libraries. Use DES_crypt().
3142 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3143 involves renaming the source and generated shared-libs for
3144 both. The engines will accept the corrected or legacy ids
3145 ('ncipher' and '4758_cca' respectively) when binding. NB,
3146 this only applies when building 'shared'.
3147 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3149 *) Add attribute functions to EVP_PKEY structure. Modify
3150 PKCS12_create() to recognize a CSP name attribute and
3151 use it. Make -CSP option work again in pkcs12 utility.
3154 *) Add new functionality to the bn blinding code:
3155 - automatic re-creation of the BN_BLINDING parameters after
3156 a fixed number of uses (currently 32)
3157 - add new function for parameter creation
3158 - introduce flags to control the update behaviour of the
3159 BN_BLINDING parameters
3160 - hide BN_BLINDING structure
3161 Add a second BN_BLINDING slot to the RSA structure to improve
3162 performance when a single RSA object is shared among several
3166 *) Add support for DTLS.
3167 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3169 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3170 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3173 *) Remove buggy and incompletet DH cert support from
3174 ssl/ssl_rsa.c and ssl/s3_both.c
3177 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3178 the apps/openssl applications.
3181 *) Compile clean with "-Wall -Wmissing-prototypes
3182 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
3183 DEBUG_SAFESTACK must also be set.
3186 *) Change ./Configure so that certain algorithms can be disabled by default.
3187 The new counterpiece to "no-xxx" is "enable-xxx".
3189 The patented RC5 and MDC2 algorithms will now be disabled unless
3190 "enable-rc5" and "enable-mdc2", respectively, are specified.
3192 (IDEA remains enabled despite being patented. This is because IDEA
3193 is frequently required for interoperability, and there is no license
3194 fee for non-commercial use. As before, "no-idea" can be used to
3195 avoid this algorithm.)
3199 *) Add processing of proxy certificates (see RFC 3820). This work was
3200 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
3201 EGEE (Enabling Grids for E-science in Europe).
3204 *) RC4 performance overhaul on modern architectures/implementations, such
3205 as Intel P4, IA-64 and AMD64.
3208 *) New utility extract-section.pl. This can be used specify an alternative
3209 section number in a pod file instead of having to treat each file as
3210 a separate case in Makefile. This can be done by adding two lines to the
3213 =for comment openssl_section:XXX
3215 The blank line is mandatory.
3219 *) New arguments -certform, -keyform and -pass for s_client and s_server
3220 to allow alternative format key and certificate files and passphrase
3224 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
3225 update associated structures and add various utility functions.
3227 Add new policy related verify parameters, include policy checking in
3228 standard verify code. Enhance 'smime' application with extra parameters
3229 to support policy checking and print out.
3232 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
3233 Nehemiah processors. These extensions support AES encryption in hardware
3234 as well as RNG (though RNG support is currently disabled).
3235 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
3237 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
3240 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
3241 [Andy Polyakov and a number of other people]
3243 *) Improved PowerPC platform support. Most notably BIGNUM assembler
3244 implementation contributed by IBM.
3245 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
3247 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
3248 exponent rather than 'unsigned long'. There is a corresponding change to
3249 the new 'rsa_keygen' element of the RSA_METHOD structure.
3250 [Jelte Jansen, Geoff Thorpe]
3252 *) Functionality for creating the initial serial number file is now
3253 moved from CA.pl to the 'ca' utility with a new option -create_serial.
3255 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
3256 number file to 1, which is bound to cause problems. To avoid
3257 the problems while respecting compatibility between different 0.9.7
3258 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
3259 CA.pl for serial number initialization. With the new release 0.9.8,
3260 we can fix the problem directly in the 'ca' utility.)
3263 *) Reduced header interdepencies by declaring more opaque objects in
3264 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
3265 give fewer recursive includes, which could break lazy source code - so
3266 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
3267 developers should define this symbol when building and using openssl to
3268 ensure they track the recommended behaviour, interfaces, [etc], but
3269 backwards-compatible behaviour prevails when this isn't defined.
3272 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
3275 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
3276 This will generate a random key of the appropriate length based on the
3277 cipher context. The EVP_CIPHER can provide its own random key generation
3278 routine to support keys of a specific form. This is used in the des and
3279 3des routines to generate a key of the correct parity. Update S/MIME
3280 code to use new functions and hence generate correct parity DES keys.
3281 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
3282 valid (weak or incorrect parity).
3285 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
3286 as looking them up. This is useful when the verified structure may contain
3287 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
3288 present unless the new PKCS7_NO_CRL flag is asserted.
3291 *) Extend ASN1 oid configuration module. It now additionally accepts the
3294 shortName = some long name, 1.2.3.4
3297 *) Reimplemented the BN_CTX implementation. There is now no more static
3298 limitation on the number of variables it can handle nor the depth of the
3299 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
3300 information can now expand as required, and rather than having a single
3301 static array of bignums, BN_CTX now uses a linked-list of such arrays
3302 allowing it to expand on demand whilst maintaining the usefulness of
3303 BN_CTX's "bundling".
3306 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
3307 to allow all RSA operations to function using a single BN_CTX.
3310 *) Preliminary support for certificate policy evaluation and checking. This
3311 is initially intended to pass the tests outlined in "Conformance Testing
3312 of Relying Party Client Certificate Path Processing Logic" v1.07.
3315 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
3316 remained unused and not that useful. A variety of other little bignum
3317 tweaks and fixes have also been made continuing on from the audit (see
3321 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
3322 associated ASN1, EVP and SSL functions and old ASN1 macros.
3325 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
3326 and this should never fail. So the return value from the use of
3327 BN_set_word() (which can fail due to needless expansion) is now deprecated;
3328 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
3331 *) BN_CTX_get() should return zero-valued bignums, providing the same
3332 initialised value as BN_new().
3333 [Geoff Thorpe, suggested by Ulf Möller]
3335 *) Support for inhibitAnyPolicy certificate extension.
3338 *) An audit of the BIGNUM code is underway, for which debugging code is
3339 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
3340 is considered valid when processing BIGNUMs, and causes execution to
3341 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
3342 further steps are taken to deliberately pollute unused data in BIGNUM
3343 structures to try and expose faulty code further on. For now, openssl will
3344 (in its default mode of operation) continue to tolerate the inconsistent
3345 forms that it has tolerated in the past, but authors and packagers should
3346 consider trying openssl and their own applications when compiled with
3347 these debugging symbols defined. It will help highlight potential bugs in
3348 their own code, and will improve the test coverage for OpenSSL itself. At
3349 some point, these tighter rules will become openssl's default to improve
3350 maintainability, though the assert()s and other overheads will remain only
3351 in debugging configurations. See bn.h for more details.
3352 [Geoff Thorpe, Nils Larsch, Ulf Möller]
3354 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
3355 that can only be obtained through BN_CTX_new() (which implicitly
3356 initialises it). The presence of this function only made it possible
3357 to overwrite an existing structure (and cause memory leaks).
3360 *) Because of the callback-based approach for implementing LHASH as a
3361 template type, lh_insert() adds opaque objects to hash-tables and
3362 lh_doall() or lh_doall_arg() are typically used with a destructor callback
3363 to clean up those corresponding objects before destroying the hash table
3364 (and losing the object pointers). So some over-zealous constifications in
3365 LHASH have been relaxed so that lh_insert() does not take (nor store) the
3366 objects as "const" and the lh_doall[_arg] callback wrappers are not
3367 prototyped to have "const" restrictions on the object pointers they are
3368 given (and so aren't required to cast them away any more).
3371 *) The tmdiff.h API was so ugly and minimal that our own timing utility
3372 (speed) prefers to use its own implementation. The two implementations
3373 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
3374 its object type properly exposed (MS_TM) instead of casting to/from "char
3375 *". This may still change yet if someone realises MS_TM and "ms_time_***"
3376 aren't necessarily the greatest nomenclatures - but this is what was used
3377 internally to the implementation so I've used that for now.
3380 *) Ensure that deprecated functions do not get compiled when
3381 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
3382 the self-tests were still using deprecated key-generation functions so
3383 these have been updated also.
3386 *) Reorganise PKCS#7 code to separate the digest location functionality
3387 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
3388 New function PKCS7_set_digest() to set the digest type for PKCS#7
3389 digestedData type. Add additional code to correctly generate the
3390 digestedData type and add support for this type in PKCS7 initialization
3394 *) New function PKCS7_set0_type_other() this initializes a PKCS7
3395 structure of type "other".
3398 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
3399 sure the loop does correctly stop and breaking ("division by zero")
3400 modulus operations are not performed. The (pre-generated) prime
3401 table crypto/bn/bn_prime.h was already correct, but it could not be
3402 re-generated on some platforms because of the "division by zero"
3403 situation in the script.
3404 [Ralf S. Engelschall]
3406 *) Update support for ECC-based TLS ciphersuites according to
3407 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
3408 SHA-1 now is only used for "small" curves (where the
3409 representation of a field element takes up to 24 bytes); for
3410 larger curves, the field element resulting from ECDH is directly
3411 used as premaster secret.
3412 [Douglas Stebila (Sun Microsystems Laboratories)]
3414 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
3415 curve secp160r1 to the tests.
3416 [Douglas Stebila (Sun Microsystems Laboratories)]
3418 *) Add the possibility to load symbols globally with DSO.
3419 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
3421 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
3422 control of the error stack.
3425 *) Add support for STORE in ENGINE.
3428 *) Add the STORE type. The intention is to provide a common interface
3429 to certificate and key stores, be they simple file-based stores, or
3430 HSM-type store, or LDAP stores, or...
3431 NOTE: The code is currently UNTESTED and isn't really used anywhere.
3434 *) Add a generic structure called OPENSSL_ITEM. This can be used to
3435 pass a list of arguments to any function as well as provide a way
3436 for a function to pass data back to the caller.
3439 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
3440 works like BUF_strdup() but can be used to duplicate a portion of
3441 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
3445 *) Add the function sk_find_ex() which works like sk_find(), but will
3446 return an index to an element even if an exact match couldn't be
3447 found. The index is guaranteed to point at the element where the
3448 searched-for key would be inserted to preserve sorting order.
3451 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
3452 takes an extra flags argument for optional functionality. Currently,
3453 the following flags are defined:
3455 OBJ_BSEARCH_VALUE_ON_NOMATCH
3456 This one gets OBJ_bsearch_ex() to return a pointer to the first
3457 element where the comparing function returns a negative or zero
3460 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
3461 This one gets OBJ_bsearch_ex() to return a pointer to the first
3462 element where the comparing function returns zero. This is useful
3463 if there are more than one element where the comparing function
3467 *) Make it possible to create self-signed certificates with 'openssl ca'
3468 in such a way that the self-signed certificate becomes part of the
3469 CA database and uses the same mechanisms for serial number generation
3470 as all other certificate signing. The new flag '-selfsign' enables
3471 this functionality. Adapt CA.sh and CA.pl.in.
3474 *) Add functionality to check the public key of a certificate request
3475 against a given private. This is useful to check that a certificate
3476 request can be signed by that key (self-signing).
3479 *) Make it possible to have multiple active certificates with the same
3480 subject in the CA index file. This is done only if the keyword
3481 'unique_subject' is set to 'no' in the main CA section (default
3482 if 'CA_default') of the configuration file. The value is saved
3483 with the database itself in a separate index attribute file,
3484 named like the index file with '.attr' appended to the name.
3487 *) Generate muti valued AVAs using '+' notation in config files for
3491 *) Support for nameConstraints certificate extension.
3494 *) Support for policyConstraints certificate extension.
3497 *) Support for policyMappings certificate extension.
3500 *) Make sure the default DSA_METHOD implementation only uses its
3501 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3502 and change its own handlers to be NULL so as to remove unnecessary
3503 indirection. This lets alternative implementations fallback to the
3504 default implementation more easily.
3507 *) Support for directoryName in GeneralName related extensions
3511 *) Make it possible to link applications using Makefile.shared.
3512 Make that possible even when linking against static libraries!
3515 *) Support for single pass processing for S/MIME signing. This now
3516 means that S/MIME signing can be done from a pipe, in addition
3517 cleartext signing (multipart/signed type) is effectively streaming
3518 and the signed data does not need to be all held in memory.
3520 This is done with a new flag PKCS7_STREAM. When this flag is set
3521 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3522 is done after the data is output (and digests calculated) in
3523 SMIME_write_PKCS7().
3526 *) Add full support for -rpath/-R, both in shared libraries and
3527 applications, at least on the platforms where it's known how
3531 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3532 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3533 will now compute a table of multiples of the generator that
3534 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3535 faster (notably in the case of a single point multiplication,
3536 scalar * generator).
3537 [Nils Larsch, Bodo Moeller]
3539 *) IPv6 support for certificate extensions. The various extensions
3540 which use the IP:a.b.c.d can now take IPv6 addresses using the
3541 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3545 *) Added an ENGINE that implements RSA by performing private key
3546 exponentiations with the GMP library. The conversions to and from
3547 GMP's mpz_t format aren't optimised nor are any montgomery forms
3548 cached, and on x86 it appears OpenSSL's own performance has caught up.
3549 However there are likely to be other architectures where GMP could
3550 provide a boost. This ENGINE is not built in by default, but it can be
3551 specified at Configure time and should be accompanied by the necessary
3552 linker additions, eg;
3553 ./config -DOPENSSL_USE_GMP -lgmp
3556 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3557 testing availability of engines with "-t" - the old behaviour is
3558 produced by increasing the feature's verbosity with "-tt".
3561 *) ECDSA routines: under certain error conditions uninitialized BN objects
3562 could be freed. Solution: make sure initialization is performed early
3563 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3567 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3568 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3569 software implementations. For DSA and DH, parameter generation can
3570 also be overriden by providing the appropriate method callbacks.
3573 *) Change the "progress" mechanism used in key-generation and
3574 primality testing to functions that take a new BN_GENCB pointer in
3575 place of callback/argument pairs. The new API functions have "_ex"
3576 postfixes and the older functions are reimplemented as wrappers for
3577 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3578 declarations of the old functions to help (graceful) attempts to
3579 migrate to the new functions. Also, the new key-generation API
3580 functions operate on a caller-supplied key-structure and return
3581 success/failure rather than returning a key or NULL - this is to
3582 help make "keygen" another member function of RSA_METHOD etc.
3584 Example for using the new callback interface:
3586 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3590 BN_GENCB_set(&my_cb, my_callback, my_arg);
3592 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3593 /* For the meaning of a, b in calls to my_callback(), see the
3594 * documentation of the function that calls the callback.
3595 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3596 * my_callback should return 1 if it wants BN_is_prime_ex()
3597 * to continue, or 0 to stop.
3602 *) Change the ZLIB compression method to be stateful, and make it
3603 available to TLS with the number defined in
3604 draft-ietf-tls-compression-04.txt.
3607 *) Add the ASN.1 structures and functions for CertificatePair, which
3608 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
3610 CertificatePair ::= SEQUENCE {
3611 forward [0] Certificate OPTIONAL,
3612 reverse [1] Certificate OPTIONAL,
3613 -- at least one of the pair shall be present -- }
3615 Also implement the PEM functions to read and write certificate
3616 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
3618 This needed to be defined, mostly for the sake of the LDAP
3619 attribute crossCertificatePair, but may prove useful elsewhere as
3623 *) Make it possible to inhibit symlinking of shared libraries in
3624 Makefile.shared, for Cygwin's sake.
3627 *) Extend the BIGNUM API by creating a function
3628 void BN_set_negative(BIGNUM *a, int neg);
3629 and a macro that behave like
3630 int BN_is_negative(const BIGNUM *a);
3632 to avoid the need to access 'a->neg' directly in applications.
3635 *) Implement fast modular reduction for pseudo-Mersenne primes
3636 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
3637 EC_GROUP_new_curve_GFp() will now automatically use this
3639 [Nils Larsch <nla@trustcenter.de>]
3641 *) Add new lock type (CRYPTO_LOCK_BN).
3644 *) Change the ENGINE framework to automatically load engines
3645 dynamically from specific directories unless they could be
3646 found to already be built in or loaded. Move all the
3647 current engines except for the cryptodev one to a new
3649 The engines in engines/ are built as shared libraries if
3650 the "shared" options was given to ./Configure or ./config.
3651 Otherwise, they are inserted in libcrypto.a.
3652 /usr/local/ssl/engines is the default directory for dynamic
3653 engines, but that can be overriden at configure time through
3654 the usual use of --prefix and/or --openssldir, and at run
3655 time with the environment variable OPENSSL_ENGINES.
3656 [Geoff Thorpe and Richard Levitte]
3658 *) Add Makefile.shared, a helper makefile to build shared
3659 libraries. Addapt Makefile.org.
3662 *) Add version info to Win32 DLLs.
3663 [Peter 'Luna' Runestig" <peter@runestig.com>]
3665 *) Add new 'medium level' PKCS#12 API. Certificates and keys
3666 can be added using this API to created arbitrary PKCS#12
3667 files while avoiding the low level API.
3669 New options to PKCS12_create(), key or cert can be NULL and
3670 will then be omitted from the output file. The encryption
3671 algorithm NIDs can be set to -1 for no encryption, the mac
3672 iteration count can be set to 0 to omit the mac.
3674 Enhance pkcs12 utility by making the -nokeys and -nocerts
3675 options work when creating a PKCS#12 file. New option -nomac
3676 to omit the mac, NONE can be set for an encryption algorithm.
3677 New code is modified to use the enhanced PKCS12_create()
3678 instead of the low level API.
3681 *) Extend ASN1 encoder to support indefinite length constructed
3682 encoding. This can output sequences tags and octet strings in
3683 this form. Modify pk7_asn1.c to support indefinite length
3684 encoding. This is experimental and needs additional code to
3685 be useful, such as an ASN1 bio and some enhanced streaming
3688 Extend template encode functionality so that tagging is passed
3689 down to the template encoder.
3692 *) Let 'openssl req' fail if an argument to '-newkey' is not
3693 recognized instead of using RSA as a default.
3696 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
3697 As these are not official, they are not included in "ALL";
3698 the "ECCdraft" ciphersuite group alias can be used to select them.
3699 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
3701 *) Add ECDH engine support.
3702 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
3704 *) Add ECDH in new directory crypto/ecdh/.
3705 [Douglas Stebila (Sun Microsystems Laboratories)]
3707 *) Let BN_rand_range() abort with an error after 100 iterations
3708 without success (which indicates a broken PRNG).
3711 *) Change BN_mod_sqrt() so that it verifies that the input value
3712 is really the square of the return value. (Previously,
3713 BN_mod_sqrt would show GIGO behaviour.)
3716 *) Add named elliptic curves over binary fields from X9.62, SECG,
3717 and WAP/WTLS; add OIDs that were still missing.
3719 [Sheueling Chang Shantz and Douglas Stebila
3720 (Sun Microsystems Laboratories)]
3722 *) Extend the EC library for elliptic curves over binary fields
3723 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
3726 EC_GF2m_simple_method
3730 EC_GROUP_new_curve_GF2m
3731 EC_GROUP_set_curve_GF2m
3732 EC_GROUP_get_curve_GF2m
3733 EC_POINT_set_affine_coordinates_GF2m
3734 EC_POINT_get_affine_coordinates_GF2m
3735 EC_POINT_set_compressed_coordinates_GF2m
3737 Point compression for binary fields is disabled by default for
3738 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
3741 As binary polynomials are represented as BIGNUMs, various members
3742 of the EC_GROUP and EC_POINT data structures can be shared
3743 between the implementations for prime fields and binary fields;
3744 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
3745 are essentially identical to their ..._GFp counterparts.
3746 (For simplicity, the '..._GFp' prefix has been dropped from
3747 various internal method names.)
3749 An internal 'field_div' method (similar to 'field_mul' and
3750 'field_sqr') has been added; this is used only for binary fields.
3752 [Sheueling Chang Shantz and Douglas Stebila
3753 (Sun Microsystems Laboratories)]
3755 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
3756 through methods ('mul', 'precompute_mult').
3758 The generic implementations (now internally called 'ec_wNAF_mul'
3759 and 'ec_wNAF_precomputed_mult') remain the default if these
3760 methods are undefined.
3762 [Sheueling Chang Shantz and Douglas Stebila
3763 (Sun Microsystems Laboratories)]
3765 *) New function EC_GROUP_get_degree, which is defined through
3766 EC_METHOD. For curves over prime fields, this returns the bit
3767 length of the modulus.
3769 [Sheueling Chang Shantz and Douglas Stebila
3770 (Sun Microsystems Laboratories)]
3772 *) New functions EC_GROUP_dup, EC_POINT_dup.
3773 (These simply call ..._new and ..._copy).
3775 [Sheueling Chang Shantz and Douglas Stebila
3776 (Sun Microsystems Laboratories)]
3778 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3779 Polynomials are represented as BIGNUMs (where the sign bit is not
3780 used) in the following functions [macros]:
3783 BN_GF2m_sub [= BN_GF2m_add]
3784 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3785 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3786 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3788 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3789 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3790 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3791 BN_GF2m_cmp [= BN_ucmp]
3793 (Note that only the 'mod' functions are actually for fields GF(2^m).
3794 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3796 For some functions, an the irreducible polynomial defining a
3797 field can be given as an 'unsigned int[]' with strictly
3798 decreasing elements giving the indices of those bits that are set;
3799 i.e., p[] represents the polynomial
3800 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3802 p[0] > p[1] > ... > p[k] = 0.
3803 This applies to the following functions:
3808 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3809 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3811 BN_GF2m_mod_sqrt_arr
3812 BN_GF2m_mod_solve_quad_arr
3816 Conversion can be performed by the following functions:
3821 bntest.c has additional tests for binary polynomial arithmetic.
3823 Two implementations for BN_GF2m_mod_div() are available.
3824 The default algorithm simply uses BN_GF2m_mod_inv() and
3825 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3826 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3827 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3829 [Sheueling Chang Shantz and Douglas Stebila
3830 (Sun Microsystems Laboratories)]
3832 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3833 functionality is disabled at compile-time.
3834 [Douglas Stebila <douglas.stebila@sun.com>]
3836 *) Change default behaviour of 'openssl asn1parse' so that more
3837 information is visible when viewing, e.g., a certificate:
3839 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3840 mode the content of non-printable OCTET STRINGs is output in a
3841 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3842 avoid the appearance of a printable string.
3843 [Nils Larsch <nla@trustcenter.de>]
3845 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3847 EC_GROUP_set_asn1_flag()
3848 EC_GROUP_get_asn1_flag()
3849 EC_GROUP_set_point_conversion_form()
3850 EC_GROUP_get_point_conversion_form()
3851 These control ASN1 encoding details:
3852 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3853 has been set to OPENSSL_EC_NAMED_CURVE.
3854 - Points are encoded in uncompressed form by default; options for
3855 asn1_for are as for point2oct, namely
3856 POINT_CONVERSION_COMPRESSED
3857 POINT_CONVERSION_UNCOMPRESSED
3858 POINT_CONVERSION_HYBRID
3860 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3863 EC_GROUP_get0_seed()
3864 EC_GROUP_get_seed_len()
3865 This is used only for ASN1 purposes (so far).
3866 [Nils Larsch <nla@trustcenter.de>]
3868 *) Add 'field_type' member to EC_METHOD, which holds the NID
3869 of the appropriate field type OID. The new function
3870 EC_METHOD_get_field_type() returns this value.
3871 [Nils Larsch <nla@trustcenter.de>]
3876 EC_POINT_point2hex()
3877 EC_POINT_hex2point()
3878 providing useful interfaces to EC_POINT_point2oct() and
3879 EC_POINT_oct2point().
3880 [Nils Larsch <nla@trustcenter.de>]
3882 *) Change internals of the EC library so that the functions
3883 EC_GROUP_set_generator()
3884 EC_GROUP_get_generator()
3885 EC_GROUP_get_order()
3886 EC_GROUP_get_cofactor()
3887 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3888 to methods, which would lead to unnecessary code duplication when
3889 adding different types of curves.
3890 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3892 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3893 arithmetic, and such that modified wNAFs are generated
3894 (which avoid length expansion in many cases).
3897 *) Add a function EC_GROUP_check_discriminant() (defined via
3898 EC_METHOD) that verifies that the curve discriminant is non-zero.
3900 Add a function EC_GROUP_check() that makes some sanity tests
3901 on a EC_GROUP, its generator and order. This includes
3902 EC_GROUP_check_discriminant().
3903 [Nils Larsch <nla@trustcenter.de>]
3905 *) Add ECDSA in new directory crypto/ecdsa/.
3907 Add applications 'openssl ecparam' and 'openssl ecdsa'
3908 (these are based on 'openssl dsaparam' and 'openssl dsa').
3910 ECDSA support is also included in various other files across the
3911 library. Most notably,
3912 - 'openssl req' now has a '-newkey ecdsa:file' option;
3913 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3914 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3915 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3916 them suitable for ECDSA where domain parameters must be
3917 extracted before the specific public key;
3918 - ECDSA engine support has been added.
3919 [Nils Larsch <nla@trustcenter.de>]
3921 *) Include some named elliptic curves, and add OIDs from X9.62,
3922 SECG, and WAP/WTLS. Each curve can be obtained from the new
3924 EC_GROUP_new_by_curve_name(),
3925 and the list of available named curves can be obtained with
3926 EC_get_builtin_curves().
3927 Also add a 'curve_name' member to EC_GROUP objects, which can be
3929 EC_GROUP_set_curve_name()
3930 EC_GROUP_get_curve_name()
3931 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3933 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3934 was actually never needed) and in BN_mul(). The removal in BN_mul()
3935 required a small change in bn_mul_part_recursive() and the addition
3936 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3937 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3938 bn_sub_words() and bn_add_words() except they take arrays with
3942 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3944 *) Cleanse PEM buffers before freeing them since they may contain
3946 [Benjamin Bennett <ben@psc.edu>]
3948 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3949 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3950 authentication-only ciphersuites.
3953 *) Since AES128 and AES256 share a single mask bit in the logic of
3954 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3955 kludge to work properly if AES128 is available and AES256 isn't.
3958 *) Expand security boundary to match 1.1.1 module.
3961 *) Remove redundant features: hash file source, editing of test vectors
3962 modify fipsld to use external fips_premain.c signature.
3965 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3966 run algorithm test programs.
3969 *) Make algorithm test programs more tolerant of whitespace.
3972 *) Have SSL/TLS server implementation tolerate "mismatched" record
3973 protocol version while receiving ClientHello even if the
3974 ClientHello is fragmented. (The server can't insist on the
3975 particular protocol version it has chosen before the ServerHello
3976 message has informed the client about his choice.)
3979 *) Load error codes if they are not already present instead of using a
3980 static variable. This allows them to be cleanly unloaded and reloaded.
3983 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3985 *) Introduce limits to prevent malicious keys being able to
3986 cause a denial of service. (CVE-2006-2940)
3987 [Steve Henson, Bodo Moeller]
3989 *) Fix ASN.1 parsing of certain invalid structures that can result
3990 in a denial of service. (CVE-2006-2937) [Steve Henson]
3992 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3993 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3995 *) Fix SSL client code which could crash if connecting to a
3996 malicious SSLv2 server. (CVE-2006-4343)
3997 [Tavis Ormandy and Will Drewry, Google Security Team]
3999 *) Change ciphersuite string processing so that an explicit
4000 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
4001 will no longer include "AES128-SHA"), and any other similar
4002 ciphersuite (same bitmap) from *other* protocol versions (so that
4003 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
4004 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
4005 changes from 0.9.8b and 0.9.8d.
4008 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4010 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4011 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4013 *) Change the Unix randomness entropy gathering to use poll() when
4014 possible instead of select(), since the latter has some
4015 undesirable limitations.
4016 [Darryl Miles via Richard Levitte and Bodo Moeller]
4018 *) Disable rogue ciphersuites:
4020 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4021 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4022 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4024 The latter two were purportedly from
4025 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4028 Also deactive the remaining ciphersuites from
4029 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4030 unofficial, and the ID has long expired.
4033 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4034 dual-core machines) and other potential thread-safety issues.
4037 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4039 *) Adapt fipsld and the build system to link against the validated FIPS
4040 module in FIPS mode.
4043 *) Fixes for VC++ 2005 build under Windows.
4046 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4047 from a Windows bash shell such as MSYS. It is autodetected from the
4048 "config" script when run from a VC++ environment. Modify standard VC++
4049 build to use fipscanister.o from the GNU make build.
4052 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4054 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4055 The value now differs depending on if you build for FIPS or not.
4056 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4057 safely run with a non-FIPSed libcrypto, as it may crash because of
4058 the difference induced by this change.
4061 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4063 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4064 (part of SSL_OP_ALL). This option used to disable the
4065 countermeasure against man-in-the-middle protocol-version
4066 rollback in the SSL 2.0 server implementation, which is a bad
4067 idea. (CVE-2005-2969)
4069 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4070 for Information Security, National Institute of Advanced Industrial
4071 Science and Technology [AIST], Japan)]
4073 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4074 mainly for FIPS compliance and not fully integrated at this stage.
4077 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4078 the exponentiation using a fixed-length exponent. (Otherwise,
4079 the information leaked through timing could expose the secret key
4080 after many signatures; cf. Bleichenbacher's attack on DSA with
4084 *) Make a new fixed-window mod_exp implementation the default for
4085 RSA, DSA, and DH private-key operations so that the sequence of
4086 squares and multiplies and the memory access pattern are
4087 independent of the particular secret key. This will mitigate
4088 cache-timing and potential related attacks.
4090 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4091 and this is automatically used by BN_mod_exp_mont() if the new flag
4092 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4093 will use this BN flag for private exponents unless the flag
4094 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4095 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4097 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4099 *) Change the client implementation for SSLv23_method() and
4100 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4101 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4102 (Previously, the SSL 2.0 backwards compatible Client Hello
4103 message format would be used even with SSL_OP_NO_SSLv2.)
4106 *) Add support for smime-type MIME parameter in S/MIME messages which some
4110 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4111 a threadsafe manner. Modify rsa code to use new function and add calls
4112 to dsa and dh code (which had race conditions before).
4115 *) Include the fixed error library code in the C error file definitions
4116 instead of fixing them up at runtime. This keeps the error code
4117 structures constant.
4120 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4122 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4125 *) Fixes for newer kerberos headers. NB: the casts are needed because
4126 the 'length' field is signed on one version and unsigned on another
4127 with no (?) obvious way to tell the difference, without these VC++
4128 complains. Also the "definition" of FAR (blank) is no longer included
4129 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4130 some needed definitions.
4133 *) Undo Cygwin change.
4136 *) Added support for proxy certificates according to RFC 3820.
4137 Because they may be a security thread to unaware applications,
4138 they must be explicitely allowed in run-time. See
4139 docs/HOWTO/proxy_certificates.txt for further information.
4142 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4144 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4145 server and client random values. Previously
4146 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4147 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4149 This change has negligible security impact because:
4151 1. Server and client random values still have 24 bytes of pseudo random
4154 2. Server and client random values are sent in the clear in the initial
4157 3. The master secret is derived using the premaster secret (48 bytes in
4158 size for static RSA ciphersuites) as well as client server and random
4161 The OpenSSL team would like to thank the UK NISCC for bringing this issue
4164 [Stephen Henson, reported by UK NISCC]
4166 *) Use Windows randomness collection on Cygwin.
4169 *) Fix hang in EGD/PRNGD query when communication socket is closed
4170 prematurely by EGD/PRNGD.
4171 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
4173 *) Prompt for pass phrases when appropriate for PKCS12 input format.
4176 *) Back-port of selected performance improvements from development
4177 branch, as well as improved support for PowerPC platforms.
4180 *) Add lots of checks for memory allocation failure, error codes to indicate
4181 failure and freeing up memory if a failure occurs.
4182 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
4184 *) Add new -passin argument to dgst.
4187 *) Perform some character comparisons of different types in X509_NAME_cmp:
4188 this is needed for some certificates that reencode DNs into UTF8Strings
4189 (in violation of RFC3280) and can't or wont issue name rollover
4193 *) Make an explicit check during certificate validation to see that
4194 the CA setting in each certificate on the chain is correct. As a
4195 side effect always do the following basic checks on extensions,
4196 not just when there's an associated purpose to the check:
4198 - if there is an unhandled critical extension (unless the user
4199 has chosen to ignore this fault)
4200 - if the path length has been exceeded (if one is set at all)
4201 - that certain extensions fit the associated purpose (if one has
4205 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
4207 *) Avoid a race condition when CRLs are checked in a multi threaded
4208 environment. This would happen due to the reordering of the revoked
4209 entries during signature checking and serial number lookup. Now the
4210 encoding is cached and the serial number sort performed under a lock.
4211 Add new STACK function sk_is_sorted().
4214 *) Add Delta CRL to the extension code.
4217 *) Various fixes to s3_pkt.c so alerts are sent properly.
4218 [David Holmes <d.holmes@f5.com>]
4220 *) Reduce the chances of duplicate issuer name and serial numbers (in
4221 violation of RFC3280) using the OpenSSL certificate creation utilities.
4222 This is done by creating a random 64 bit value for the initial serial
4223 number when a serial number file is created or when a self signed
4224 certificate is created using 'openssl req -x509'. The initial serial
4225 number file is created using 'openssl x509 -next_serial' in CA.pl
4226 rather than being initialized to 1.
4229 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
4231 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4232 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4233 [Joe Orton, Steve Henson]
4235 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
4237 [Joe Orton, Steve Henson]
4239 *) Make it possible to have multiple active certificates with the same
4240 subject in the CA index file. This is done only if the keyword
4241 'unique_subject' is set to 'no' in the main CA section (default
4242 if 'CA_default') of the configuration file. The value is saved
4243 with the database itself in a separate index attribute file,
4244 named like the index file with '.attr' appended to the name.
4247 *) X509 verify fixes. Disable broken certificate workarounds when
4248 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
4249 keyUsage extension present. Don't accept CRLs with unhandled critical
4250 extensions: since verify currently doesn't process CRL extensions this
4251 rejects a CRL with *any* critical extensions. Add new verify error codes
4255 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
4256 A clarification of RFC2560 will require the use of OCTET STRINGs and
4257 some implementations cannot handle the current raw format. Since OpenSSL
4258 copies and compares OCSP nonces as opaque blobs without any attempt at
4259 parsing them this should not create any compatibility issues.
4262 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
4263 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
4264 this HMAC (and other) operations are several times slower than OpenSSL
4268 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
4269 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4271 *) Use the correct content when signing type "other".
4274 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
4276 *) Fix various bugs revealed by running the NISCC test suite:
4278 Stop out of bounds reads in the ASN1 code when presented with
4279 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4281 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
4283 If verify callback ignores invalid public key errors don't try to check
4284 certificate signature with the NULL public key.
4288 *) New -ignore_err option in ocsp application to stop the server
4289 exiting on the first error in a request.
4292 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4293 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4297 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4298 extra data after the compression methods not only for TLS 1.0
4299 but also for SSL 3.0 (as required by the specification).
4300 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4302 *) Change X509_certificate_type() to mark the key as exported/exportable
4303 when it's 512 *bits* long, not 512 bytes.
4306 *) Change AES_cbc_encrypt() so it outputs exact multiple of
4307 blocks during encryption.
4310 *) Various fixes to base64 BIO and non blocking I/O. On write
4311 flushes were not handled properly if the BIO retried. On read
4312 data was not being buffered properly and had various logic bugs.
4313 This also affects blocking I/O when the data being decoded is a
4317 *) Various S/MIME bugfixes and compatibility changes:
4318 output correct application/pkcs7 MIME type if
4319 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
4320 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
4321 of files as .eml work). Correctly handle very long lines in MIME
4325 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
4327 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4328 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4329 a protocol version number mismatch like a decryption error
4330 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4333 *) Turn on RSA blinding by default in the default implementation
4334 to avoid a timing attack. Applications that don't want it can call
4335 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4336 They would be ill-advised to do so in most cases.
4337 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4339 *) Change RSA blinding code so that it works when the PRNG is not
4340 seeded (in this case, the secret RSA exponent is abused as
4341 an unpredictable seed -- if it is not unpredictable, there
4342 is no point in blinding anyway). Make RSA blinding thread-safe
4343 by remembering the creator's thread ID in rsa->blinding and
4344 having all other threads use local one-time blinding factors
4345 (this requires more computation than sharing rsa->blinding, but
4346 avoids excessive locking; and if an RSA object is not shared
4347 between threads, blinding will still be very fast).
4350 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
4351 ENGINE as defaults for all supported algorithms irrespective of
4352 the 'flags' parameter. 'flags' is now honoured, so applications
4353 should make sure they are passing it correctly.
4356 *) Target "mingw" now allows native Windows code to be generated in
4357 the Cygwin environment as well as with the MinGW compiler.
4360 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
4362 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4363 via timing by performing a MAC computation even if incorrrect
4364 block cipher padding has been found. This is a countermeasure
4365 against active attacks where the attacker has to distinguish
4366 between bad padding and a MAC verification error. (CVE-2003-0078)
4368 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4369 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4370 Martin Vuagnoux (EPFL, Ilion)]
4372 *) Make the no-err option work as intended. The intention with no-err
4373 is not to have the whole error stack handling routines removed from
4374 libcrypto, it's only intended to remove all the function name and
4375 reason texts, thereby removing some of the footprint that may not
4376 be interesting if those errors aren't displayed anyway.
4378 NOTE: it's still possible for any application or module to have it's
4379 own set of error texts inserted. The routines are there, just not
4380 used by default when no-err is given.
4383 *) Add support for FreeBSD on IA64.
4384 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
4386 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
4387 Kerberos function mit_des_cbc_cksum(). Before this change,
4388 the value returned by DES_cbc_cksum() was like the one from
4389 mit_des_cbc_cksum(), except the bytes were swapped.
4390 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
4392 *) Allow an application to disable the automatic SSL chain building.
4393 Before this a rather primitive chain build was always performed in
4394 ssl3_output_cert_chain(): an application had no way to send the
4395 correct chain if the automatic operation produced an incorrect result.
4397 Now the chain builder is disabled if either:
4399 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
4401 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
4403 The reasoning behind this is that an application would not want the
4404 auto chain building to take place if extra chain certificates are
4405 present and it might also want a means of sending no additional
4406 certificates (for example the chain has two certificates and the
4410 *) Add the possibility to build without the ENGINE framework.
4411 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4413 *) Under Win32 gmtime() can return NULL: check return value in
4414 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
4417 *) DSA routines: under certain error conditions uninitialized BN objects
4418 could be freed. Solution: make sure initialization is performed early
4419 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
4420 Nils Larsch <nla@trustcenter.de> via PR#459)
4423 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
4424 checked on reconnect on the client side, therefore session resumption
4425 could still fail with a "ssl session id is different" error. This
4426 behaviour is masked when SSL_OP_ALL is used due to
4427 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
4428 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4429 followup to PR #377.
4432 *) IA-32 assembler support enhancements: unified ELF targets, support
4433 for SCO/Caldera platforms, fix for Cygwin shared build.
4436 *) Add support for FreeBSD on sparc64. As a consequence, support for
4437 FreeBSD on non-x86 processors is separate from x86 processors on
4438 the config script, much like the NetBSD support.
4439 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
4441 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
4443 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
4446 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
4447 code (06) was taken as the first octet of the session ID and the last
4448 octet was ignored consequently. As a result SSLv2 client side session
4449 caching could not have worked due to the session ID mismatch between
4451 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4455 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
4456 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
4460 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
4461 seems that in spite of existing for more than a year, many application
4462 author have done nothing to provide the necessary callbacks, which
4463 means that this particular engine will not work properly anywhere.
4464 This is a very unfortunate situation which forces us, in the name
4465 of usability, to give the hw_ncipher.c a static lock, which is part
4467 NOTE: This is for the 0.9.7 series ONLY. This hack will never
4468 appear in 0.9.8 or later. We EXPECT application authors to have
4469 dealt properly with this when 0.9.8 is released (unless we actually
4470 make such changes in the libcrypto locking code that changes will
4471 have to be made anyway).
4474 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
4475 octets have been read, EOF or an error occurs. Without this change
4476 some truncated ASN1 structures will not produce an error.
4479 *) Disable Heimdal support, since it hasn't been fully implemented.
4480 Still give the possibility to force the use of Heimdal, but with
4481 warnings and a request that patches get sent to openssl-dev.
4484 *) Add the VC-CE target, introduce the WINCE sysname, and add
4485 INSTALL.WCE and appropriate conditionals to make it build.
4486 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4488 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4489 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4490 edit numbers of the version.
4491 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4493 *) Introduce safe string copy and catenation functions
4494 (BUF_strlcpy() and BUF_strlcat()).
4495 [Ben Laurie (CHATS) and Richard Levitte]
4497 *) Avoid using fixed-size buffers for one-line DNs.
4498 [Ben Laurie (CHATS)]
4500 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4501 resizing buffers containing secrets, and use where appropriate.
4502 [Ben Laurie (CHATS)]
4504 *) Avoid using fixed size buffers for configuration file location.
4505 [Ben Laurie (CHATS)]
4507 *) Avoid filename truncation for various CA files.
4508 [Ben Laurie (CHATS)]
4510 *) Use sizeof in preference to magic numbers.
4511 [Ben Laurie (CHATS)]
4513 *) Avoid filename truncation in cert requests.
4514 [Ben Laurie (CHATS)]
4516 *) Add assertions to check for (supposedly impossible) buffer
4518 [Ben Laurie (CHATS)]
4520 *) Don't cache truncated DNS entries in the local cache (this could
4521 potentially lead to a spoofing attack).
4522 [Ben Laurie (CHATS)]
4524 *) Fix various buffers to be large enough for hex/decimal
4525 representations in a platform independent manner.
4526 [Ben Laurie (CHATS)]
4528 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4529 resizing buffers containing secrets, and use where appropriate.
4530 [Ben Laurie (CHATS)]
4532 *) Add BIO_indent() to avoid much slightly worrying code to do
4534 [Ben Laurie (CHATS)]
4536 *) Convert sprintf()/BIO_puts() to BIO_printf().
4537 [Ben Laurie (CHATS)]
4539 *) buffer_gets() could terminate with the buffer only half
4541 [Ben Laurie (CHATS)]
4543 *) Add assertions to prevent user-supplied crypto functions from
4544 overflowing internal buffers by having large block sizes, etc.
4545 [Ben Laurie (CHATS)]
4547 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4549 [Ben Laurie (CHATS)]
4551 *) Eliminate unused copy of key in RC4.
4552 [Ben Laurie (CHATS)]
4554 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4555 [Ben Laurie (CHATS)]
4557 *) Fix off-by-one error in EGD path.
4558 [Ben Laurie (CHATS)]
4560 *) If RANDFILE path is too long, ignore instead of truncating.
4561 [Ben Laurie (CHATS)]
4563 *) Eliminate unused and incorrectly sized X.509 structure
4565 [Ben Laurie (CHATS)]
4567 *) Eliminate unused and dangerous function knumber().
4568 [Ben Laurie (CHATS)]
4570 *) Eliminate unused and dangerous structure, KSSL_ERR.
4571 [Ben Laurie (CHATS)]
4573 *) Protect against overlong session ID context length in an encoded
4574 session object. Since these are local, this does not appear to be
4576 [Ben Laurie (CHATS)]
4578 *) Change from security patch (see 0.9.6e below) that did not affect
4579 the 0.9.6 release series:
4581 Remote buffer overflow in SSL3 protocol - an attacker could
4582 supply an oversized master key in Kerberos-enabled versions.
4584 [Ben Laurie (CHATS)]
4586 *) Change the SSL kerb5 codes to match RFC 2712.
4589 *) Make -nameopt work fully for req and add -reqopt switch.
4590 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4592 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4593 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4595 *) Make sure tests can be performed even if the corresponding algorithms
4596 have been removed entirely. This was also the last step to make
4597 OpenSSL compilable with DJGPP under all reasonable conditions.
4598 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
4600 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
4601 to allow version independent disabling of normally unselected ciphers,
4602 which may be activated as a side-effect of selecting a single cipher.
4604 (E.g., cipher list string "RSA" enables ciphersuites that are left
4605 out of "ALL" because they do not provide symmetric encryption.
4606 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
4607 [Lutz Jaenicke, Bodo Moeller]
4609 *) Add appropriate support for separate platform-dependent build
4610 directories. The recommended way to make a platform-dependent
4611 build directory is the following (tested on Linux), maybe with
4614 # Place yourself outside of the OpenSSL source tree. In
4615 # this example, the environment variable OPENSSL_SOURCE
4616 # is assumed to contain the absolute OpenSSL source directory.
4617 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
4618 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
4619 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
4620 mkdir -p `dirname $F`
4621 ln -s $OPENSSL_SOURCE/$F $F
4624 To be absolutely sure not to disturb the source tree, a "make clean"
4625 is a good thing. If it isn't successfull, don't worry about it,
4626 it probably means the source directory is very clean.
4629 *) Make sure any ENGINE control commands make local copies of string
4630 pointers passed to them whenever necessary. Otherwise it is possible
4631 the caller may have overwritten (or deallocated) the original string
4632 data when a later ENGINE operation tries to use the stored values.
4633 [Götz Babin-Ebell <babinebell@trustcenter.de>]
4635 *) Improve diagnostics in file reading and command-line digests.
4636 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
4638 *) Add AES modes CFB and OFB to the object database. Correct an
4639 error in AES-CFB decryption.
4642 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
4643 allows existing EVP_CIPHER_CTX structures to be reused after
4644 calling EVP_*Final(). This behaviour is used by encryption
4645 BIOs and some applications. This has the side effect that
4646 applications must explicitly clean up cipher contexts with
4647 EVP_CIPHER_CTX_cleanup() or they will leak memory.
4650 *) Check the values of dna and dnb in bn_mul_recursive before calling
4651 bn_mul_comba (a non zero value means the a or b arrays do not contain
4652 n2 elements) and fallback to bn_mul_normal if either is not zero.
4655 *) Fix escaping of non-ASCII characters when using the -subj option
4656 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
4659 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
4660 form for "surname", serialNumber has no short form.
4661 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
4662 therefore remove "mail" short name for "internet 7".
4663 The OID for unique identifiers in X509 certificates is
4664 x500UniqueIdentifier, not uniqueIdentifier.
4665 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
4668 *) Add an "init" command to the ENGINE config module and auto initialize
4669 ENGINEs. Without any "init" command the ENGINE will be initialized
4670 after all ctrl commands have been executed on it. If init=1 the
4671 ENGINE is initailized at that point (ctrls before that point are run
4672 on the uninitialized ENGINE and after on the initialized one). If
4673 init=0 then the ENGINE will not be iniatialized at all.
4676 *) Fix the 'app_verify_callback' interface so that the user-defined
4677 argument is actually passed to the callback: In the
4678 SSL_CTX_set_cert_verify_callback() prototype, the callback
4679 declaration has been changed from
4682 int (*cb)(X509_STORE_CTX *,void *);
4683 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
4684 i=s->ctx->app_verify_callback(&ctx)
4685 has been changed into
4686 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
4688 To update applications using SSL_CTX_set_cert_verify_callback(),
4689 a dummy argument can be added to their callback functions.
4690 [D. K. Smetters <smetters@parc.xerox.com>]
4692 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
4693 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
4695 *) Add and OPENSSL_LOAD_CONF define which will cause
4696 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
4697 This allows older applications to transparently support certain
4698 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
4699 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
4700 load the config file and OPENSSL_add_all_algorithms_conf() which will
4701 always load it have also been added.
4704 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
4705 Adjust NIDs and EVP layer.
4706 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4708 *) Config modules support in openssl utility.
4710 Most commands now load modules from the config file,
4711 though in a few (such as version) this isn't done
4712 because it couldn't be used for anything.
4714 In the case of ca and req the config file used is
4715 the same as the utility itself: that is the -config
4716 command line option can be used to specify an
4720 *) Move default behaviour from OPENSSL_config(). If appname is NULL
4721 use "openssl_conf" if filename is NULL use default openssl config file.
4724 *) Add an argument to OPENSSL_config() to allow the use of an alternative
4725 config section name. Add a new flag to tolerate a missing config file
4726 and move code to CONF_modules_load_file().
4729 *) Support for crypto accelerator cards from Accelerated Encryption
4730 Processing, www.aep.ie. (Use engine 'aep')
4731 The support was copied from 0.9.6c [engine] and adapted/corrected
4732 to work with the new engine framework.
4733 [AEP Inc. and Richard Levitte]
4735 *) Support for SureWare crypto accelerator cards from Baltimore
4736 Technologies. (Use engine 'sureware')
4737 The support was copied from 0.9.6c [engine] and adapted
4738 to work with the new engine framework.
4741 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
4742 make the newer ENGINE framework commands for the CHIL engine work.
4743 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
4745 *) Make it possible to produce shared libraries on ReliantUNIX.
4746 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
4748 *) Add the configuration target debug-linux-ppro.
4749 Make 'openssl rsa' use the general key loading routines
4750 implemented in apps.c, and make those routines able to
4751 handle the key format FORMAT_NETSCAPE and the variant
4753 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4755 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4756 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4758 *) Add -keyform to rsautl, and document -engine.
4759 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
4761 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
4762 BIO_R_NO_SUCH_FILE error code rather than the generic
4763 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
4766 *) Add new functions
4768 ERR_peek_last_error_line
4769 ERR_peek_last_error_line_data.
4770 These are similar to
4773 ERR_peek_error_line_data,
4774 but report on the latest error recorded rather than the first one
4775 still in the error queue.
4776 [Ben Laurie, Bodo Moeller]
4778 *) default_algorithms option in ENGINE config module. This allows things
4780 default_algorithms = ALL
4781 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4784 *) Prelminary ENGINE config module.
4787 *) New experimental application configuration code.
4790 *) Change the AES code to follow the same name structure as all other
4791 symmetric ciphers, and behave the same way. Move everything to
4792 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4793 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4795 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4796 [Ben Laurie and Theo de Raadt]
4798 *) Add option to output public keys in req command.
4799 [Massimiliano Pala madwolf@openca.org]
4801 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4802 (up to about 10% better than before for P-192 and P-224).
4805 *) New functions/macros
4807 SSL_CTX_set_msg_callback(ctx, cb)
4808 SSL_CTX_set_msg_callback_arg(ctx, arg)
4809 SSL_set_msg_callback(ssl, cb)
4810 SSL_set_msg_callback_arg(ssl, arg)
4812 to request calling a callback function
4814 void cb(int write_p, int version, int content_type,
4815 const void *buf, size_t len, SSL *ssl, void *arg)
4817 whenever a protocol message has been completely received
4818 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4819 protocol version according to which the SSL library interprets
4820 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4821 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4822 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4823 specification (change_cipher_spec(20), alert(21), handshake(22)).
4824 'buf' and 'len' point to the actual message, 'ssl' to the
4825 SSL object, and 'arg' is the application-defined value set by
4826 SSL[_CTX]_set_msg_callback_arg().
4828 'openssl s_client' and 'openssl s_server' have new '-msg' options
4829 to enable a callback that displays all protocol messages.
4832 *) Change the shared library support so shared libraries are built as
4833 soon as the corresponding static library is finished, and thereby get
4834 openssl and the test programs linked against the shared library.
4835 This still only happens when the keyword "shard" has been given to
4836 the configuration scripts.
4838 NOTE: shared library support is still an experimental thing, and
4839 backward binary compatibility is still not guaranteed.
4840 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4842 *) Add support for Subject Information Access extension.
4843 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4845 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4846 additional bytes when new memory had to be allocated, not just
4847 when reusing an existing buffer.
4850 *) New command line and configuration option 'utf8' for the req command.
4851 This allows field values to be specified as UTF8 strings.
4854 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4855 runs for the former and machine-readable output for the latter.
4858 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4859 of the e-mail address in the DN (i.e., it will go into a certificate
4860 extension only). The new configuration file option 'email_in_dn = no'
4861 has the same effect.
4862 [Massimiliano Pala madwolf@openca.org]
4864 *) Change all functions with names starting with des_ to be starting
4865 with DES_ instead. Add wrappers that are compatible with libdes,
4866 but are named _ossl_old_des_*. Finally, add macros that map the
4867 des_* symbols to the corresponding _ossl_old_des_* if libdes
4868 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4869 desired, the des_* symbols will be mapped to DES_*, with one
4872 Since we provide two compatibility mappings, the user needs to
4873 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4874 compatibility is desired. The default (i.e., when that macro
4875 isn't defined) is OpenSSL 0.9.6c compatibility.
4877 There are also macros that enable and disable the support of old
4878 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4879 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4880 are defined, the default will apply: to support the old des routines.
4882 In either case, one must include openssl/des.h to get the correct
4883 definitions. Do not try to just include openssl/des_old.h, that
4886 NOTE: This is a major break of an old API into a new one. Software
4887 authors are encouraged to switch to the DES_ style functions. Some
4888 time in the future, des_old.h and the libdes compatibility functions
4889 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4890 default), and then completely removed.
4893 *) Test for certificates which contain unsupported critical extensions.
4894 If such a certificate is found during a verify operation it is
4895 rejected by default: this behaviour can be overridden by either
4896 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4897 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4898 X509_supported_extension() has also been added which returns 1 if a
4899 particular extension is supported.
4902 *) Modify the behaviour of EVP cipher functions in similar way to digests
4903 to retain compatibility with existing code.
4906 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4907 compatibility with existing code. In particular the 'ctx' parameter does
4908 not have to be to be initialized before the call to EVP_DigestInit() and
4909 it is tidied up after a call to EVP_DigestFinal(). New function
4910 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4911 EVP_MD_CTX_copy() changed to not require the destination to be
4912 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4913 requires the destination to be valid.
4915 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4916 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4919 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4920 so that complete 'Handshake' protocol structures are kept in memory
4921 instead of overwriting 'msg_type' and 'length' with 'body' data.
4924 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4925 [Massimo Santin via Richard Levitte]
4927 *) Major restructuring to the underlying ENGINE code. This includes
4928 reduction of linker bloat, separation of pure "ENGINE" manipulation
4929 (initialisation, etc) from functionality dealing with implementations
4930 of specific crypto iterfaces. This change also introduces integrated
4931 support for symmetric ciphers and digest implementations - so ENGINEs
4932 can now accelerate these by providing EVP_CIPHER and EVP_MD
4933 implementations of their own. This is detailed in crypto/engine/README
4934 as it couldn't be adequately described here. However, there are a few
4935 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4936 were changed in the original introduction of ENGINE code have now
4937 reverted back - the hooking from this code to ENGINE is now a good
4938 deal more passive and at run-time, operations deal directly with
4939 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4940 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4941 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4942 they were not being used by the framework as there is no concept of a
4943 BIGNUM_METHOD and they could not be generalised to the new
4944 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4945 ENGINE_cpy() has been removed as it cannot be consistently defined in
4949 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4952 *) Change mkdef.pl to sort symbols that get the same entry number,
4953 and make sure the automatically generated functions ERR_load_*
4954 become part of libeay.num as well.
4957 *) New function SSL_renegotiate_pending(). This returns true once
4958 renegotiation has been requested (either SSL_renegotiate() call
4959 or HelloRequest/ClientHello receveived from the peer) and becomes
4960 false once a handshake has been completed.
4961 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4962 sends a HelloRequest, but does not ensure that a handshake takes
4963 place. SSL_renegotiate_pending() is useful for checking if the
4964 client has followed the request.)
4967 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4968 By default, clients may request session resumption even during
4969 renegotiation (if session ID contexts permit); with this option,
4970 session resumption is possible only in the first handshake.
4972 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4973 more bits available for options that should not be part of
4974 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4977 *) Add some demos for certificate and certificate request creation.
4980 *) Make maximum certificate chain size accepted from the peer application
4981 settable (SSL*_get/set_max_cert_list()), as proposed by
4982 "Douglas E. Engert" <deengert@anl.gov>.
4985 *) Add support for shared libraries for Unixware-7
4986 (Boyd Lynn Gerber <gerberb@zenez.com>).
4989 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4990 be done prior to destruction. Use this to unload error strings from
4991 ENGINEs that load their own error strings. NB: This adds two new API
4992 functions to "get" and "set" this destroy handler in an ENGINE.
4995 *) Alter all existing ENGINE implementations (except "openssl" and
4996 "openbsd") to dynamically instantiate their own error strings. This
4997 makes them more flexible to be built both as statically-linked ENGINEs
4998 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4999 Also, add stub code to each that makes building them as self-contained
5000 shared-libraries easier (see README.ENGINE).
5003 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
5004 implementations into applications that are completely implemented in
5005 self-contained shared-libraries. The "dynamic" ENGINE exposes control
5006 commands that can be used to configure what shared-library to load and
5007 to control aspects of the way it is handled. Also, made an update to
5008 the README.ENGINE file that brings its information up-to-date and
5009 provides some information and instructions on the "dynamic" ENGINE
5010 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5013 *) Make it possible to unload ranges of ERR strings with a new
5014 "ERR_unload_strings" function.
5017 *) Add a copy() function to EVP_MD.
5020 *) Make EVP_MD routines take a context pointer instead of just the
5021 md_data void pointer.
5024 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5025 that the digest can only process a single chunk of data
5026 (typically because it is provided by a piece of
5027 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5028 is only going to provide a single chunk of data, and hence the
5029 framework needn't accumulate the data for oneshot drivers.
5032 *) As with "ERR", make it possible to replace the underlying "ex_data"
5033 functions. This change also alters the storage and management of global
5034 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5035 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5036 index counters. The API functions that use this state have been changed
5037 to take a "class_index" rather than pointers to the class's local STACK
5038 and counter, and there is now an API function to dynamically create new
5039 classes. This centralisation allows us to (a) plug a lot of the
5040 thread-safety problems that existed, and (b) makes it possible to clean
5041 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5042 such data would previously have always leaked in application code and
5043 workarounds were in place to make the memory debugging turn a blind eye
5044 to it. Application code that doesn't use this new function will still
5045 leak as before, but their memory debugging output will announce it now
5046 rather than letting it slide.
5048 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5049 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5050 has a return value to indicate success or failure.
5053 *) Make it possible to replace the underlying "ERR" functions such that the
5054 global state (2 LHASH tables and 2 locks) is only used by the "default"
5055 implementation. This change also adds two functions to "get" and "set"
5056 the implementation prior to it being automatically set the first time
5057 any other ERR function takes place. Ie. an application can call "get",
5058 pass the return value to a module it has just loaded, and that module
5059 can call its own "set" function using that value. This means the
5060 module's "ERR" operations will use (and modify) the error state in the
5061 application and not in its own statically linked copy of OpenSSL code.
5064 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5065 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5066 the operation, and provides a more encapsulated way for external code
5067 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5068 to use these functions rather than manually incrementing the counts.
5070 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5073 *) Add EVP test program.
5076 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5079 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5080 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5081 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5082 These allow a CRL to be built without having to access X509_CRL fields
5083 directly. Modify 'ca' application to use new functions.
5086 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5087 bug workarounds. Rollback attack detection is a security feature.
5088 The problem will only arise on OpenSSL servers when TLSv1 is not
5089 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5090 Software authors not wanting to support TLSv1 will have special reasons
5091 for their choice and can explicitly enable this option.
5092 [Bodo Moeller, Lutz Jaenicke]
5094 *) Rationalise EVP so it can be extended: don't include a union of
5095 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5096 (similar to those existing for EVP_CIPHER_CTX).
5101 EVP_MD_CTX_init(&md); /* new function call */
5102 EVP_DigestInit(&md, EVP_sha1());
5103 EVP_DigestUpdate(&md, in, len);
5104 EVP_DigestFinal(&md, out, NULL);
5105 EVP_MD_CTX_cleanup(&md); /* new function call */
5109 *) Make DES key schedule conform to the usual scheme, as well as
5110 correcting its structure. This means that calls to DES functions
5111 now have to pass a pointer to a des_key_schedule instead of a
5112 plain des_key_schedule (which was actually always a pointer
5115 des_key_schedule ks;
5117 des_set_key_checked(..., &ks);
5118 des_ncbc_encrypt(..., &ks, ...);
5120 (Note that a later change renames 'des_...' into 'DES_...'.)
5123 *) Initial reduction of linker bloat: the use of some functions, such as
5124 PEM causes large amounts of unused functions to be linked in due to
5125 poor organisation. For example pem_all.c contains every PEM function
5126 which has a knock on effect of linking in large amounts of (unused)
5127 ASN1 code. Grouping together similar functions and splitting unrelated
5128 functions prevents this.
5131 *) Cleanup of EVP macros.
5134 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5135 correct _ecb suffix.
5138 *) Add initial OCSP responder support to ocsp application. The
5139 revocation information is handled using the text based index
5140 use by the ca application. The responder can either handle
5141 requests generated internally, supplied in files (for example
5142 via a CGI script) or using an internal minimal server.
5145 *) Add configuration choices to get zlib compression for TLS.
5148 *) Changes to Kerberos SSL for RFC 2712 compliance:
5149 1. Implemented real KerberosWrapper, instead of just using
5150 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5151 2. Implemented optional authenticator field of KerberosWrapper.
5153 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5154 and authenticator structs; see crypto/krb5/.
5156 Generalized Kerberos calls to support multiple Kerberos libraries.
5157 [Vern Staats <staatsvr@asc.hpc.mil>,
5158 Jeffrey Altman <jaltman@columbia.edu>
5159 via Richard Levitte]
5161 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
5162 already does with RSA. testdsa.h now has 'priv_key/pub_key'
5163 values for each of the key sizes rather than having just
5164 parameters (and 'speed' generating keys each time).
5167 *) Speed up EVP routines.
5170 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
5171 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
5172 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
5173 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
5175 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
5176 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
5177 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
5180 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
5182 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
5185 *) Added the OS2-EMX target.
5186 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
5188 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
5189 to support NCONF routines in extension code. New function CONF_set_nconf()
5190 to allow functions which take an NCONF to also handle the old LHASH
5191 structure: this means that the old CONF compatible routines can be
5192 retained (in particular wrt extensions) without having to duplicate the
5193 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
5196 *) Enhance the general user interface with mechanisms for inner control
5197 and with possibilities to have yes/no kind of prompts.
5200 *) Change all calls to low level digest routines in the library and
5201 applications to use EVP. Add missing calls to HMAC_cleanup() and
5202 don't assume HMAC_CTX can be copied using memcpy().
5203 [Verdon Walker <VWalker@novell.com>, Steve Henson]
5205 *) Add the possibility to control engines through control names but with
5206 arbitrary arguments instead of just a string.
5207 Change the key loaders to take a UI_METHOD instead of a callback
5208 function pointer. NOTE: this breaks binary compatibility with earlier
5209 versions of OpenSSL [engine].
5210 Adapt the nCipher code for these new conditions and add a card insertion
5214 *) Enhance the general user interface with mechanisms to better support
5215 dialog box interfaces, application-defined prompts, the possibility
5216 to use defaults (for example default passwords from somewhere else)
5217 and interrupts/cancellations.
5220 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
5221 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
5224 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
5225 tidy up some unnecessarily weird code in 'sk_new()').
5226 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
5228 *) Change the key loading routines for ENGINEs to use the same kind
5229 callback (pem_password_cb) as all other routines that need this
5233 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
5234 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
5235 than this minimum value is recommended.
5238 *) New random seeder for OpenVMS, using the system process statistics
5239 that are easily reachable.
5242 *) Windows apparently can't transparently handle global
5243 variables defined in DLLs. Initialisations such as:
5245 const ASN1_ITEM *it = &ASN1_INTEGER_it;
5247 wont compile. This is used by the any applications that need to
5248 declare their own ASN1 modules. This was fixed by adding the option
5249 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
5250 needed for static libraries under Win32.
5253 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
5254 setting of purpose and trust fields. New X509_STORE trust and
5255 purpose functions and tidy up setting in other SSL functions.
5258 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
5259 structure. These are inherited by X509_STORE_CTX when it is
5260 initialised. This allows various defaults to be set in the
5261 X509_STORE structure (such as flags for CRL checking and custom
5262 purpose or trust settings) for functions which only use X509_STORE_CTX
5263 internally such as S/MIME.
5265 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
5266 trust settings if they are not set in X509_STORE. This allows X509_STORE
5267 purposes and trust (in S/MIME for example) to override any set by default.
5269 Add command line options for CRL checking to smime, s_client and s_server
5273 *) Initial CRL based revocation checking. If the CRL checking flag(s)
5274 are set then the CRL is looked up in the X509_STORE structure and
5275 its validity and signature checked, then if the certificate is found
5276 in the CRL the verify fails with a revoked error.
5278 Various new CRL related callbacks added to X509_STORE_CTX structure.
5280 Command line options added to 'verify' application to support this.
5282 This needs some additional work, such as being able to handle multiple
5283 CRLs with different times, extension based lookup (rather than just
5284 by subject name) and ultimately more complete V2 CRL extension
5288 *) Add a general user interface API (crypto/ui/). This is designed
5289 to replace things like des_read_password and friends (backward
5290 compatibility functions using this new API are provided).
5291 The purpose is to remove prompting functions from the DES code
5292 section as well as provide for prompting through dialog boxes in
5293 a window system and the like.
5296 *) Add "ex_data" support to ENGINE so implementations can add state at a
5297 per-structure level rather than having to store it globally.
5300 *) Make it possible for ENGINE structures to be copied when retrieved by
5301 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
5302 This causes the "original" ENGINE structure to act like a template,
5303 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
5304 operational state can be localised to each ENGINE structure, despite the
5305 fact they all share the same "methods". New ENGINE structures returned in
5306 this case have no functional references and the return value is the single
5307 structural reference. This matches the single structural reference returned
5308 by ENGINE_by_id() normally, when it is incremented on the pre-existing
5312 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
5313 needs to match any other type at all we need to manually clear the
5317 *) Changes to the "openssl engine" utility to include;
5318 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
5319 about an ENGINE's available control commands.
5320 - executing control commands from command line arguments using the
5321 '-pre' and '-post' switches. '-post' is only used if '-t' is
5322 specified and the ENGINE is successfully initialised. The syntax for
5323 the individual commands are colon-separated, for example;
5324 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
5327 *) New dynamic control command support for ENGINEs. ENGINEs can now
5328 declare their own commands (numbers), names (strings), descriptions,
5329 and input types for run-time discovery by calling applications. A
5330 subset of these commands are implicitly classed as "executable"
5331 depending on their input type, and only these can be invoked through
5332 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
5333 can be based on user input, config files, etc). The distinction is
5334 that "executable" commands cannot return anything other than a boolean
5335 result and can only support numeric or string input, whereas some
5336 discoverable commands may only be for direct use through
5337 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
5338 pointers, or other custom uses. The "executable" commands are to
5339 support parameterisations of ENGINE behaviour that can be
5340 unambiguously defined by ENGINEs and used consistently across any
5341 OpenSSL-based application. Commands have been added to all the
5342 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
5343 control over shared-library paths without source code alterations.
5346 *) Changed all ENGINE implementations to dynamically allocate their
5347 ENGINEs rather than declaring them statically. Apart from this being
5348 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
5349 this also allows the implementations to compile without using the
5350 internal engine_int.h header.
5353 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
5354 'const' value. Any code that should be able to modify a RAND_METHOD
5355 should already have non-const pointers to it (ie. they should only
5356 modify their own ones).
5359 *) Made a variety of little tweaks to the ENGINE code.
5360 - "atalla" and "ubsec" string definitions were moved from header files
5361 to C code. "nuron" string definitions were placed in variables
5362 rather than hard-coded - allowing parameterisation of these values
5363 later on via ctrl() commands.
5364 - Removed unused "#if 0"'d code.
5365 - Fixed engine list iteration code so it uses ENGINE_free() to release
5366 structural references.
5367 - Constified the RAND_METHOD element of ENGINE structures.
5368 - Constified various get/set functions as appropriate and added
5369 missing functions (including a catch-all ENGINE_cpy that duplicates
5370 all ENGINE values onto a new ENGINE except reference counts/state).
5371 - Removed NULL parameter checks in get/set functions. Setting a method
5372 or function to NULL is a way of cancelling out a previously set
5373 value. Passing a NULL ENGINE parameter is just plain stupid anyway
5374 and doesn't justify the extra error symbols and code.
5375 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
5376 flags from engine_int.h to engine.h.
5377 - Changed prototypes for ENGINE handler functions (init(), finish(),
5378 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
5381 *) Implement binary inversion algorithm for BN_mod_inverse in addition
5382 to the algorithm using long division. The binary algorithm can be
5383 used only if the modulus is odd. On 32-bit systems, it is faster
5384 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
5385 roughly 5-15% for 256-bit moduli), so we use it only for moduli
5386 up to 450 bits. In 64-bit environments, the binary algorithm
5387 appears to be advantageous for much longer moduli; here we use it
5388 for moduli up to 2048 bits.
5391 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
5392 could not support the combine flag in choice fields.
5395 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
5396 extensions from a certificate request to the certificate.
5399 *) Allow multiple 'certopt' and 'nameopt' options to be separated
5400 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
5401 file: this allows the display of the certificate about to be
5402 signed to be customised, to allow certain fields to be included
5403 or excluded and extension details. The old system didn't display
5404 multicharacter strings properly, omitted fields not in the policy
5405 and couldn't display additional details such as extensions.
5408 *) Function EC_POINTs_mul for multiple scalar multiplication
5409 of an arbitrary number of elliptic curve points
5410 \sum scalars[i]*points[i],
5411 optionally including the generator defined for the EC_GROUP:
5412 scalar*generator + \sum scalars[i]*points[i].
5414 EC_POINT_mul is a simple wrapper function for the typical case
5415 that the point list has just one item (besides the optional
5419 *) First EC_METHODs for curves over GF(p):
5421 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
5422 operations and provides various method functions that can also
5423 operate with faster implementations of modular arithmetic.
5425 EC_GFp_mont_method() reuses most functions that are part of
5426 EC_GFp_simple_method, but uses Montgomery arithmetic.
5428 [Bodo Moeller; point addition and point doubling
5429 implementation directly derived from source code provided by
5430 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
5432 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
5433 crypto/ec/ec_lib.c):
5435 Curves are EC_GROUP objects (with an optional group generator)
5436 based on EC_METHODs that are built into the library.
5438 Points are EC_POINT objects based on EC_GROUP objects.
5440 Most of the framework would be able to handle curves over arbitrary
5441 finite fields, but as there are no obvious types for fields other
5442 than GF(p), some functions are limited to that for now.
5445 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
5446 that the file contains a complete HTTP response.
5449 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
5450 change the def and num file printf format specifier from "%-40sXXX"
5451 to "%-39s XXX". The latter will always guarantee a space after the
5452 field while the former will cause them to run together if the field
5453 is 40 of more characters long.
5456 *) Constify the cipher and digest 'method' functions and structures
5457 and modify related functions to take constant EVP_MD and EVP_CIPHER
5461 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
5462 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
5465 *) Modify EVP_Digest*() routines so they now return values. Although the
5466 internal software routines can never fail additional hardware versions
5470 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
5472 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
5473 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
5476 ERR_R_NESTED_ASN1_ERROR
5478 ERR_R_MISSING_ASN1_EOS
5479 were 4 .. 9, conflicting with
5480 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5482 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5483 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5485 Add new error code 'ERR_R_INTERNAL_ERROR'.
5488 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5492 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5493 sets the subject name for a new request or supersedes the
5494 subject name in a given request. Formats that can be parsed are
5495 'CN=Some Name, OU=myOU, C=IT'
5497 'CN=Some Name/OU=myOU/C=IT'.
5499 Add options '-batch' and '-verbose' to 'openssl req'.
5500 [Massimiliano Pala <madwolf@hackmasters.net>]
5502 *) Introduce the possibility to access global variables through
5503 functions on platform were that's the best way to handle exporting
5504 global variables in shared libraries. To enable this functionality,
5505 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5506 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5507 is normally done by Configure or something similar).
5509 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5510 in the source file (foo.c) like this:
5512 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5513 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5515 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5516 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5518 OPENSSL_DECLARE_GLOBAL(int,foo);
5519 #define foo OPENSSL_GLOBAL_REF(foo)
5520 OPENSSL_DECLARE_GLOBAL(double,bar);
5521 #define bar OPENSSL_GLOBAL_REF(bar)
5523 The #defines are very important, and therefore so is including the
5524 header file everywhere where the defined globals are used.
5526 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5527 of ASN.1 items, but that structure is a bit different.
5529 The largest change is in util/mkdef.pl which has been enhanced with
5530 better and easier to understand logic to choose which symbols should
5531 go into the Windows .def files as well as a number of fixes and code
5532 cleanup (among others, algorithm keywords are now sorted
5533 lexicographically to avoid constant rewrites).
5536 *) In BN_div() keep a copy of the sign of 'num' before writing the
5537 result to 'rm' because if rm==num the value will be overwritten
5538 and produce the wrong result if 'num' is negative: this caused
5539 problems with BN_mod() and BN_nnmod().
5542 *) Function OCSP_request_verify(). This checks the signature on an
5543 OCSP request and verifies the signer certificate. The signer
5544 certificate is just checked for a generic purpose and OCSP request
5548 *) Add OCSP_check_validity() function to check the validity of OCSP
5549 responses. OCSP responses are prepared in real time and may only
5550 be a few seconds old. Simply checking that the current time lies
5551 between thisUpdate and nextUpdate max reject otherwise valid responses
5552 caused by either OCSP responder or client clock inaccuracy. Instead
5553 we allow thisUpdate and nextUpdate to fall within a certain period of
5554 the current time. The age of the response can also optionally be
5555 checked. Two new options -validity_period and -status_age added to
5559 *) If signature or public key algorithm is unrecognized print out its
5560 OID rather that just UNKNOWN.
5563 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5564 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5565 ID to be generated from the issuer certificate alone which can then be
5566 passed to OCSP_id_issuer_cmp().
5569 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5570 ASN1 modules to export functions returning ASN1_ITEM pointers
5571 instead of the ASN1_ITEM structures themselves. This adds several
5572 new macros which allow the underlying ASN1 function/structure to
5573 be accessed transparently. As a result code should not use ASN1_ITEM
5574 references directly (such as &X509_it) but instead use the relevant
5575 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5576 use of the new ASN1 code on platforms where exporting structures
5577 is problematical (for example in shared libraries) but exporting
5578 functions returning pointers to structures is not.
5581 *) Add support for overriding the generation of SSL/TLS session IDs.
5582 These callbacks can be registered either in an SSL_CTX or per SSL.
5583 The purpose of this is to allow applications to control, if they wish,
5584 the arbitrary values chosen for use as session IDs, particularly as it
5585 can be useful for session caching in multiple-server environments. A
5586 command-line switch for testing this (and any client code that wishes
5587 to use such a feature) has been added to "s_server".
5588 [Geoff Thorpe, Lutz Jaenicke]
5590 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5591 of the form '#if defined(...) || defined(...) || ...' and
5592 '#if !defined(...) && !defined(...) && ...'. This also avoids
5593 the growing number of special cases it was previously handling.
5596 *) Make all configuration macros available for application by making
5597 sure they are available in opensslconf.h, by giving them names starting
5598 with "OPENSSL_" to avoid conflicts with other packages and by making
5599 sure e_os2.h will cover all platform-specific cases together with
5601 Additionally, it is now possible to define configuration/platform-
5602 specific names (called "system identities"). In the C code, these
5603 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
5604 macro with the name beginning with "OPENSSL_SYS_", which is determined
5605 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
5609 *) New option -set_serial to 'req' and 'x509' this allows the serial
5610 number to use to be specified on the command line. Previously self
5611 signed certificates were hard coded with serial number 0 and the
5612 CA options of 'x509' had to use a serial number in a file which was
5616 *) New options to 'ca' utility to support V2 CRL entry extensions.
5617 Currently CRL reason, invalidity date and hold instruction are
5618 supported. Add new CRL extensions to V3 code and some new objects.
5621 *) New function EVP_CIPHER_CTX_set_padding() this is used to
5622 disable standard block padding (aka PKCS#5 padding) in the EVP
5623 API, which was previously mandatory. This means that the data is
5624 not padded in any way and so the total length much be a multiple
5625 of the block size, otherwise an error occurs.
5628 *) Initial (incomplete) OCSP SSL support.
5631 *) New function OCSP_parse_url(). This splits up a URL into its host,
5632 port and path components: primarily to parse OCSP URLs. New -url
5633 option to ocsp utility.
5636 *) New nonce behavior. The return value of OCSP_check_nonce() now
5637 reflects the various checks performed. Applications can decide
5638 whether to tolerate certain situations such as an absent nonce
5639 in a response when one was present in a request: the ocsp application
5640 just prints out a warning. New function OCSP_add1_basic_nonce()
5641 this is to allow responders to include a nonce in a response even if
5642 the request is nonce-less.
5645 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
5646 skipped when using openssl x509 multiple times on a single input file,
5647 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
5650 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
5651 set string type: to handle setting ASN1_TIME structures. Fix ca
5652 utility to correctly initialize revocation date of CRLs.
5655 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
5656 the clients preferred ciphersuites and rather use its own preferences.
5657 Should help to work around M$ SGC (Server Gated Cryptography) bug in
5658 Internet Explorer by ensuring unchanged hash method during stepup.
5659 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
5662 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
5663 to aes and add a new 'exist' option to print out symbols that don't
5667 *) Additional options to ocsp utility to allow flags to be set and
5668 additional certificates supplied.
5671 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
5672 OCSP client a number of certificate to only verify the response
5676 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
5677 handle the new API. Currently only ECB, CBC modes supported. Add new
5680 Add TLS AES ciphersuites as described in RFC3268, "Advanced
5681 Encryption Standard (AES) Ciphersuites for Transport Layer
5682 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
5683 not enabled by default and were not part of the "ALL" ciphersuite
5684 alias because they were not yet official; they could be
5685 explicitly requested by specifying the "AESdraft" ciphersuite
5686 group alias. In the final release of OpenSSL 0.9.7, the group
5687 alias is called "AES" and is part of "ALL".)
5688 [Ben Laurie, Steve Henson, Bodo Moeller]
5690 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
5691 request to response.
5694 *) Functions for OCSP responders. OCSP_request_onereq_count(),
5695 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
5696 extract information from a certificate request. OCSP_response_create()
5697 creates a response and optionally adds a basic response structure.
5698 OCSP_basic_add1_status() adds a complete single response to a basic
5699 response and returns the OCSP_SINGLERESP structure just added (to allow
5700 extensions to be included for example). OCSP_basic_add1_cert() adds a
5701 certificate to a basic response and OCSP_basic_sign() signs a basic
5702 response with various flags. New helper functions ASN1_TIME_check()
5703 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
5704 (converts ASN1_TIME to GeneralizedTime).
5707 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
5708 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
5709 structure from a certificate. X509_pubkey_digest() digests the public_key
5710 contents: this is used in various key identifiers.
5713 *) Make sk_sort() tolerate a NULL argument.
5714 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
5716 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
5717 passed by the function are trusted implicitly. If any of them signed the
5718 response then it is assumed to be valid and is not verified.
5721 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
5722 to data. This was previously part of the PKCS7 ASN1 code. This
5723 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
5724 [Steve Henson, reported by Kenneth R. Robinette
5725 <support@securenetterm.com>]
5727 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
5728 routines: without these tracing memory leaks is very painful.
5729 Fix leaks in PKCS12 and PKCS7 routines.
5732 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
5733 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
5734 effectively meant GeneralizedTime would never be used. Now it
5735 is initialised to -1 but X509_time_adj() now has to check the value
5736 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
5737 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
5738 [Steve Henson, reported by Kenneth R. Robinette
5739 <support@securenetterm.com>]
5741 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
5742 result in a zero length in the ASN1_INTEGER structure which was
5743 not consistent with the structure when d2i_ASN1_INTEGER() was used
5744 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
5745 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
5746 where it did not print out a minus for negative ASN1_INTEGER.
5749 *) Add summary printout to ocsp utility. The various functions which
5750 convert status values to strings have been renamed to:
5751 OCSP_response_status_str(), OCSP_cert_status_str() and
5752 OCSP_crl_reason_str() and are no longer static. New options
5753 to verify nonce values and to disable verification. OCSP response
5754 printout format cleaned up.
5757 *) Add additional OCSP certificate checks. These are those specified
5758 in RFC2560. This consists of two separate checks: the CA of the
5759 certificate being checked must either be the OCSP signer certificate
5760 or the issuer of the OCSP signer certificate. In the latter case the
5761 OCSP signer certificate must contain the OCSP signing extended key
5762 usage. This check is performed by attempting to match the OCSP
5763 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
5764 in the OCSP_CERTID structures of the response.
5767 *) Initial OCSP certificate verification added to OCSP_basic_verify()
5768 and related routines. This uses the standard OpenSSL certificate
5769 verify routines to perform initial checks (just CA validity) and
5770 to obtain the certificate chain. Then additional checks will be
5771 performed on the chain. Currently the root CA is checked to see
5772 if it is explicitly trusted for OCSP signing. This is used to set
5773 a root CA as a global signing root: that is any certificate that
5774 chains to that CA is an acceptable OCSP signing certificate.
5777 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5778 extensions from a separate configuration file.
5779 As when reading extensions from the main configuration file,
5780 the '-extensions ...' option may be used for specifying the
5782 [Massimiliano Pala <madwolf@comune.modena.it>]
5784 *) New OCSP utility. Allows OCSP requests to be generated or
5785 read. The request can be sent to a responder and the output
5786 parsed, outputed or printed in text form. Not complete yet:
5787 still needs to check the OCSP response validity.
5790 *) New subcommands for 'openssl ca':
5791 'openssl ca -status <serial>' prints the status of the cert with
5792 the given serial number (according to the index file).
5793 'openssl ca -updatedb' updates the expiry status of certificates
5795 [Massimiliano Pala <madwolf@comune.modena.it>]
5797 *) New '-newreq-nodes' command option to CA.pl. This is like
5798 '-newreq', but calls 'openssl req' with the '-nodes' option
5799 so that the resulting key is not encrypted.
5800 [Damien Miller <djm@mindrot.org>]
5802 *) New configuration for the GNU Hurd.
5803 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5805 *) Initial code to implement OCSP basic response verify. This
5806 is currently incomplete. Currently just finds the signer's
5807 certificate and verifies the signature on the response.
5810 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5811 value of OPENSSLDIR. This is available via the new '-d' option
5812 to 'openssl version', and is also included in 'openssl version -a'.
5815 *) Allowing defining memory allocation callbacks that will be given
5816 file name and line number information in additional arguments
5817 (a const char* and an int). The basic functionality remains, as
5818 well as the original possibility to just replace malloc(),
5819 realloc() and free() by functions that do not know about these
5820 additional arguments. To register and find out the current
5821 settings for extended allocation functions, the following
5822 functions are provided:
5824 CRYPTO_set_mem_ex_functions
5825 CRYPTO_set_locked_mem_ex_functions
5826 CRYPTO_get_mem_ex_functions
5827 CRYPTO_get_locked_mem_ex_functions
5829 These work the same way as CRYPTO_set_mem_functions and friends.
5830 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5831 extended allocation function is enabled.
5832 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5833 a conventional allocation function is enabled.
5834 [Richard Levitte, Bodo Moeller]
5836 *) Finish off removing the remaining LHASH function pointer casts.
5837 There should no longer be any prototype-casting required when using
5838 the LHASH abstraction, and any casts that remain are "bugs". See
5839 the callback types and macros at the head of lhash.h for details
5840 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5843 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5844 If /dev/[u]random devices are not available or do not return enough
5845 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5847 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5848 /etc/entropy will be queried once each in this sequence, quering stops
5849 when enough entropy was collected without querying more sockets.
5852 *) Change the Unix RAND_poll() variant to be able to poll several
5853 random devices, as specified by DEVRANDOM, until a sufficient amount
5854 of data has been collected. We spend at most 10 ms on each file
5855 (select timeout) and read in non-blocking mode. DEVRANDOM now
5856 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5857 (previously it was just the string "/dev/urandom"), so on typical
5858 platforms the 10 ms delay will never occur.
5859 Also separate out the Unix variant to its own file, rand_unix.c.
5860 For VMS, there's a currently-empty rand_vms.c.
5863 *) Move OCSP client related routines to ocsp_cl.c. These
5864 provide utility functions which an application needing
5865 to issue a request to an OCSP responder and analyse the
5866 response will typically need: as opposed to those which an
5867 OCSP responder itself would need which will be added later.
5869 OCSP_request_sign() signs an OCSP request with an API similar
5870 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5871 response. OCSP_response_get1_basic() extracts basic response
5872 from response. OCSP_resp_find_status(): finds and extracts status
5873 information from an OCSP_CERTID structure (which will be created
5874 when the request structure is built). These are built from lower
5875 level functions which work on OCSP_SINGLERESP structures but
5876 wont normally be used unless the application wishes to examine
5877 extensions in the OCSP response for example.
5879 Replace nonce routines with a pair of functions.
5880 OCSP_request_add1_nonce() adds a nonce value and optionally
5881 generates a random value. OCSP_check_nonce() checks the
5882 validity of the nonce in an OCSP response.
5885 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5886 This doesn't copy the supplied OCSP_CERTID and avoids the
5887 need to free up the newly created id. Change return type
5888 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5889 This can then be used to add extensions to the request.
5890 Deleted OCSP_request_new(), since most of its functionality
5891 is now in OCSP_REQUEST_new() (and the case insensitive name
5892 clash) apart from the ability to set the request name which
5893 will be added elsewhere.
5896 *) Update OCSP API. Remove obsolete extensions argument from
5897 various functions. Extensions are now handled using the new
5898 OCSP extension code. New simple OCSP HTTP function which
5899 can be used to send requests and parse the response.
5902 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5903 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5904 uses the special reorder version of SET OF to sort the attributes
5905 and reorder them to match the encoded order. This resolves a long
5906 standing problem: a verify on a PKCS7 structure just after signing
5907 it used to fail because the attribute order did not match the
5908 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5909 it uses the received order. This is necessary to tolerate some broken
5910 software that does not order SET OF. This is handled by encoding
5911 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5912 to produce the required SET OF.
5915 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5916 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5917 files to get correct declarations of the ASN.1 item variables.
5920 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5921 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5922 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5923 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5924 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5925 ASN1_ITEM and no wrapper functions.
5928 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5929 replace the old function pointer based I/O routines. Change most of
5930 the *_d2i_bio() and *_d2i_fp() functions to use these.
5933 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5934 lines, recognice more "algorithms" that can be deselected, and make
5935 it complain about algorithm deselection that isn't recognised.
5938 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5939 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5940 to use new functions. Add NO_ASN1_OLD which can be set to remove
5941 some old style ASN1 functions: this can be used to determine if old
5942 code will still work when these eventually go away.
5945 *) New extension functions for OCSP structures, these follow the
5946 same conventions as certificates and CRLs.
5949 *) New function X509V3_add1_i2d(). This automatically encodes and
5950 adds an extension. Its behaviour can be customised with various
5951 flags to append, replace or delete. Various wrappers added for
5952 certifcates and CRLs.
5955 *) Fix to avoid calling the underlying ASN1 print routine when
5956 an extension cannot be parsed. Correct a typo in the
5957 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5960 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5961 entries for variables.
5964 *) Add functionality to apps/openssl.c for detecting locking
5965 problems: As the program is single-threaded, all we have
5966 to do is register a locking callback using an array for
5967 storing which locks are currently held by the program.
5970 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5971 SSL_get_ex_data_X509_STORE_idx(), which is used in
5972 ssl_verify_cert_chain() and thus can be called at any time
5973 during TLS/SSL handshakes so that thread-safety is essential.
5974 Unfortunately, the ex_data design is not at all suited
5975 for multi-threaded use, so it probably should be abolished.
5978 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5979 [Broadcom, tweaked and integrated by Geoff Thorpe]
5981 *) Move common extension printing code to new function
5982 X509V3_print_extensions(). Reorganise OCSP print routines and
5983 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5986 *) New function X509_signature_print() to remove duplication in some
5990 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5991 set (this was treated exactly the same as SET OF previously). This
5992 is used to reorder the STACK representing the structure to match the
5993 encoding. This will be used to get round a problem where a PKCS7
5994 structure which was signed could not be verified because the STACK
5995 order did not reflect the encoded order.
5998 *) Reimplement the OCSP ASN1 module using the new code.
6001 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
6002 for its ASN1 operations. The old style function pointers still exist
6003 for now but they will eventually go away.
6006 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
6007 completely replaces the old ASN1 functionality with a table driven
6008 encoder and decoder which interprets an ASN1_ITEM structure describing
6009 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6010 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6011 has also been converted to the new form.
6014 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6015 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6016 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6017 for negative moduli.
6020 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6021 of not touching the result's sign bit.
6024 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6028 *) Changed the LHASH code to use prototypes for callbacks, and created
6029 macros to declare and implement thin (optionally static) functions
6030 that provide type-safety and avoid function pointer casting for the
6031 type-specific callbacks.
6034 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6036 [Veers Staats <staatsvr@asc.hpc.mil>,
6037 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6039 *) Reformat the FAQ so the different questions and answers can be divided
6040 in sections depending on the subject.
6043 *) Have the zlib compression code load ZLIB.DLL dynamically under
6047 *) New function BN_mod_sqrt for computing square roots modulo a prime
6048 (using the probabilistic Tonelli-Shanks algorithm unless
6049 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6050 be handled deterministically).
6051 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6053 *) Make BN_mod_inverse faster by explicitly handling small quotients
6054 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6055 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6058 *) New function BN_kronecker.
6061 *) Fix BN_gcd so that it works on negative inputs; the result is
6062 positive unless both parameters are zero.
6063 Previously something reasonably close to an infinite loop was
6064 possible because numbers could be growing instead of shrinking
6065 in the implementation of Euclid's algorithm.
6068 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6069 sign of the number in question.
6071 Fix BN_is_word(a,w) to work correctly for w == 0.
6073 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6074 because its test if the absolute value of 'a' equals 'w'.
6075 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6076 it exists mostly for use in the implementations of BN_is_zero(),
6077 BN_is_one(), and BN_is_word().
6080 *) New function BN_swap.
6083 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6084 the exponentiation functions are more likely to produce reasonable
6085 results on negative inputs.
6088 *) Change BN_mod_mul so that the result is always non-negative.
6089 Previously, it could be negative if one of the factors was negative;
6090 I don't think anyone really wanted that behaviour.
6093 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6094 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6095 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6096 and add new functions:
6105 BN_mod_lshift1_quick
6109 These functions always generate non-negative results.
6111 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6112 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6114 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6115 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6116 be reduced modulo m.
6117 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6120 The following entry accidentily appeared in the CHANGES file
6121 distributed with OpenSSL 0.9.7. The modifications described in
6122 it do *not* apply to OpenSSL 0.9.7.
6124 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6125 was actually never needed) and in BN_mul(). The removal in BN_mul()
6126 required a small change in bn_mul_part_recursive() and the addition
6127 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6128 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6129 bn_sub_words() and bn_add_words() except they take arrays with
6134 *) In 'openssl passwd', verify passwords read from the terminal
6135 unless the '-salt' option is used (which usually means that
6136 verification would just waste user's time since the resulting
6137 hash is going to be compared with some given password hash)
6138 or the new '-noverify' option is used.
6140 This is an incompatible change, but it does not affect
6141 non-interactive use of 'openssl passwd' (passwords on the command
6142 line, '-stdin' option, '-in ...' option) and thus should not
6146 *) Remove all references to RSAref, since there's no more need for it.
6149 *) Make DSO load along a path given through an environment variable
6150 (SHLIB_PATH) with shl_load().
6153 *) Constify the ENGINE code as a result of BIGNUM constification.
6154 Also constify the RSA code and most things related to it. In a
6155 few places, most notable in the depth of the ASN.1 code, ugly
6156 casts back to non-const were required (to be solved at a later
6160 *) Make it so the openssl application has all engines loaded by default.
6163 *) Constify the BIGNUM routines a little more.
6166 *) Add the following functions:
6168 ENGINE_load_cswift()
6170 ENGINE_load_atalla()
6172 ENGINE_load_builtin_engines()
6174 That way, an application can itself choose if external engines that
6175 are built-in in OpenSSL shall ever be used or not. The benefit is
6176 that applications won't have to be linked with libdl or other dso
6177 libraries unless it's really needed.
6179 Changed 'openssl engine' to load all engines on demand.
6180 Changed the engine header files to avoid the duplication of some
6181 declarations (they differed!).
6184 *) 'openssl engine' can now list capabilities.
6187 *) Better error reporting in 'openssl engine'.
6190 *) Never call load_dh_param(NULL) in s_server.
6193 *) Add engine application. It can currently list engines by name and
6194 identity, and test if they are actually available.
6197 *) Improve RPM specification file by forcing symbolic linking and making
6198 sure the installed documentation is also owned by root.root.
6199 [Damien Miller <djm@mindrot.org>]
6201 *) Give the OpenSSL applications more possibilities to make use of
6202 keys (public as well as private) handled by engines.
6205 *) Add OCSP code that comes from CertCo.
6208 *) Add VMS support for the Rijndael code.
6211 *) Added untested support for Nuron crypto accelerator.
6214 *) Add support for external cryptographic devices. This code was
6215 previously distributed separately as the "engine" branch.
6216 [Geoff Thorpe, Richard Levitte]
6218 *) Rework the filename-translation in the DSO code. It is now possible to
6219 have far greater control over how a "name" is turned into a filename
6220 depending on the operating environment and any oddities about the
6221 different shared library filenames on each system.
6224 *) Support threads on FreeBSD-elf in Configure.
6227 *) Fix for SHA1 assembly problem with MASM: it produces
6228 warnings about corrupt line number information when assembling
6229 with debugging information. This is caused by the overlapping
6231 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
6234 NCONF_get_number() has no error checking at all. As a replacement,
6235 NCONF_get_number_e() is defined (_e for "error checking") and is
6236 promoted strongly. The old NCONF_get_number is kept around for
6237 binary backward compatibility.
6238 Make it possible for methods to load from something other than a BIO,
6239 by providing a function pointer that is given a name instead of a BIO.
6240 For example, this could be used to load configuration data from an
6244 *) Fix for non blocking accept BIOs. Added new I/O special reason
6245 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
6246 with non blocking I/O was not possible because no retry code was
6247 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
6251 *) Added the beginnings of Rijndael support.
6254 *) Fix for bug in DirectoryString mask setting. Add support for
6255 X509_NAME_print_ex() in 'req' and X509_print_ex() function
6256 to allow certificate printing to more controllable, additional
6257 'certopt' option to 'x509' to allow new printing options to be
6261 *) Clean old EAY MD5 hack from e_os.h.
6264 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
6266 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6267 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6268 [Joe Orton, Steve Henson]
6270 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
6272 *) Fix additional bug revealed by the NISCC test suite:
6274 Stop bug triggering large recursion when presented with
6275 certain ASN.1 tags (CVE-2003-0851)
6278 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
6280 *) Fix various bugs revealed by running the NISCC test suite:
6282 Stop out of bounds reads in the ASN1 code when presented with
6283 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6285 If verify callback ignores invalid public key errors don't try to check
6286 certificate signature with the NULL public key.
6290 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6291 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6295 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6296 extra data after the compression methods not only for TLS 1.0
6297 but also for SSL 3.0 (as required by the specification).
6298 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6300 *) Change X509_certificate_type() to mark the key as exported/exportable
6301 when it's 512 *bits* long, not 512 bytes.
6304 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
6306 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6307 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6308 a protocol version number mismatch like a decryption error
6309 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6312 *) Turn on RSA blinding by default in the default implementation
6313 to avoid a timing attack. Applications that don't want it can call
6314 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6315 They would be ill-advised to do so in most cases.
6316 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6318 *) Change RSA blinding code so that it works when the PRNG is not
6319 seeded (in this case, the secret RSA exponent is abused as
6320 an unpredictable seed -- if it is not unpredictable, there
6321 is no point in blinding anyway). Make RSA blinding thread-safe
6322 by remembering the creator's thread ID in rsa->blinding and
6323 having all other threads use local one-time blinding factors
6324 (this requires more computation than sharing rsa->blinding, but
6325 avoids excessive locking; and if an RSA object is not shared
6326 between threads, blinding will still be very fast).
6329 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
6331 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6332 via timing by performing a MAC computation even if incorrrect
6333 block cipher padding has been found. This is a countermeasure
6334 against active attacks where the attacker has to distinguish
6335 between bad padding and a MAC verification error. (CVE-2003-0078)
6337 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6338 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6339 Martin Vuagnoux (EPFL, Ilion)]
6341 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
6343 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
6344 memory from it's contents. This is done with a counter that will
6345 place alternating values in each byte. This can be used to solve
6346 two issues: 1) the removal of calls to memset() by highly optimizing
6347 compilers, and 2) cleansing with other values than 0, since those can
6348 be read through on certain media, for example a swap space on disk.
6351 *) Bugfix: client side session caching did not work with external caching,
6352 because the session->cipher setting was not restored when reloading
6353 from the external cache. This problem was masked, when
6354 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
6355 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
6358 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
6359 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
6360 [Zeev Lieber <zeev-l@yahoo.com>]
6362 *) Undo an undocumented change introduced in 0.9.6e which caused
6363 repeated calls to OpenSSL_add_all_ciphers() and
6364 OpenSSL_add_all_digests() to be ignored, even after calling
6368 *) Change the default configuration reader to deal with last line not
6369 being properly terminated.
6372 *) Change X509_NAME_cmp() so it applies the special rules on handling
6373 DN values that are of type PrintableString, as well as RDNs of type
6374 emailAddress where the value has the type ia5String.
6375 [stefank@valicert.com via Richard Levitte]
6377 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
6378 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
6379 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
6380 the bitwise-OR of the two for use by the majority of applications
6381 wanting this behaviour, and update the docs. The documented
6382 behaviour and actual behaviour were inconsistent and had been
6383 changing anyway, so this is more a bug-fix than a behavioural
6385 [Geoff Thorpe, diagnosed by Nadav Har'El]
6387 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
6388 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
6391 *) Fix initialization code race conditions in
6392 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
6393 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
6394 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
6395 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
6396 ssl2_get_cipher_by_char(),
6397 ssl3_get_cipher_by_char().
6398 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
6400 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
6401 the cached sessions are flushed, as the remove_cb() might use ex_data
6402 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
6403 (see [openssl.org #212]).
6404 [Geoff Thorpe, Lutz Jaenicke]
6406 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
6407 length, instead of the encoding length to d2i_ASN1_OBJECT.
6410 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
6412 *) [In 0.9.6g-engine release:]
6413 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
6414 [Lynn Gazis <lgazis@rainbow.com>]
6416 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
6418 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
6419 and get fix the header length calculation.
6420 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
6421 Alon Kantor <alonk@checkpoint.com> (and others),
6424 *) Use proper error handling instead of 'assertions' in buffer
6425 overflow checks added in 0.9.6e. This prevents DoS (the
6426 assertions could call abort()).
6427 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
6429 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
6431 *) Add various sanity checks to asn1_get_length() to reject
6432 the ASN1 length bytes if they exceed sizeof(long), will appear
6433 negative or the content length exceeds the length of the
6435 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6437 *) Fix cipher selection routines: ciphers without encryption had no flags
6438 for the cipher strength set and where therefore not handled correctly
6439 by the selection routines (PR #130).
6442 *) Fix EVP_dsa_sha macro.
6446 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
6447 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
6448 that was added in OpenSSL 0.9.6d.
6450 As the countermeasure turned out to be incompatible with some
6451 broken SSL implementations, the new option is part of SSL_OP_ALL.
6452 SSL_OP_ALL is usually employed when compatibility with weird SSL
6453 implementations is desired (e.g. '-bugs' option to 's_client' and
6454 's_server'), so the new option is automatically set in many
6458 *) Changes in security patch:
6460 Changes marked "(CHATS)" were sponsored by the Defense Advanced
6461 Research Projects Agency (DARPA) and Air Force Research Laboratory,
6462 Air Force Materiel Command, USAF, under agreement number
6465 *) Add various sanity checks to asn1_get_length() to reject
6466 the ASN1 length bytes if they exceed sizeof(long), will appear
6467 negative or the content length exceeds the length of the
6468 supplied buffer. (CVE-2002-0659)
6469 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6471 *) Assertions for various potential buffer overflows, not known to
6473 [Ben Laurie (CHATS)]
6475 *) Various temporary buffers to hold ASCII versions of integers were
6476 too small for 64 bit platforms. (CVE-2002-0655)
6477 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6479 *) Remote buffer overflow in SSL3 protocol - an attacker could
6480 supply an oversized session ID to a client. (CVE-2002-0656)
6481 [Ben Laurie (CHATS)]
6483 *) Remote buffer overflow in SSL2 protocol - an attacker could
6484 supply an oversized client master key. (CVE-2002-0656)
6485 [Ben Laurie (CHATS)]
6487 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6489 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6490 encoded as NULL) with id-dsa-with-sha1.
6491 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6493 *) Check various X509_...() return values in apps/req.c.
6494 [Nils Larsch <nla@trustcenter.de>]
6496 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6497 an end-of-file condition would erronously be flagged, when the CRLF
6498 was just at the end of a processed block. The bug was discovered when
6499 processing data through a buffering memory BIO handing the data to a
6500 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6501 <ptsekov@syntrex.com> and Nedelcho Stanev.
6504 *) Implement a countermeasure against a vulnerability recently found
6505 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6506 before application data chunks to avoid the use of known IVs
6507 with data potentially chosen by the attacker.
6510 *) Fix length checks in ssl3_get_client_hello().
6513 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6514 to prevent ssl3_read_internal() from incorrectly assuming that
6515 ssl3_read_bytes() found application data while handshake
6516 processing was enabled when in fact s->s3->in_read_app_data was
6517 merely automatically cleared during the initial handshake.
6518 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6520 *) Fix object definitions for Private and Enterprise: they were not
6521 recognized in their shortname (=lowercase) representation. Extend
6522 obj_dat.pl to issue an error when using undefined keywords instead
6523 of silently ignoring the problem (Svenning Sorensen
6524 <sss@sss.dnsalias.net>).
6527 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6528 generators, i.e. generators other than 2 and 5. (Previously, the
6529 code did not properly initialise the 'add' and 'rem' values to
6530 BN_generate_prime().)
6532 In the new general case, we do not insist that 'generator' is
6533 actually a primitive root: This requirement is rather pointless;
6534 a generator of the order-q subgroup is just as good, if not
6538 *) Map new X509 verification errors to alerts. Discovered and submitted by
6539 Tom Wu <tom@arcot.com>.
6542 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6543 returning non-zero before the data has been completely received
6544 when using non-blocking I/O.
6545 [Bodo Moeller; problem pointed out by John Hughes]
6547 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6548 [Ben Laurie, Lutz Jaenicke]
6550 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6551 Yoram Zahavi <YoramZ@gilian.com>).
6554 *) Add information about CygWin 1.3 and on, and preserve proper
6555 configuration for the versions before that.
6556 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6558 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6559 check whether we deal with a copy of a session and do not delete from
6560 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6561 <izhar@checkpoint.com>.
6564 *) Do not store session data into the internal session cache, if it
6565 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6566 flag is set). Proposed by Aslam <aslam@funk.com>.
6569 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6573 *) [In 0.9.6d-engine release:]
6574 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6575 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6577 *) Add the configuration target linux-s390x.
6578 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6580 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6581 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6582 variable as an indication that a ClientHello message has been
6583 received. As the flag value will be lost between multiple
6584 invocations of ssl3_accept when using non-blocking I/O, the
6585 function may not be aware that a handshake has actually taken
6586 place, thus preventing a new session from being added to the
6589 To avoid this problem, we now set s->new_session to 2 instead of
6590 using a local variable.
6591 [Lutz Jaenicke, Bodo Moeller]
6593 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6594 if the SSL_R_LENGTH_MISMATCH error is detected.
6595 [Geoff Thorpe, Bodo Moeller]
6597 *) New 'shared_ldflag' column in Configure platform table.
6600 *) Fix EVP_CIPHER_mode macro.
6601 ["Dan S. Camper" <dan@bti.net>]
6603 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
6604 type, we must throw them away by setting rr->length to 0.
6605 [D P Chang <dpc@qualys.com>]
6607 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
6609 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
6610 <Dominikus.Scherkl@biodata.com>. (The previous implementation
6611 worked incorrectly for those cases where range = 10..._2 and
6612 3*range is two bits longer than range.)
6615 *) Only add signing time to PKCS7 structures if it is not already
6619 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
6620 OBJ_ld_ce should be OBJ_id_ce.
6621 Also some ip-pda OIDs in crypto/objects/objects.txt were
6622 incorrect (cf. RFC 3039).
6623 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
6625 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
6626 returns early because it has nothing to do.
6627 [Andy Schneider <andy.schneider@bjss.co.uk>]
6629 *) [In 0.9.6c-engine release:]
6630 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
6631 [Andy Schneider <andy.schneider@bjss.co.uk>]
6633 *) [In 0.9.6c-engine release:]
6634 Add support for Cryptographic Appliance's keyserver technology.
6635 (Use engine 'keyclient')
6636 [Cryptographic Appliances and Geoff Thorpe]
6638 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
6639 is called via tools/c89.sh because arguments have to be
6640 rearranged (all '-L' options must appear before the first object
6642 [Richard Shapiro <rshapiro@abinitio.com>]
6644 *) [In 0.9.6c-engine release:]
6645 Add support for Broadcom crypto accelerator cards, backported
6647 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
6649 *) [In 0.9.6c-engine release:]
6650 Add support for SureWare crypto accelerator cards from
6651 Baltimore Technologies. (Use engine 'sureware')
6652 [Baltimore Technologies and Mark Cox]
6654 *) [In 0.9.6c-engine release:]
6655 Add support for crypto accelerator cards from Accelerated
6656 Encryption Processing, www.aep.ie. (Use engine 'aep')
6657 [AEP Inc. and Mark Cox]
6659 *) Add a configuration entry for gcc on UnixWare.
6660 [Gary Benson <gbenson@redhat.com>]
6662 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
6663 messages are stored in a single piece (fixed-length part and
6664 variable-length part combined) and fix various bugs found on the way.
6667 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
6668 instead. BIO_gethostbyname() does not know what timeouts are
6669 appropriate, so entries would stay in cache even when they have
6671 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
6673 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
6674 faced with a pathologically small ClientHello fragment that does
6675 not contain client_version: Instead of aborting with an error,
6676 simply choose the highest available protocol version (i.e.,
6677 TLS 1.0 unless it is disabled). In practice, ClientHello
6678 messages are never sent like this, but this change gives us
6679 strictly correct behaviour at least for TLS.
6682 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
6683 never resets s->method to s->ctx->method when called from within
6684 one of the SSL handshake functions.
6685 [Bodo Moeller; problem pointed out by Niko Baric]
6687 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
6688 (sent using the client's version number) if client_version is
6689 smaller than the protocol version in use. Also change
6690 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
6691 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
6692 the client will at least see that alert.
6695 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
6699 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
6700 client receives HelloRequest while in a handshake.
6701 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
6703 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
6704 should end in 'break', not 'goto end' which circuments various
6705 cleanups done in state SSL_ST_OK. But session related stuff
6706 must be disabled for SSL_ST_OK in the case that we just sent a
6709 Also avoid some overhead by not calling ssl_init_wbio_buffer()
6710 before just sending a HelloRequest.
6711 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
6713 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
6714 reveal whether illegal block cipher padding was found or a MAC
6715 verification error occured. (Neither SSLerr() codes nor alerts
6716 are directly visible to potential attackers, but the information
6717 may leak via logfiles.)
6719 Similar changes are not required for the SSL 2.0 implementation
6720 because the number of padding bytes is sent in clear for SSL 2.0,
6721 and the extra bytes are just ignored. However ssl/s2_pkt.c
6722 failed to verify that the purported number of padding bytes is in
6726 *) Add OpenUNIX-8 support including shared libraries
6727 (Boyd Lynn Gerber <gerberb@zenez.com>).
6730 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
6731 'wristwatch attack' using huge encoding parameters (cf.
6732 James H. Manger's CRYPTO 2001 paper). Note that the
6733 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
6734 encoding parameters and hence was not vulnerable.
6737 *) BN_sqr() bug fix.
6738 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
6740 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
6741 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
6742 followed by modular reduction.
6743 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
6745 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
6746 equivalent based on BN_pseudo_rand() instead of BN_rand().
6749 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
6750 This function was broken, as the check for a new client hello message
6751 to handle SGC did not allow these large messages.
6752 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
6755 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
6758 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
6759 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
6762 *) Rework the configuration and shared library support for Tru64 Unix.
6763 The configuration part makes use of modern compiler features and
6764 still retains old compiler behavior for those that run older versions
6765 of the OS. The shared library support part includes a variant that
6766 uses the RPATH feature, and is available through the special
6767 configuration target "alpha-cc-rpath", which will never be selected
6769 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
6771 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
6772 with the same message size as in ssl3_get_certificate_request().
6773 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
6774 messages might inadvertently be reject as too long.
6775 [Petr Lampa <lampa@fee.vutbr.cz>]
6777 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6780 *) Modified SSL library such that the verify_callback that has been set
6781 specificly for an SSL object with SSL_set_verify() is actually being
6782 used. Before the change, a verify_callback set with this function was
6783 ignored and the verify_callback() set in the SSL_CTX at the time of
6784 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6785 to allow the necessary settings.
6788 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6789 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6790 done automatically (in contradiction to the requirements of the C
6791 standard). This made problems when used from OpenSSH.
6794 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6795 dh->length and always used
6797 BN_rand_range(priv_key, dh->p).
6799 BN_rand_range() is not necessary for Diffie-Hellman, and this
6800 specific range makes Diffie-Hellman unnecessarily inefficient if
6801 dh->length (recommended exponent length) is much smaller than the
6802 length of dh->p. We could use BN_rand_range() if the order of
6803 the subgroup was stored in the DH structure, but we only have
6808 BN_rand(priv_key, l, ...)
6810 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6816 RSA_eay_public_encrypt
6817 RSA_eay_private_decrypt
6818 RSA_eay_private_encrypt (signing)
6819 RSA_eay_public_decrypt (signature verification)
6821 (default implementations for RSA_public_encrypt,
6822 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6823 always reject numbers >= n.
6826 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6827 to synchronize access to 'locking_thread'. This is necessary on
6828 systems where access to 'locking_thread' (an 'unsigned long'
6829 variable) is not atomic.
6832 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6833 *before* setting the 'crypto_lock_rand' flag. The previous code had
6834 a race condition if 0 is a valid thread ID.
6835 [Travis Vitek <vitek@roguewave.com>]
6837 *) Add support for shared libraries under Irix.
6838 [Albert Chin-A-Young <china@thewrittenword.com>]
6840 *) Add configuration option to build on Linux on both big-endian and
6842 [Ralf Baechle <ralf@uni-koblenz.de>]
6844 *) Add the possibility to create shared libraries on HP-UX.
6847 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6849 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6850 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6851 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6852 PRNG state recovery was possible based on the output of
6853 one PRNG request appropriately sized to gain knowledge on
6854 'md' followed by enough consecutive 1-byte PRNG requests
6855 to traverse all of 'state'.
6857 1. When updating 'md_local' (the current thread's copy of 'md')
6858 during PRNG output generation, hash all of the previous
6859 'md_local' value, not just the half used for PRNG output.
6861 2. Make the number of bytes from 'state' included into the hash
6862 independent from the number of PRNG bytes requested.
6864 The first measure alone would be sufficient to avoid
6865 Markku-Juhani's attack. (Actually it had never occurred
6866 to me that the half of 'md_local' used for chaining was the
6867 half from which PRNG output bytes were taken -- I had always
6868 assumed that the secret half would be used.) The second
6869 measure makes sure that additional data from 'state' is never
6870 mixed into 'md_local' in small portions; this heuristically
6871 further strengthens the PRNG.
6874 *) Fix crypto/bn/asm/mips3.s.
6877 *) When only the key is given to "enc", the IV is undefined. Print out
6878 an error message in this case.
6881 *) Handle special case when X509_NAME is empty in X509 printing routines.
6884 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6885 positive and less than q.
6888 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6889 used: it isn't thread safe and the add_lock_callback should handle
6891 [Paul Rose <Paul.Rose@bridge.com>]
6893 *) Verify that incoming data obeys the block size in
6894 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6898 [Ulf Möller, Bodo Möller]
6900 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6901 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6902 when fixing the server behaviour for backwards-compatible 'client
6903 hello' messages. (Note that the attack is impractical against
6904 SSL 3.0 and TLS 1.0 anyway because length and version checking
6905 means that the probability of guessing a valid ciphertext is
6906 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6909 Before 0.9.5, the countermeasure (hide the error by generating a
6910 random 'decryption result') did not work properly because
6911 ERR_clear_error() was missing, meaning that SSL_get_error() would
6912 detect the supposedly ignored error.
6914 Both problems are now fixed.
6917 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6918 (previously it was 1024).
6921 *) Fix for compatibility mode trust settings: ignore trust settings
6922 unless some valid trust or reject settings are present.
6925 *) Fix for blowfish EVP: its a variable length cipher.
6928 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6929 parameters in DSA public key structures and return an error in the
6930 DSA routines if parameters are absent.
6933 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6934 in the current directory if neither $RANDFILE nor $HOME was set.
6935 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6936 caused some confusion to Windows users who haven't defined $HOME.
6937 Thus RAND_file_name() is changed again: e_os.h can define a
6938 DEFAULT_HOME, which will be used if $HOME is not set.
6939 For Windows, we use "C:"; on other platforms, we still require
6940 environment variables.
6942 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6943 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6944 having multiple threads call RAND_poll() concurrently.
6947 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6948 combination of a flag and a thread ID variable.
6949 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6950 flag), *other* threads can enter ssleay_add_bytes without obeying
6951 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6952 that they do not hold after the first thread unsets add_do_not_lock).
6955 *) Change bctest again: '-x' expressions are not available in all
6959 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6961 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6962 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6964 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6965 the default extension for executables, if any. Also, make the perl
6966 scripts that use symlink() to test if it really exists and use "cp"
6967 if it doesn't. All this made OpenSSL compilable and installable in
6971 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6972 If SEQUENCE is length is indefinite just set c->slen to the total
6973 amount of data available.
6974 [Steve Henson, reported by shige@FreeBSD.org]
6975 [This change does not apply to 0.9.7.]
6977 *) Change bctest to avoid here-documents inside command substitution
6978 (workaround for FreeBSD /bin/sh bug).
6979 For compatibility with Ultrix, avoid shell functions (introduced
6980 in the bctest version that searches along $PATH).
6983 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6984 with des_encrypt() defined on some operating systems, like Solaris
6988 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6989 On the Importance of Eliminating Errors in Cryptographic
6990 Computations, J. Cryptology 14 (2001) 2, 101-119,
6991 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6994 *) MIPS assembler BIGNUM division bug fix.
6997 *) Disabled incorrect Alpha assembler code.
7000 *) Fix PKCS#7 decode routines so they correctly update the length
7001 after reading an EOC for the EXPLICIT tag.
7003 [This change does not apply to 0.9.7.]
7005 *) Fix bug in PKCS#12 key generation routines. This was triggered
7006 if a 3DES key was generated with a 0 initial byte. Include
7007 PKCS12_BROKEN_KEYGEN compilation option to retain the old
7008 (but broken) behaviour.
7011 *) Enhance bctest to search for a working bc along $PATH and print
7013 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7015 *) Fix memory leaks in err.c: free err_data string if necessary;
7016 don't write to the wrong index in ERR_set_error_data.
7019 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7023 *) Replace rdtsc with _emit statements for VC++ version 5.
7024 [Jeremy Cooper <jeremy@baymoo.org>]
7026 *) Make it possible to reuse SSLv2 sessions.
7029 *) In copy_email() check for >= 0 as a return value for
7030 X509_NAME_get_index_by_NID() since 0 is a valid index.
7031 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7033 *) Avoid coredump with unsupported or invalid public keys by checking if
7034 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7035 PKCS7_verify() fails with non detached data.
7038 *) Don't use getenv in library functions when run as setuid/setgid.
7039 New function OPENSSL_issetugid().
7042 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7043 due to incorrect handling of multi-threading:
7045 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7047 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7049 3. Count how many times MemCheck_off() has been called so that
7050 nested use can be treated correctly. This also avoids
7051 inband-signalling in the previous code (which relied on the
7052 assumption that thread ID 0 is impossible).
7055 *) Add "-rand" option also to s_client and s_server.
7058 *) Fix CPU detection on Irix 6.x.
7059 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7060 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7062 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7065 [This change does not apply to 0.9.7.]
7067 *) Use the cached encoding of an X509_NAME structure rather than
7068 copying it. This is apparently the reason for the libsafe "errors"
7069 but the code is actually correct.
7072 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7073 Bleichenbacher's DSA attack.
7074 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7075 to be set and top=0 forces the highest bit to be set; top=-1 is new
7076 and leaves the highest bit random.
7077 [Ulf Moeller, Bodo Moeller]
7079 *) In the NCONF_...-based implementations for CONF_... queries
7080 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7081 a temporary CONF structure with the data component set to NULL
7082 (which gives segmentation faults in lh_retrieve).
7083 Instead, use NULL for the CONF pointer in CONF_get_string and
7084 CONF_get_number (which may use environment variables) and directly
7085 return NULL from CONF_get_section.
7088 *) Fix potential buffer overrun for EBCDIC.
7091 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7092 keyUsage if basicConstraints absent for a CA.
7095 *) Make SMIME_write_PKCS7() write mail header values with a format that
7096 is more generally accepted (no spaces before the semicolon), since
7097 some programs can't parse those values properly otherwise. Also make
7098 sure BIO's that break lines after each write do not create invalid
7102 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7103 macros previously used would not encode an empty SEQUENCE OF
7104 and break the signature.
7106 [This change does not apply to 0.9.7.]
7108 *) Zero the premaster secret after deriving the master secret in
7112 *) Add some EVP_add_digest_alias registrations (as found in
7113 OpenSSL_add_all_digests()) to SSL_library_init()
7114 aka OpenSSL_add_ssl_algorithms(). This provides improved
7115 compatibility with peers using X.509 certificates
7116 with unconventional AlgorithmIdentifier OIDs.
7119 *) Fix for Irix with NO_ASM.
7120 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7122 *) ./config script fixes.
7123 [Ulf Moeller, Richard Levitte]
7125 *) Fix 'openssl passwd -1'.
7128 *) Change PKCS12_key_gen_asc() so it can cope with non null
7129 terminated strings whose length is passed in the passlen
7130 parameter, for example from PEM callbacks. This was done
7131 by adding an extra length parameter to asc2uni().
7132 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7134 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7135 call failed, free the DSA structure.
7138 *) Fix to uni2asc() to cope with zero length Unicode strings.
7139 These are present in some PKCS#12 files.
7142 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7143 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7144 when writing a 32767 byte record.
7145 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7147 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7148 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7150 (RSA objects have a reference count access to which is protected
7151 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7152 so they are meant to be shared between threads.)
7153 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7154 "Reddie, Steven" <Steven.Reddie@ca.com>]
7156 *) Fix a deadlock in CRYPTO_mem_leaks().
7159 *) Use better test patterns in bntest.
7162 *) rand_win.c fix for Borland C.
7165 *) BN_rshift bugfix for n == 0.
7168 *) Add a 'bctest' script that checks for some known 'bc' bugs
7169 so that 'make test' does not abort just because 'bc' is broken.
7172 *) Store verify_result within SSL_SESSION also for client side to
7173 avoid potential security hole. (Re-used sessions on the client side
7174 always resulted in verify_result==X509_V_OK, not using the original
7175 result of the server certificate verification.)
7178 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
7179 SSL3_RT_APPLICATION_DATA, return 0.
7180 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
7184 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
7185 releases, have been re-implemented by renaming the previous
7186 implementations of ssl2_read and ssl3_read to ssl2_read_internal
7187 and ssl3_read_internal, respectively, and adding 'peek' parameters
7188 to them. The new ssl[23]_{read,peek} functions are calls to
7189 ssl[23]_read_internal with the 'peek' flag set appropriately.
7190 A 'peek' parameter has also been added to ssl3_read_bytes, which
7191 does the actual work for ssl3_read_internal.
7194 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
7195 the method-specific "init()" handler. Also clean up ex_data after
7196 calling the method-specific "finish()" handler. Previously, this was
7197 happening the other way round.
7200 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
7201 The previous value, 12, was not always sufficient for BN_mod_exp().
7204 *) Make sure that shared libraries get the internal name engine with
7205 the full version number and not just 0. This should mark the
7206 shared libraries as not backward compatible. Of course, this should
7207 be changed again when we can guarantee backward binary compatibility.
7210 *) Fix typo in get_cert_by_subject() in by_dir.c
7211 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
7213 *) Rework the system to generate shared libraries:
7215 - Make note of the expected extension for the shared libraries and
7216 if there is a need for symbolic links from for example libcrypto.so.0
7217 to libcrypto.so.0.9.7. There is extended info in Configure for
7220 - Make as few rebuilds of the shared libraries as possible.
7222 - Still avoid linking the OpenSSL programs with the shared libraries.
7224 - When installing, install the shared libraries separately from the
7228 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
7230 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
7231 and not in SSL_clear because the latter is also used by the
7232 accept/connect functions; previously, the settings made by
7233 SSL_set_read_ahead would be lost during the handshake.
7234 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
7236 *) Correct util/mkdef.pl to be selective about disabled algorithms.
7237 Previously, it would create entries for disableed algorithms no
7241 *) Added several new manual pages for SSL_* function.
7244 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
7246 *) In ssl23_get_client_hello, generate an error message when faced
7247 with an initial SSL 3.0/TLS record that is too small to contain the
7248 first two bytes of the ClientHello message, i.e. client_version.
7249 (Note that this is a pathologic case that probably has never happened
7250 in real life.) The previous approach was to use the version number
7251 from the record header as a substitute; but our protocol choice
7252 should not depend on that one because it is not authenticated
7253 by the Finished messages.
7256 *) More robust randomness gathering functions for Windows.
7257 [Jeffrey Altman <jaltman@columbia.edu>]
7259 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
7260 not set then we don't setup the error code for issuer check errors
7261 to avoid possibly overwriting other errors which the callback does
7262 handle. If an application does set the flag then we assume it knows
7263 what it is doing and can handle the new informational codes
7267 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
7268 a general "ANY" type, as such it should be able to decode anything
7269 including tagged types. However it didn't check the class so it would
7270 wrongly interpret tagged types in the same way as their universal
7271 counterpart and unknown types were just rejected. Changed so that the
7272 tagged and unknown types are handled in the same way as a SEQUENCE:
7273 that is the encoding is stored intact. There is also a new type
7274 "V_ASN1_OTHER" which is used when the class is not universal, in this
7275 case we have no idea what the actual type is so we just lump them all
7279 *) On VMS, stdout may very well lead to a file that is written to
7280 in a record-oriented fashion. That means that every write() will
7281 write a separate record, which will be read separately by the
7282 programs trying to read from it. This can be very confusing.
7284 The solution is to put a BIO filter in the way that will buffer
7285 text until a linefeed is reached, and then write everything a
7286 line at a time, so every record written will be an actual line,
7287 not chunks of lines and not (usually doesn't happen, but I've
7288 seen it once) several lines in one record. BIO_f_linebuffer() is
7291 Currently, it's a VMS-only method, because that's where it has
7292 been tested well enough.
7295 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
7296 it can return incorrect results.
7297 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
7298 but it was in 0.9.6-beta[12].)
7301 *) Disable the check for content being present when verifying detached
7302 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
7303 include zero length content when signing messages.
7306 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
7307 BIO_ctrl (for BIO pairs).
7310 *) Add DSO method for VMS.
7313 *) Bug fix: Montgomery multiplication could produce results with the
7317 *) Add RPM specification openssl.spec and modify it to build three
7318 packages. The default package contains applications, application
7319 documentation and run-time libraries. The devel package contains
7320 include files, static libraries and function documentation. The
7321 doc package contains the contents of the doc directory. The original
7322 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
7325 *) Add a large number of documentation files for many SSL routines.
7326 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7328 *) Add a configuration entry for Sony News 4.
7329 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
7331 *) Don't set the two most significant bits to one when generating a
7332 random number < q in the DSA library.
7335 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
7336 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
7337 the underlying transport is blocking) if a handshake took place.
7338 (The default behaviour is needed by applications such as s_client
7339 and s_server that use select() to determine when to use SSL_read;
7340 but for applications that know in advance when to expect data, it
7341 just makes things more complicated.)
7344 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
7348 *) Add a few more EBCDIC conditionals that make `req' and `x509'
7349 work better on such systems.
7350 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7352 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
7353 Update PKCS12_parse() so it copies the friendlyName and the
7354 keyid to the certificates aux info.
7357 *) Fix bug in PKCS7_verify() which caused an infinite loop
7358 if there was more than one signature.
7359 [Sven Uszpelkat <su@celocom.de>]
7361 *) Major change in util/mkdef.pl to include extra information
7362 about each symbol, as well as presentig variables as well
7363 as functions. This change means that there's n more need
7364 to rebuild the .num files when some algorithms are excluded.
7367 *) Allow the verify time to be set by an application,
7368 rather than always using the current time.
7371 *) Phase 2 verify code reorganisation. The certificate
7372 verify code now looks up an issuer certificate by a
7373 number of criteria: subject name, authority key id
7374 and key usage. It also verifies self signed certificates
7375 by the same criteria. The main comparison function is
7376 X509_check_issued() which performs these checks.
7378 Lot of changes were necessary in order to support this
7379 without completely rewriting the lookup code.
7381 Authority and subject key identifier are now cached.
7383 The LHASH 'certs' is X509_STORE has now been replaced
7384 by a STACK_OF(X509_OBJECT). This is mainly because an
7385 LHASH can't store or retrieve multiple objects with
7386 the same hash value.
7388 As a result various functions (which were all internal
7389 use only) have changed to handle the new X509_STORE
7390 structure. This will break anything that messed round
7391 with X509_STORE internally.
7393 The functions X509_STORE_add_cert() now checks for an
7394 exact match, rather than just subject name.
7396 The X509_STORE API doesn't directly support the retrieval
7397 of multiple certificates matching a given criteria, however
7398 this can be worked round by performing a lookup first
7399 (which will fill the cache with candidate certificates)
7400 and then examining the cache for matches. This is probably
7401 the best we can do without throwing out X509_LOOKUP
7402 entirely (maybe later...).
7404 The X509_VERIFY_CTX structure has been enhanced considerably.
7406 All certificate lookup operations now go via a get_issuer()
7407 callback. Although this currently uses an X509_STORE it
7408 can be replaced by custom lookups. This is a simple way
7409 to bypass the X509_STORE hackery necessary to make this
7410 work and makes it possible to use more efficient techniques
7411 in future. A very simple version which uses a simple
7412 STACK for its trusted certificate store is also provided
7413 using X509_STORE_CTX_trusted_stack().
7415 The verify_cb() and verify() callbacks now have equivalents
7416 in the X509_STORE_CTX structure.
7418 X509_STORE_CTX also has a 'flags' field which can be used
7419 to customise the verify behaviour.
7422 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
7423 excludes S/MIME capabilities.
7426 *) When a certificate request is read in keep a copy of the
7427 original encoding of the signed data and use it when outputing
7428 again. Signatures then use the original encoding rather than
7429 a decoded, encoded version which may cause problems if the
7430 request is improperly encoded.
7433 *) For consistency with other BIO_puts implementations, call
7434 buffer_write(b, ...) directly in buffer_puts instead of calling
7437 In BIO_puts, increment b->num_write as in BIO_write.
7438 [Peter.Sylvester@EdelWeb.fr]
7440 *) Fix BN_mul_word for the case where the word is 0. (We have to use
7441 BN_zero, we may not return a BIGNUM with an array consisting of
7445 *) Avoid calling abort() from within the library when problems are
7446 detected, except if preprocessor symbols have been defined
7447 (such as REF_CHECK, BN_DEBUG etc.).
7450 *) New openssl application 'rsautl'. This utility can be
7451 used for low level RSA operations. DER public key
7452 BIO/fp routines also added.
7455 *) New Configure entry and patches for compiling on QNX 4.
7456 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
7458 *) A demo state-machine implementation was sponsored by
7459 Nuron (http://www.nuron.com/) and is now available in
7460 demos/state_machine.
7463 *) New options added to the 'dgst' utility for signature
7464 generation and verification.
7467 *) Unrecognized PKCS#7 content types are now handled via a
7468 catch all ASN1_TYPE structure. This allows unsupported
7469 types to be stored as a "blob" and an application can
7470 encode and decode it manually.
7473 *) Fix various signed/unsigned issues to make a_strex.c
7475 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7477 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7478 length if passed a buffer. ASN1_INTEGER_to_BN failed
7479 if passed a NULL BN and its argument was negative.
7480 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7482 *) Modification to PKCS#7 encoding routines to output definite
7483 length encoding. Since currently the whole structures are in
7484 memory there's not real point in using indefinite length
7485 constructed encoding. However if OpenSSL is compiled with
7486 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7489 *) Added BIO_vprintf() and BIO_vsnprintf().
7492 *) Added more prefixes to parse for in the the strings written
7493 through a logging bio, to cover all the levels that are available
7494 through syslog. The prefixes are now:
7496 PANIC, EMERG, EMR => LOG_EMERG
7497 ALERT, ALR => LOG_ALERT
7498 CRIT, CRI => LOG_CRIT
7499 ERROR, ERR => LOG_ERR
7500 WARNING, WARN, WAR => LOG_WARNING
7501 NOTICE, NOTE, NOT => LOG_NOTICE
7502 INFO, INF => LOG_INFO
7503 DEBUG, DBG => LOG_DEBUG
7505 and as before, if none of those prefixes are present at the
7506 beginning of the string, LOG_ERR is chosen.
7508 On Win32, the LOG_* levels are mapped according to this:
7510 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7511 LOG_WARNING => EVENTLOG_WARNING_TYPE
7512 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7516 *) Made it possible to reconfigure with just the configuration
7517 argument "reconf" or "reconfigure". The command line arguments
7518 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7519 and are retrieved from there when reconfiguring.
7523 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7525 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7528 *) The obj_dat.pl script was messing up the sorting of object
7529 names. The reason was that it compared the quoted version
7530 of strings as a result "OCSP" > "OCSP Signing" because
7531 " > SPACE. Changed script to store unquoted versions of
7532 names and add quotes on output. It was also omitting some
7533 names from the lookup table if they were given a default
7534 value (that is if SN is missing it is given the same
7535 value as LN and vice versa), these are now added on the
7536 grounds that if an object has a name we should be able to
7537 look it up. Finally added warning output when duplicate
7538 short or long names are found.
7541 *) Changes needed for Tandem NSK.
7542 [Scott Uroff <scott@xypro.com>]
7544 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7545 RSA_padding_check_SSLv23(), special padding was never detected
7546 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7547 version rollback attacks was not effective.
7549 In s23_clnt.c, don't use special rollback-attack detection padding
7550 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7551 client; similarly, in s23_srvr.c, don't do the rollback check if
7552 SSL 2.0 is the only protocol enabled in the server.
7555 *) Make it possible to get hexdumps of unprintable data with 'openssl
7556 asn1parse'. By implication, the functions ASN1_parse_dump() and
7557 BIO_dump_indent() are added.
7560 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7561 these print out strings and name structures based on various
7562 flags including RFC2253 support and proper handling of
7563 multibyte characters. Added options to the 'x509' utility
7564 to allow the various flags to be set.
7567 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7568 Also change the functions X509_cmp_current_time() and
7569 X509_gmtime_adj() work with an ASN1_TIME structure,
7570 this will enable certificates using GeneralizedTime in validity
7571 dates to be checked.
7574 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7575 negative public key encodings) on by default,
7576 NO_NEG_PUBKEY_BUG can be set to disable it.
7579 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7580 content octets. An i2c_ASN1_OBJECT is unnecessary because
7581 the encoding can be trivially obtained from the structure.
7584 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7585 not read locks (CRYPTO_r_[un]lock).
7588 *) A first attempt at creating official support for shared
7589 libraries through configuration. I've kept it so the
7590 default is static libraries only, and the OpenSSL programs
7591 are always statically linked for now, but there are
7592 preparations for dynamic linking in place.
7593 This has been tested on Linux and Tru64.
7596 *) Randomness polling function for Win9x, as described in:
7597 Peter Gutmann, Software Generation of Practically Strong
7601 *) Fix so PRNG is seeded in req if using an already existing
7605 *) New options to smime application. -inform and -outform
7606 allow alternative formats for the S/MIME message including
7607 PEM and DER. The -content option allows the content to be
7608 specified separately. This should allow things like Netscape
7609 form signing output easier to verify.
7612 *) Fix the ASN1 encoding of tags using the 'long form'.
7615 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
7616 STRING types. These convert content octets to and from the
7617 underlying type. The actual tag and length octets are
7618 already assumed to have been read in and checked. These
7619 are needed because all other string types have virtually
7620 identical handling apart from the tag. By having versions
7621 of the ASN1 functions that just operate on content octets
7622 IMPLICIT tagging can be handled properly. It also allows
7623 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
7624 and ASN1_INTEGER are identical apart from the tag.
7627 *) Change the handling of OID objects as follows:
7629 - New object identifiers are inserted in objects.txt, following
7630 the syntax given in objects.README.
7631 - objects.pl is used to process obj_mac.num and create a new
7633 - obj_dat.pl is used to create a new obj_dat.h, using the data in
7636 This is currently kind of a hack, and the perl code in objects.pl
7637 isn't very elegant, but it works as I intended. The simplest way
7638 to check that it worked correctly is to look in obj_dat.h and
7639 check the array nid_objs and make sure the objects haven't moved
7640 around (this is important!). Additions are OK, as well as
7641 consistent name changes.
7644 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
7647 *) Addition of the command line parameter '-rand file' to 'openssl req'.
7648 The given file adds to whatever has already been seeded into the
7649 random pool through the RANDFILE configuration file option or
7650 environment variable, or the default random state file.
7653 *) mkstack.pl now sorts each macro group into lexical order.
7654 Previously the output order depended on the order the files
7655 appeared in the directory, resulting in needless rewriting
7659 *) Patches to make OpenSSL compile under Win32 again. Mostly
7660 work arounds for the VC++ problem that it treats func() as
7661 func(void). Also stripped out the parts of mkdef.pl that
7662 added extra typesafe functions: these no longer exist.
7665 *) Reorganisation of the stack code. The macros are now all
7666 collected in safestack.h . Each macro is defined in terms of
7667 a "stack macro" of the form SKM_<name>(type, a, b). The
7668 DEBUG_SAFESTACK is now handled in terms of function casts,
7669 this has the advantage of retaining type safety without the
7670 use of additional functions. If DEBUG_SAFESTACK is not defined
7671 then the non typesafe macros are used instead. Also modified the
7672 mkstack.pl script to handle the new form. Needs testing to see
7673 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
7674 the default if no major problems. Similar behaviour for ASN1_SET_OF
7675 and PKCS12_STACK_OF.
7678 *) When some versions of IIS use the 'NET' form of private key the
7679 key derivation algorithm is different. Normally MD5(password) is
7680 used as a 128 bit RC4 key. In the modified case
7681 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
7682 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
7683 as the old Netscape_RSA functions except they have an additional
7684 'sgckey' parameter which uses the modified algorithm. Also added
7685 an -sgckey command line option to the rsa utility. Thanks to
7686 Adrian Peck <bertie@ncipher.com> for posting details of the modified
7687 algorithm to openssl-dev.
7690 *) The evp_local.h macros were using 'c.##kname' which resulted in
7691 invalid expansion on some systems (SCO 5.0.5 for example).
7692 Corrected to 'c.kname'.
7693 [Phillip Porch <root@theporch.com>]
7695 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
7696 a STACK of email addresses from a certificate or request, these look
7697 in the subject name and the subject alternative name extensions and
7698 omit any duplicate addresses.
7701 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
7702 This makes DSA verification about 2 % faster.
7705 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
7706 (meaning that now 2^5 values will be precomputed, which is only 4 KB
7707 plus overhead for 1024 bit moduli).
7708 This makes exponentiations about 0.5 % faster for 1024 bit
7709 exponents (as measured by "openssl speed rsa2048").
7712 *) Rename memory handling macros to avoid conflicts with other
7714 Malloc => OPENSSL_malloc
7715 Malloc_locked => OPENSSL_malloc_locked
7716 Realloc => OPENSSL_realloc
7717 Free => OPENSSL_free
7720 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
7721 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
7724 *) CygWin32 support.
7725 [John Jarvie <jjarvie@newsguy.com>]
7727 *) The type-safe stack code has been rejigged. It is now only compiled
7728 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
7729 by default all type-specific stack functions are "#define"d back to
7730 standard stack functions. This results in more streamlined output
7731 but retains the type-safety checking possibilities of the original
7735 *) The STACK code has been cleaned up, and certain type declarations
7736 that didn't make a lot of sense have been brought in line. This has
7737 also involved a cleanup of sorts in safestack.h to more correctly
7738 map type-safe stack functions onto their plain stack counterparts.
7739 This work has also resulted in a variety of "const"ifications of
7740 lots of the code, especially "_cmp" operations which should normally
7741 be prototyped with "const" parameters anyway.
7744 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
7745 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
7746 (The PRNG state consists of two parts, the large pool 'state' and 'md',
7747 where all of 'md' is used each time the PRNG is used, but 'state'
7748 is used only indexed by a cyclic counter. As entropy may not be
7749 well distributed from the beginning, 'md' is important as a
7750 chaining variable. However, the output function chains only half
7751 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
7752 all of 'md', and seeding with STATE_SIZE dummy bytes will result
7753 in all of 'state' being rewritten, with the new values depending
7754 on virtually all of 'md'. This overcomes the 80 bit limitation.)
7757 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
7758 the handshake is continued after ssl_verify_cert_chain();
7759 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
7760 can lead to 'unexplainable' connection aborts later.
7761 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
7763 *) Major EVP API cipher revision.
7764 Add hooks for extra EVP features. This allows various cipher
7765 parameters to be set in the EVP interface. Support added for variable
7766 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
7767 setting of RC2 and RC5 parameters.
7769 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
7772 Remove lots of duplicated code from the EVP library. For example *every*
7773 cipher init() function handles the 'iv' in the same way according to the
7774 cipher mode. They also all do nothing if the 'key' parameter is NULL and
7775 for CFB and OFB modes they zero ctx->num.
7777 New functionality allows removal of S/MIME code RC2 hack.
7779 Most of the routines have the same form and so can be declared in terms
7782 By shifting this to the top level EVP_CipherInit() it can be removed from
7783 all individual ciphers. If the cipher wants to handle IVs or keys
7784 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7787 Change lots of functions like EVP_EncryptUpdate() to now return a
7788 value: although software versions of the algorithms cannot fail
7789 any installed hardware versions can.
7792 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7793 this option is set, tolerate broken clients that send the negotiated
7794 protocol version number instead of the requested protocol version
7798 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7799 i.e. non-zero for export ciphersuites, zero otherwise.
7800 Previous versions had this flag inverted, inconsistent with
7801 rsa_tmp_cb (..._TMP_RSA_CB).
7802 [Bodo Moeller; problem reported by Amit Chopra]
7804 *) Add missing DSA library text string. Work around for some IIS
7805 key files with invalid SEQUENCE encoding.
7808 *) Add a document (doc/standards.txt) that list all kinds of standards
7809 and so on that are implemented in OpenSSL.
7812 *) Enhance c_rehash script. Old version would mishandle certificates
7813 with the same subject name hash and wouldn't handle CRLs at all.
7814 Added -fingerprint option to crl utility, to support new c_rehash
7818 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7821 *) Fix for SSL server purpose checking. Server checking was
7822 rejecting certificates which had extended key usage present
7823 but no ssl client purpose.
7824 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7826 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7827 is a little unclear about how a blank password is handled.
7828 Since the password in encoded as a BMPString with terminating
7829 double NULL a zero length password would end up as just the
7830 double NULL. However no password at all is different and is
7831 handled differently in the PKCS#12 key generation code. NS
7832 treats a blank password as zero length. MSIE treats it as no
7833 password on export: but it will try both on import. We now do
7834 the same: PKCS12_parse() tries zero length and no password if
7835 the password is set to "" or NULL (NULL is now a valid password:
7836 it wasn't before) as does the pkcs12 application.
7839 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7840 perror when PEM_read_bio_X509_REQ fails, the error message must
7841 be obtained from the error queue.
7844 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7845 it in ERR_remove_state if appropriate, and change ERR_get_state
7846 accordingly to avoid race conditions (this is necessary because
7847 thread_hash is no longer constant once set).
7850 *) Bugfix for linux-elf makefile.one.
7853 *) RSA_get_default_method() will now cause a default
7854 RSA_METHOD to be chosen if one doesn't exist already.
7855 Previously this was only set during a call to RSA_new()
7856 or RSA_new_method(NULL) meaning it was possible for
7857 RSA_get_default_method() to return NULL.
7860 *) Added native name translation to the existing DSO code
7861 that will convert (if the flag to do so is set) filenames
7862 that are sufficiently small and have no path information
7863 into a canonical native form. Eg. "blah" converted to
7864 "libblah.so" or "blah.dll" etc.
7867 *) New function ERR_error_string_n(e, buf, len) which is like
7868 ERR_error_string(e, buf), but writes at most 'len' bytes
7869 including the 0 terminator. For ERR_error_string_n, 'buf'
7871 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7873 *) CONF library reworked to become more general. A new CONF
7874 configuration file reader "class" is implemented as well as a
7875 new functions (NCONF_*, for "New CONF") to handle it. The now
7876 old CONF_* functions are still there, but are reimplemented to
7877 work in terms of the new functions. Also, a set of functions
7878 to handle the internal storage of the configuration data is
7879 provided to make it easier to write new configuration file
7880 reader "classes" (I can definitely see something reading a
7881 configuration file in XML format, for example), called _CONF_*,
7882 or "the configuration storage API"...
7884 The new configuration file reading functions are:
7886 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7887 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7889 NCONF_default, NCONF_WIN32
7891 NCONF_dump_fp, NCONF_dump_bio
7893 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7894 NCONF_new creates a new CONF object. This works in the same way
7895 as other interfaces in OpenSSL, like the BIO interface.
7896 NCONF_dump_* dump the internal storage of the configuration file,
7897 which is useful for debugging. All other functions take the same
7898 arguments as the old CONF_* functions wth the exception of the
7899 first that must be a `CONF *' instead of a `LHASH *'.
7901 To make it easer to use the new classes with the old CONF_* functions,
7902 the function CONF_set_default_method is provided.
7905 *) Add '-tls1' option to 'openssl ciphers', which was already
7906 mentioned in the documentation but had not been implemented.
7907 (This option is not yet really useful because even the additional
7908 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7911 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7912 OpenSSL-based applications) load shared libraries and bind to
7913 them in a portable way.
7914 [Geoff Thorpe, with contributions from Richard Levitte]
7916 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7918 *) Make sure _lrotl and _lrotr are only used with MSVC.
7920 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7921 (the default implementation of RAND_status).
7923 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7924 to '-clrext' (= clear extensions), as intended and documented.
7925 [Bodo Moeller; inconsistency pointed out by Michael Attili
7928 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7929 was larger than the MD block size.
7930 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7932 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7933 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7934 using the passed key: if the passed key was a private key the result
7935 of X509_print(), for example, would be to print out all the private key
7939 *) des_quad_cksum() byte order bug fix.
7940 [Ulf Möller, using the problem description in krb4-0.9.7, where
7941 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7943 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7945 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7947 *) For easily testing in shell scripts whether some command
7948 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7949 returns with exit code 0 iff no command of the given name is available.
7950 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7951 the output goes to stdout and nothing is printed to stderr.
7952 Additional arguments are always ignored.
7954 Since for each cipher there is a command of the same name,
7955 the 'no-cipher' compilation switches can be tested this way.
7957 ('openssl no-XXX' is not able to detect pseudo-commands such
7958 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7961 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7964 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7965 is set; it will be thrown away anyway because each handshake creates
7967 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7968 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7969 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7970 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7973 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7974 'Q' and 'R' lose their special meanings (quit/renegotiate).
7975 This is part of what -quiet does; unlike -quiet, -ign_eof
7976 does not suppress any output.
7979 *) Add compatibility options to the purpose and trust code. The
7980 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7981 accepts a certificate or CA, this was the previous behaviour,
7982 with all the associated security issues.
7984 X509_TRUST_COMPAT is the old trust behaviour: only and
7985 automatically trust self signed roots in certificate store. A
7986 new trust setting X509_TRUST_DEFAULT is used to specify that
7987 a purpose has no associated trust setting and it should instead
7988 use the value in the default purpose.
7991 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7992 and fix a memory leak.
7995 *) In util/mkerr.pl (which implements 'make errors'), preserve
7996 reason strings from the previous version of the .c file, as
7997 the default to have only downcase letters (and digits) in
7998 automatically generated reasons codes is not always appropriate.
8001 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
8002 using strerror. Previously, ERR_reason_error_string() returned
8003 library names as reason strings for SYSerr; but SYSerr is a special
8004 case where small numbers are errno values, not library numbers.
8007 *) Add '-dsaparam' option to 'openssl dhparam' application. This
8008 converts DSA parameters into DH parameters. (When creating parameters,
8009 DSA_generate_parameters is used.)
8012 *) Include 'length' (recommended exponent length) in C code generated
8013 by 'openssl dhparam -C'.
8016 *) The second argument to set_label in perlasm was already being used
8017 so couldn't be used as a "file scope" flag. Moved to third argument
8021 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8022 instead of RAND_bytes for encryption IVs and salts.
8025 *) Include RAND_status() into RAND_METHOD instead of implementing
8026 it only for md_rand.c Otherwise replacing the PRNG by calling
8027 RAND_set_rand_method would be impossible.
8030 *) Don't let DSA_generate_key() enter an infinite loop if the random
8031 number generation fails.
8034 *) New 'rand' application for creating pseudo-random output.
8037 *) Added configuration support for Linux/IA64
8038 [Rolf Haberrecker <rolf@suse.de>]
8040 *) Assembler module support for Mingw32.
8043 *) Shared library support for HPUX (in shlib/).
8044 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8046 *) Shared library support for Solaris gcc.
8047 [Lutz Behnke <behnke@trustcenter.de>]
8049 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8051 *) PKCS7_encrypt() was adding text MIME headers twice because they
8052 were added manually and by SMIME_crlf_copy().
8055 *) In bntest.c don't call BN_rand with zero bits argument.
8056 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8058 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8059 case was implemented. This caused BN_div_recp() to fail occasionally.
8062 *) Add an optional second argument to the set_label() in the perl
8063 assembly language builder. If this argument exists and is set
8064 to 1 it signals that the assembler should use a symbol whose
8065 scope is the entire file, not just the current function. This
8066 is needed with MASM which uses the format label:: for this scope.
8067 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8069 *) Change the ASN1 types so they are typedefs by default. Before
8070 almost all types were #define'd to ASN1_STRING which was causing
8071 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8075 *) Change names of new functions to the new get1/get0 naming
8076 convention: After 'get1', the caller owns a reference count
8077 and has to call ..._free; 'get0' returns a pointer to some
8078 data structure without incrementing reference counters.
8079 (Some of the existing 'get' functions increment a reference
8080 counter, some don't.)
8081 Similarly, 'set1' and 'add1' functions increase reference
8082 counters or duplicate objects.
8085 *) Allow for the possibility of temp RSA key generation failure:
8086 the code used to assume it always worked and crashed on failure.
8089 *) Fix potential buffer overrun problem in BIO_printf().
8090 [Ulf Möller, using public domain code by Patrick Powell; problem
8091 pointed out by David Sacerdote <das33@cornell.edu>]
8093 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8094 RAND_egd() and RAND_status(). In the command line application,
8095 the EGD socket can be specified like a seed file using RANDFILE
8099 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8100 Some CAs (e.g. Verisign) distribute certificates in this form.
8103 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8104 list to exclude them. This means that no special compilation option
8105 is needed to use anonymous DH: it just needs to be included in the
8109 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8110 EVP_MD_type. The old functionality is available in a new macro called
8111 EVP_MD_md(). Change code that uses it and update docs.
8114 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8115 where the 'void *' argument is replaced by a function pointer argument.
8116 Previously 'void *' was abused to point to functions, which works on
8117 many platforms, but is not correct. As these functions are usually
8118 called by macros defined in OpenSSL header files, most source code
8119 should work without changes.
8122 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8123 sections with information on -D... compiler switches used for
8124 compiling the library so that applications can see them. To enable
8125 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8126 must be defined. E.g.,
8127 #define OPENSSL_ALGORITHM_DEFINES
8128 #include <openssl/opensslconf.h>
8129 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8130 [Richard Levitte, Ulf and Bodo Möller]
8132 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8136 *) Change the 'other' type in certificate aux info to a STACK_OF
8137 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8138 the required ASN1 format: arbitrary types determined by an OID.
8141 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8142 argument to 'req'. This is not because the function is newer or
8143 better than others it just uses the work 'NEW' in the certificate
8144 request header lines. Some software needs this.
8147 *) Reorganise password command line arguments: now passwords can be
8148 obtained from various sources. Delete the PEM_cb function and make
8149 it the default behaviour: i.e. if the callback is NULL and the
8150 usrdata argument is not NULL interpret it as a null terminated pass
8151 phrase. If usrdata and the callback are NULL then the pass phrase
8152 is prompted for as usual.
8155 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8156 the support is automatically enabled. The resulting binaries will
8157 autodetect the card and use it if present.
8158 [Ben Laurie and Compaq Inc.]
8160 *) Work around for Netscape hang bug. This sends certificate request
8161 and server done in one record. Since this is perfectly legal in the
8162 SSL/TLS protocol it isn't a "bug" option and is on by default. See
8163 the bugs/SSLv3 entry for more info.
8166 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
8169 *) Add -rand argument to smime and pkcs12 applications and read/write
8173 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
8176 *) Add command line password options to the remaining applications.
8179 *) Bug fix for BN_div_recp() for numerators with an even number of
8183 *) More tests in bntest.c, and changed test_bn output.
8186 *) ./config recognizes MacOS X now.
8189 *) Bug fix for BN_div() when the first words of num and divsor are
8190 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
8193 *) Add support for various broken PKCS#8 formats, and command line
8194 options to produce them.
8197 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
8198 get temporary BIGNUMs from a BN_CTX.
8201 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
8205 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
8206 include a #define from the old name to the new. The original intent
8207 was that statically linked binaries could for example just call
8208 SSLeay_add_all_ciphers() to just add ciphers to the table and not
8209 link with digests. This never worked becayse SSLeay_add_all_digests()
8210 and SSLeay_add_all_ciphers() were in the same source file so calling
8211 one would link with the other. They are now in separate source files.
8214 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
8217 *) Use a less unusual form of the Miller-Rabin primality test (it used
8218 a binary algorithm for exponentiation integrated into the Miller-Rabin
8219 loop, our standard modexp algorithms are faster).
8222 *) Support for the EBCDIC character set completed.
8223 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
8225 *) Source code cleanups: use const where appropriate, eliminate casts,
8226 use void * instead of char * in lhash.
8229 *) Bugfix: ssl3_send_server_key_exchange was not restartable
8230 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
8231 this the server could overwrite ephemeral keys that the client
8235 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
8236 using 50 iterations of the Rabin-Miller test.
8238 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
8239 iterations of the Rabin-Miller test as required by the appendix
8240 to FIPS PUB 186[-1]) instead of DSA_is_prime.
8241 As BN_is_prime_fasttest includes trial division, DSA parameter
8242 generation becomes much faster.
8244 This implies a change for the callback functions in DSA_is_prime
8245 and DSA_generate_parameters: The callback function is called once
8246 for each positive witness in the Rabin-Miller test, not just
8247 occasionally in the inner loop; and the parameters to the
8248 callback function now provide an iteration count for the outer
8249 loop rather than for the current invocation of the inner loop.
8250 DSA_generate_parameters additionally can call the callback
8251 function with an 'iteration count' of -1, meaning that a
8252 candidate has passed the trial division test (when q is generated
8253 from an application-provided seed, trial division is skipped).
8256 *) New function BN_is_prime_fasttest that optionally does trial
8257 division before starting the Rabin-Miller test and has
8258 an additional BN_CTX * argument (whereas BN_is_prime always
8259 has to allocate at least one BN_CTX).
8260 'callback(1, -1, cb_arg)' is called when a number has passed the
8261 trial division stage.
8264 *) Fix for bug in CRL encoding. The validity dates weren't being handled
8268 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
8271 *) New function BN_pseudo_rand().
8274 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
8275 bignum version of BN_from_montgomery() with the working code from
8276 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
8280 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
8281 made it impossible to use the same SSL_SESSION data structure in
8282 SSL2 clients in multiple threads.
8285 *) The return value of RAND_load_file() no longer counts bytes obtained
8286 by stat(). RAND_load_file(..., -1) is new and uses the complete file
8287 to seed the PRNG (previously an explicit byte count was required).
8288 [Ulf Möller, Bodo Möller]
8290 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
8291 used (char *) instead of (void *) and had casts all over the place.
8294 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
8297 *) Retain source code compatibility for BN_prime_checks macro:
8298 BN_is_prime(..., BN_prime_checks, ...) now uses
8299 BN_prime_checks_for_size to determine the appropriate number of
8300 Rabin-Miller iterations.
8303 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
8304 DH_CHECK_P_NOT_SAFE_PRIME.
8305 (Check if this is true? OpenPGP calls them "strong".)
8308 *) Merge the functionality of "dh" and "gendh" programs into a new program
8309 "dhparam". The old programs are retained for now but will handle DH keys
8310 (instead of parameters) in future.
8313 *) Make the ciphers, s_server and s_client programs check the return values
8314 when a new cipher list is set.
8317 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
8318 ciphers. Before when the 56bit ciphers were enabled the sorting was
8321 The syntax for the cipher sorting has been extended to support sorting by
8322 cipher-strength (using the strength_bits hard coded in the tables).
8323 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
8325 Fix a bug in the cipher-command parser: when supplying a cipher command
8326 string with an "undefined" symbol (neither command nor alphanumeric
8327 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
8328 an error is flagged.
8330 Due to the strength-sorting extension, the code of the
8331 ssl_create_cipher_list() function was completely rearranged. I hope that
8332 the readability was also increased :-)
8333 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8335 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
8336 for the first serial number and places 2 in the serial number file. This
8337 avoids problems when the root CA is created with serial number zero and
8338 the first user certificate has the same issuer name and serial number
8342 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
8343 the new code. Add documentation for this stuff.
8346 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
8347 X509_*() to X509at_*() on the grounds that they don't handle X509
8348 structures and behave in an analagous way to the X509v3 functions:
8349 they shouldn't be called directly but wrapper functions should be used
8352 So we also now have some wrapper functions that call the X509at functions
8353 when passed certificate requests. (TO DO: similar things can be done with
8354 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
8355 things. Some of these need some d2i or i2d and print functionality
8356 because they handle more complex structures.)
8359 *) Add missing #ifndefs that caused missing symbols when building libssl
8360 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
8361 NO_RSA in ssl/s2*.c.
8362 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
8364 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
8365 has a return value which indicates the quality of the random data
8366 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
8367 error queue. New function RAND_pseudo_bytes() generates output that is
8368 guaranteed to be unique but not unpredictable. RAND_add is like
8369 RAND_seed, but takes an extra argument for an entropy estimate
8370 (RAND_seed always assumes full entropy).
8373 *) Do more iterations of Rabin-Miller probable prime test (specifically,
8374 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
8375 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
8376 in crypto/bn/bn_prime.c for the complete table). This guarantees a
8377 false-positive rate of at most 2^-80 for random input.
8380 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
8383 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
8384 in the 0.9.5 release), this returns the chain
8385 from an X509_CTX structure with a dup of the stack and all
8386 the X509 reference counts upped: so the stack will exist
8387 after X509_CTX_cleanup() has been called. Modify pkcs12.c
8390 Also make SSL_SESSION_print() print out the verify return
8394 *) Add manpage for the pkcs12 command. Also change the default
8395 behaviour so MAC iteration counts are used unless the new
8396 -nomaciter option is used. This improves file security and
8397 only older versions of MSIE (4.0 for example) need it.
8400 *) Honor the no-xxx Configure options when creating .DEF files.
8403 *) Add PKCS#10 attributes to field table: challengePassword,
8404 unstructuredName and unstructuredAddress. These are taken from
8405 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
8406 international characters are used.
8408 More changes to X509_ATTRIBUTE code: allow the setting of types
8409 based on strings. Remove the 'loc' parameter when adding
8410 attributes because these will be a SET OF encoding which is sorted
8414 *) Initial changes to the 'req' utility to allow request generation
8415 automation. This will allow an application to just generate a template
8416 file containing all the field values and have req construct the
8419 Initial support for X509_ATTRIBUTE handling. Stacks of these are
8420 used all over the place including certificate requests and PKCS#7
8421 structures. They are currently handled manually where necessary with
8422 some primitive wrappers for PKCS#7. The new functions behave in a
8423 manner analogous to the X509 extension functions: they allow
8424 attributes to be looked up by NID and added.
8426 Later something similar to the X509V3 code would be desirable to
8427 automatically handle the encoding, decoding and printing of the
8428 more complex types. The string types like challengePassword can
8429 be handled by the string table functions.
8431 Also modified the multi byte string table handling. Now there is
8432 a 'global mask' which masks out certain types. The table itself
8433 can use the flag STABLE_NO_MASK to ignore the mask setting: this
8434 is useful when for example there is only one permissible type
8435 (as in countryName) and using the mask might result in no valid
8439 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
8440 SSL_get_peer_finished to allow applications to obtain the latest
8441 Finished messages sent to the peer or expected from the peer,
8442 respectively. (SSL_get_peer_finished is usually the Finished message
8443 actually received from the peer, otherwise the protocol will be aborted.)
8445 As the Finished message are message digests of the complete handshake
8446 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
8447 be used for external authentication procedures when the authentication
8448 provided by SSL/TLS is not desired or is not enough.
8451 *) Enhanced support for Alpha Linux is added. Now ./config checks if
8452 the host supports BWX extension and if Compaq C is present on the
8453 $PATH. Just exploiting of the BWX extension results in 20-30%
8454 performance kick for some algorithms, e.g. DES and RC4 to mention
8455 a couple. Compaq C in turn generates ~20% faster code for MD5 and
8459 *) Add support for MS "fast SGC". This is arguably a violation of the
8460 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
8461 weak crypto and after checking the certificate is SGC a second one
8462 with strong crypto. MS SGC stops the first handshake after receiving
8463 the server certificate message and sends a second client hello. Since
8464 a server will typically do all the time consuming operations before
8465 expecting any further messages from the client (server key exchange
8466 is the most expensive) there is little difference between the two.
8468 To get OpenSSL to support MS SGC we have to permit a second client
8469 hello message after we have sent server done. In addition we have to
8470 reset the MAC if we do get this second client hello.
8473 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
8474 if a DER encoded private key is RSA or DSA traditional format. Changed
8475 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
8476 format DER encoded private key. Newer code should use PKCS#8 format which
8477 has the key type encoded in the ASN1 structure. Added DER private key
8478 support to pkcs8 application.
8481 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8482 ciphersuites has been selected (as required by the SSL 3/TLS 1
8483 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8484 is set, we interpret this as a request to violate the specification
8485 (the worst that can happen is a handshake failure, and 'correct'
8486 behaviour would result in a handshake failure anyway).
8489 *) In SSL_CTX_add_session, take into account that there might be multiple
8490 SSL_SESSION structures with the same session ID (e.g. when two threads
8491 concurrently obtain them from an external cache).
8492 The internal cache can handle only one SSL_SESSION with a given ID,
8493 so if there's a conflict, we now throw out the old one to achieve
8497 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8498 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8499 some routines that use cipher OIDs: some ciphers do not have OIDs
8500 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8504 *) Simplify the trust setting structure and code. Now we just have
8505 two sequences of OIDs for trusted and rejected settings. These will
8506 typically have values the same as the extended key usage extension
8507 and any application specific purposes.
8509 The trust checking code now has a default behaviour: it will just
8510 check for an object with the same NID as the passed id. Functions can
8511 be provided to override either the default behaviour or the behaviour
8512 for a given id. SSL client, server and email already have functions
8513 in place for compatibility: they check the NID and also return "trusted"
8514 if the certificate is self signed.
8517 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8518 traditional format into an EVP_PKEY structure.
8521 *) Add a password callback function PEM_cb() which either prompts for
8522 a password if usr_data is NULL or otherwise assumes it is a null
8523 terminated password. Allow passwords to be passed on command line
8524 environment or config files in a few more utilities.
8527 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8528 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8529 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8530 Update documentation.
8533 *) Support for ASN1 "NULL" type. This could be handled before by using
8534 ASN1_TYPE but there wasn't any function that would try to read a NULL
8535 and produce an error if it couldn't. For compatibility we also have
8536 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8537 don't allocate anything because they don't need to.
8540 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8542 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8544 *) Rebuild of the memory allocation routines used by OpenSSL code and
8545 possibly others as well. The purpose is to make an interface that
8546 provide hooks so anyone can build a separate set of allocation and
8547 deallocation routines to be used by OpenSSL, for example memory
8548 pool implementations, or something else, which was previously hard
8549 since Malloc(), Realloc() and Free() were defined as macros having
8550 the values malloc, realloc and free, respectively (except for Win32
8551 compilations). The same is provided for memory debugging code.
8552 OpenSSL already comes with functionality to find memory leaks, but
8553 this gives people a chance to debug other memory problems.
8555 With these changes, a new set of functions and macros have appeared:
8557 CRYPTO_set_mem_debug_functions() [F]
8558 CRYPTO_get_mem_debug_functions() [F]
8559 CRYPTO_dbg_set_options() [F]
8560 CRYPTO_dbg_get_options() [F]
8561 CRYPTO_malloc_debug_init() [M]
8563 The memory debug functions are NULL by default, unless the library
8564 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8565 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8566 gives the standard debugging functions that come with OpenSSL) or
8567 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8568 provided by the library user) must be used. When the standard
8569 debugging functions are used, CRYPTO_dbg_set_options can be used to
8570 request additional information:
8571 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8572 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8574 Also, things like CRYPTO_set_mem_functions will always give the
8575 expected result (the new set of functions is used for allocation
8576 and deallocation) at all times, regardless of platform and compiler
8579 To finish it up, some functions that were never use in any other
8580 way than through macros have a new API and new semantic:
8583 CRYPTO_dbg_realloc()
8586 All macros of value have retained their old syntax.
8587 [Richard Levitte and Bodo Moeller]
8589 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8590 ordering of SMIMECapabilities wasn't in "strength order" and there
8591 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8595 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8596 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8597 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
8599 *) Merge in my S/MIME library for OpenSSL. This provides a simple
8600 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
8601 functionality to handle multipart/signed properly) and a utility
8602 called 'smime' to call all this stuff. This is based on code I
8603 originally wrote for Celo who have kindly allowed it to be
8604 included in OpenSSL.
8607 *) Add variants des_set_key_checked and des_set_key_unchecked of
8608 des_set_key (aka des_key_sched). Global variable des_check_key
8609 decides which of these is called by des_set_key; this way
8610 des_check_key behaves as it always did, but applications and
8611 the library itself, which was buggy for des_check_key == 1,
8612 have a cleaner way to pick the version they need.
8615 *) New function PKCS12_newpass() which changes the password of a
8619 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
8620 dynamic mix. In both cases the ids can be used as an index into the
8621 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
8622 functions so they accept a list of the field values and the
8623 application doesn't need to directly manipulate the X509_TRUST
8627 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
8631 *) Modify the way the V3 extension code looks up extensions. This now
8632 works in a similar way to the object code: we have some "standard"
8633 extensions in a static table which is searched with OBJ_bsearch()
8634 and the application can add dynamic ones if needed. The file
8635 crypto/x509v3/ext_dat.h now has the info: this file needs to be
8636 updated whenever a new extension is added to the core code and kept
8637 in ext_nid order. There is a simple program 'tabtest.c' which checks
8638 this. New extensions are not added too often so this file can readily
8639 be maintained manually.
8641 There are two big advantages in doing things this way. The extensions
8642 can be looked up immediately and no longer need to be "added" using
8643 X509V3_add_standard_extensions(): this function now does nothing.
8644 [Side note: I get *lots* of email saying the extension code doesn't
8645 work because people forget to call this function]
8646 Also no dynamic allocation is done unless new extensions are added:
8647 so if we don't add custom extensions there is no need to call
8648 X509V3_EXT_cleanup().
8651 *) Modify enc utility's salting as follows: make salting the default. Add a
8652 magic header, so unsalted files fail gracefully instead of just decrypting
8653 to garbage. This is because not salting is a big security hole, so people
8654 should be discouraged from doing it.
8657 *) Fixes and enhancements to the 'x509' utility. It allowed a message
8658 digest to be passed on the command line but it only used this
8659 parameter when signing a certificate. Modified so all relevant
8660 operations are affected by the digest parameter including the
8661 -fingerprint and -x509toreq options. Also -x509toreq choked if a
8662 DSA key was used because it didn't fix the digest.
8665 *) Initial certificate chain verify code. Currently tests the untrusted
8666 certificates for consistency with the verify purpose (which is set
8667 when the X509_STORE_CTX structure is set up) and checks the pathlength.
8669 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
8670 this is because it will reject chains with invalid extensions whereas
8671 every previous version of OpenSSL and SSLeay made no checks at all.
8673 Trust code: checks the root CA for the relevant trust settings. Trust
8674 settings have an initial value consistent with the verify purpose: e.g.
8675 if the verify purpose is for SSL client use it expects the CA to be
8676 trusted for SSL client use. However the default value can be changed to
8677 permit custom trust settings: one example of this would be to only trust
8678 certificates from a specific "secure" set of CAs.
8680 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
8681 which should be used for version portability: especially since the
8682 verify structure is likely to change more often now.
8684 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
8685 to set them. If not set then assume SSL clients will verify SSL servers
8688 Two new options to the verify program: -untrusted allows a set of
8689 untrusted certificates to be passed in and -purpose which sets the
8690 intended purpose of the certificate. If a purpose is set then the
8691 new chain verify code is used to check extension consistency.
8694 *) Support for the authority information access extension.
8697 *) Modify RSA and DSA PEM read routines to transparently handle
8698 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
8699 public keys in a format compatible with certificate
8700 SubjectPublicKeyInfo structures. Unfortunately there were already
8701 functions called *_PublicKey_* which used various odd formats so
8702 these are retained for compatibility: however the DSA variants were
8703 never in a public release so they have been deleted. Changed dsa/rsa
8704 utilities to handle the new format: note no releases ever handled public
8705 keys so we should be OK.
8707 The primary motivation for this change is to avoid the same fiasco
8708 that dogs private keys: there are several incompatible private key
8709 formats some of which are standard and some OpenSSL specific and
8710 require various evil hacks to allow partial transparent handling and
8711 even then it doesn't work with DER formats. Given the option anything
8712 other than PKCS#8 should be dumped: but the other formats have to
8713 stay in the name of compatibility.
8715 With public keys and the benefit of hindsight one standard format
8716 is used which works with EVP_PKEY, RSA or DSA structures: though
8717 it clearly returns an error if you try to read the wrong kind of key.
8719 Added a -pubkey option to the 'x509' utility to output the public key.
8720 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
8721 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
8722 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
8723 that do the same as the EVP_PKEY_assign_*() except they up the
8724 reference count of the added key (they don't "swallow" the
8728 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
8729 CRLs would fail if the file contained no certificates or no CRLs:
8730 added a new function to read in both types and return the number
8731 read: this means that if none are read it will be an error. The
8732 DER versions of the certificate and CRL reader would always fail
8733 because it isn't possible to mix certificates and CRLs in DER format
8734 without choking one or the other routine. Changed this to just read
8735 a certificate: this is the best we can do. Also modified the code
8736 in apps/verify.c to take notice of return codes: it was previously
8737 attempting to read in certificates from NULL pointers and ignoring
8738 any errors: this is one reason why the cert and CRL reader seemed
8739 to work. It doesn't check return codes from the default certificate
8740 routines: these may well fail if the certificates aren't installed.
8743 *) Code to support otherName option in GeneralName.
8746 *) First update to verify code. Change the verify utility
8747 so it warns if it is passed a self signed certificate:
8748 for consistency with the normal behaviour. X509_verify
8749 has been modified to it will now verify a self signed
8750 certificate if *exactly* the same certificate appears
8751 in the store: it was previously impossible to trust a
8752 single self signed certificate. This means that:
8753 openssl verify ss.pem
8754 now gives a warning about a self signed certificate but
8755 openssl verify -CAfile ss.pem ss.pem
8759 *) For servers, store verify_result in SSL_SESSION data structure
8760 (and add it to external session representation).
8761 This is needed when client certificate verifications fails,
8762 but an application-provided verification callback (set by
8763 SSL_CTX_set_cert_verify_callback) allows accepting the session
8764 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
8765 but returns 1): When the session is reused, we have to set
8766 ssl->verify_result to the appropriate error code to avoid
8768 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
8770 *) Fix a bug in the new PKCS#7 code: it didn't consider the
8771 case in PKCS7_dataInit() where the signed PKCS7 structure
8772 didn't contain any existing data because it was being created.
8773 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
8775 *) Add a salt to the key derivation routines in enc.c. This
8776 forms the first 8 bytes of the encrypted file. Also add a
8777 -S option to allow a salt to be input on the command line.
8780 *) New function X509_cmp(). Oddly enough there wasn't a function
8781 to compare two certificates. We do this by working out the SHA1
8782 hash and comparing that. X509_cmp() will be needed by the trust
8786 *) SSL_get1_session() is like SSL_get_session(), but increments
8787 the reference count in the SSL_SESSION returned.
8788 [Geoff Thorpe <geoff@eu.c2.net>]
8790 *) Fix for 'req': it was adding a null to request attributes.
8791 Also change the X509_LOOKUP and X509_INFO code to handle
8792 certificate auxiliary information.
8795 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8799 *) Add the possibility to add extra information to the memory leak
8800 detecting output, to form tracebacks, showing from where each
8801 allocation was originated: CRYPTO_push_info("constant string") adds
8802 the string plus current file name and line number to a per-thread
8803 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8804 is like calling CYRPTO_pop_info() until the stack is empty.
8805 Also updated memory leak detection code to be multi-thread-safe.
8808 *) Add options -text and -noout to pkcs7 utility and delete the
8809 encryption options which never did anything. Update docs.
8812 *) Add options to some of the utilities to allow the pass phrase
8813 to be included on either the command line (not recommended on
8814 OSes like Unix) or read from the environment. Update the
8815 manpages and fix a few bugs.
8818 *) Add a few manpages for some of the openssl commands.
8821 *) Fix the -revoke option in ca. It was freeing up memory twice,
8822 leaking and not finding already revoked certificates.
8825 *) Extensive changes to support certificate auxiliary information.
8826 This involves the use of X509_CERT_AUX structure and X509_AUX
8827 functions. An X509_AUX function such as PEM_read_X509_AUX()
8828 can still read in a certificate file in the usual way but it
8829 will also read in any additional "auxiliary information". By
8830 doing things this way a fair degree of compatibility can be
8831 retained: existing certificates can have this information added
8832 using the new 'x509' options.
8834 Current auxiliary information includes an "alias" and some trust
8835 settings. The trust settings will ultimately be used in enhanced
8836 certificate chain verification routines: currently a certificate
8837 can only be trusted if it is self signed and then it is trusted
8841 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8842 The problem was that one of the replacement routines had not been working
8843 since SSLeay releases. For now the offending routine has been replaced
8844 with non-optimised assembler. Even so, this now gives around 95%
8845 performance improvement for 1024 bit RSA signs.
8848 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8849 handling. Most clients have the effective key size in bits equal to
8850 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8851 A few however don't do this and instead use the size of the decrypted key
8852 to determine the RC2 key length and the AlgorithmIdentifier to determine
8853 the effective key length. In this case the effective key length can still
8854 be 40 bits but the key length can be 168 bits for example. This is fixed
8855 by manually forcing an RC2 key into the EVP_PKEY structure because the
8856 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8857 the key length and effective key length are equal.
8860 *) Add a bunch of functions that should simplify the creation of
8861 X509_NAME structures. Now you should be able to do:
8862 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8863 and have it automatically work out the correct field type and fill in
8864 the structures. The more adventurous can try:
8865 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8866 and it will (hopefully) work out the correct multibyte encoding.
8869 *) Change the 'req' utility to use the new field handling and multibyte
8870 copy routines. Before the DN field creation was handled in an ad hoc
8871 way in req, ca, and x509 which was rather broken and didn't support
8872 BMPStrings or UTF8Strings. Since some software doesn't implement
8873 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8874 using the dirstring_type option. See the new comment in the default
8875 openssl.cnf for more info.
8878 *) Make crypto/rand/md_rand.c more robust:
8879 - Assure unique random numbers after fork().
8880 - Make sure that concurrent threads access the global counter and
8881 md serializably so that we never lose entropy in them
8882 or use exactly the same state in multiple threads.
8883 Access to the large state is not always serializable because
8884 the additional locking could be a performance killer, and
8885 md should be large enough anyway.
8888 *) New file apps/app_rand.c with commonly needed functionality
8889 for handling the random seed file.
8891 Use the random seed file in some applications that previously did not:
8893 dsaparam -genkey (which also ignored its '-rand' option),
8896 x509 (when signing).
8897 Except on systems with /dev/urandom, it is crucial to have a random
8898 seed file at least for key creation, DSA signing, and for DH exchanges;
8899 for RSA signatures we could do without one.
8901 gendh and gendsa (unlike genrsa) used to read only the first byte
8902 of each file listed in the '-rand' option. The function as previously
8903 found in genrsa is now in app_rand.c and is used by all programs
8904 that support '-rand'.
8907 *) In RAND_write_file, use mode 0600 for creating files;
8908 don't just chmod when it may be too late.
8911 *) Report an error from X509_STORE_load_locations
8912 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8915 *) New function ASN1_mbstring_copy() this copies a string in either
8916 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8917 into an ASN1_STRING type. A mask of permissible types is passed
8918 and it chooses the "minimal" type to use or an error if not type
8922 *) Add function equivalents to the various macros in asn1.h. The old
8923 macros are retained with an M_ prefix. Code inside the library can
8924 use the M_ macros. External code (including the openssl utility)
8925 should *NOT* in order to be "shared library friendly".
8928 *) Add various functions that can check a certificate's extensions
8929 to see if it usable for various purposes such as SSL client,
8930 server or S/MIME and CAs of these types. This is currently
8931 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8932 verification. Also added a -purpose flag to x509 utility to
8933 print out all the purposes.
8936 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8940 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8941 for, obtain and decode and extension and obtain its critical flag.
8942 This allows all the necessary extension code to be handled in a
8943 single function call.
8946 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8947 platforms. See crypto/rc4/rc4_enc.c for further details.
8950 *) New -noout option to asn1parse. This causes no output to be produced
8951 its main use is when combined with -strparse and -out to extract data
8952 from a file (which may not be in ASN.1 format).
8955 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8956 when producing the local key id.
8957 [Richard Levitte <levitte@stacken.kth.se>]
8959 *) New option -dhparam in s_server. This allows a DH parameter file to be
8960 stated explicitly. If it is not stated then it tries the first server
8961 certificate file. The previous behaviour hard coded the filename
8965 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8966 a public key to be input or output. For example:
8967 openssl rsa -in key.pem -pubout -out pubkey.pem
8968 Also added necessary DSA public key functions to handle this.
8971 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8972 in the message. This was handled by allowing
8973 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8974 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8976 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8977 to the end of the strings whereas this didn't. This would cause problems
8978 if strings read with d2i_ASN1_bytes() were later modified.
8979 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8981 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8982 data and it contains EOF it will end up returning an error. This is
8983 caused by input 46 bytes long. The cause is due to the way base64
8984 BIOs find the start of base64 encoded data. They do this by trying a
8985 trial decode on each line until they find one that works. When they
8986 do a flag is set and it starts again knowing it can pass all the
8987 data directly through the decoder. Unfortunately it doesn't reset
8988 the context it uses. This means that if EOF is reached an attempt
8989 is made to pass two EOFs through the context and this causes the
8990 resulting error. This can also cause other problems as well. As is
8991 usual with these problems it takes *ages* to find and the fix is
8992 trivial: move one line.
8993 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8995 *) Ugly workaround to get s_client and s_server working under Windows. The
8996 old code wouldn't work because it needed to select() on sockets and the
8997 tty (for keypresses and to see if data could be written). Win32 only
8998 supports select() on sockets so we select() with a 1s timeout on the
8999 sockets and then see if any characters are waiting to be read, if none
9000 are present then we retry, we also assume we can always write data to
9001 the tty. This isn't nice because the code then blocks until we've
9002 received a complete line of data and it is effectively polling the
9003 keyboard at 1s intervals: however it's quite a bit better than not
9004 working at all :-) A dedicated Windows application might handle this
9005 with an event loop for example.
9008 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9009 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9010 will be called when RSA_sign() and RSA_verify() are used. This is useful
9011 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9012 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9013 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9014 This necessitated the support of an extra signature type NID_md5_sha1
9015 for SSL signatures and modifications to the SSL library to use it instead
9016 of calling RSA_public_decrypt() and RSA_private_encrypt().
9019 *) Add new -verify -CAfile and -CApath options to the crl program, these
9020 will lookup a CRL issuers certificate and verify the signature in a
9021 similar way to the verify program. Tidy up the crl program so it
9022 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9023 less strict. It will now permit CRL extensions even if it is not
9024 a V2 CRL: this will allow it to tolerate some broken CRLs.
9027 *) Initialize all non-automatic variables each time one of the openssl
9028 sub-programs is started (this is necessary as they may be started
9029 multiple times from the "OpenSSL>" prompt).
9030 [Lennart Bang, Bodo Moeller]
9032 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9033 removing all other RSA functionality (this is what NO_RSA does). This
9034 is so (for example) those in the US can disable those operations covered
9035 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9039 *) Non-copying interface to BIO pairs.
9040 (still largely untested)
9043 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9044 ASCII string. This was handled independently in various places before.
9047 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9048 UTF8 strings a character at a time.
9051 *) Use client_version from client hello to select the protocol
9052 (s23_srvr.c) and for RSA client key exchange verification
9053 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9056 *) Add various utility functions to handle SPKACs, these were previously
9057 handled by poking round in the structure internals. Added new function
9058 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9059 print, verify and generate SPKACs. Based on an original idea from
9060 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9063 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9066 *) Allow the config file extension section to be overwritten on the
9067 command line. Based on an original idea from Massimiliano Pala
9068 <madwolf@comune.modena.it>. The new option is called -extensions
9069 and can be applied to ca, req and x509. Also -reqexts to override
9070 the request extensions in req and -crlexts to override the crl extensions
9074 *) Add new feature to the SPKAC handling in ca. Now you can include
9075 the same field multiple times by preceding it by "XXXX." for example:
9078 this is the same syntax as used in the req config file.
9081 *) Allow certificate extensions to be added to certificate requests. These
9082 are specified in a 'req_extensions' option of the req section of the
9083 config file. They can be printed out with the -text option to req but
9084 are otherwise ignored at present.
9087 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9088 data read consists of only the final block it would not decrypted because
9089 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9090 A misplaced 'break' also meant the decrypted final block might not be
9091 copied until the next read.
9094 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9095 a few extra parameters to the DH structure: these will be useful if
9096 for example we want the value of 'q' or implement X9.42 DH.
9099 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9100 provides hooks that allow the default DSA functions or functions on a
9101 "per key" basis to be replaced. This allows hardware acceleration and
9102 hardware key storage to be handled without major modification to the
9103 library. Also added low level modexp hooks and CRYPTO_EX structure and
9104 associated functions.
9107 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9108 as "read only": it can't be written to and the buffer it points to will
9109 not be freed. Reading from a read only BIO is much more efficient than
9110 a normal memory BIO. This was added because there are several times when
9111 an area of memory needs to be read from a BIO. The previous method was
9112 to create a memory BIO and write the data to it, this results in two
9113 copies of the data and an O(n^2) reading algorithm. There is a new
9114 function BIO_new_mem_buf() which creates a read only memory BIO from
9115 an area of memory. Also modified the PKCS#7 routines to use read only
9119 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9120 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9121 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9122 but a retry condition occured while trying to read the rest.
9125 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9126 NID_pkcs7_encrypted by default: this was wrong since this should almost
9127 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9128 the encrypted data type: this is a more sensible place to put it and it
9129 allows the PKCS#12 code to be tidied up that duplicated this
9133 *) Changed obj_dat.pl script so it takes its input and output files on
9134 the command line. This should avoid shell escape redirection problems
9138 *) Initial support for certificate extension requests, these are included
9139 in things like Xenroll certificate requests. Included functions to allow
9140 extensions to be obtained and added.
9143 *) -crlf option to s_client and s_server for sending newlines as
9144 CRLF (as required by many protocols).
9147 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9149 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9150 [Ralf S. Engelschall]
9152 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9153 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9155 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
9159 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
9160 DH parameters/keys (q is lost during that conversion, but the resulting
9161 DH parameters contain its length).
9163 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
9164 much faster than DH_generate_parameters (which creates parameters
9165 where p = 2*q + 1), and also the smaller q makes DH computations
9166 much more efficient (160-bit exponentiation instead of 1024-bit
9167 exponentiation); so this provides a convenient way to support DHE
9168 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
9169 utter importance to use
9170 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9172 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9173 when such DH parameters are used, because otherwise small subgroup
9174 attacks may become possible!
9177 *) Avoid memory leak in i2d_DHparams.
9180 *) Allow the -k option to be used more than once in the enc program:
9181 this allows the same encrypted message to be read by multiple recipients.
9184 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
9185 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
9186 it will always use the numerical form of the OID, even if it has a short
9190 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
9191 method only got called if p,q,dmp1,dmq1,iqmp components were present,
9192 otherwise bn_mod_exp was called. In the case of hardware keys for example
9193 no private key components need be present and it might store extra data
9194 in the RSA structure, which cannot be accessed from bn_mod_exp.
9195 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
9196 private key operations.
9199 *) Added support for SPARC Linux.
9202 *) pem_password_cb function type incompatibly changed from
9203 typedef int pem_password_cb(char *buf, int size, int rwflag);
9205 ....(char *buf, int size, int rwflag, void *userdata);
9206 so that applications can pass data to their callbacks:
9207 The PEM[_ASN1]_{read,write}... functions and macros now take an
9208 additional void * argument, which is just handed through whenever
9209 the password callback is called.
9210 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
9212 New function SSL_CTX_set_default_passwd_cb_userdata.
9214 Compatibility note: As many C implementations push function arguments
9215 onto the stack in reverse order, the new library version is likely to
9216 interoperate with programs that have been compiled with the old
9217 pem_password_cb definition (PEM_whatever takes some data that
9218 happens to be on the stack as its last argument, and the callback
9219 just ignores this garbage); but there is no guarantee whatsoever that
9222 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
9223 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
9224 problems not only on Windows, but also on some Unix platforms.
9225 To avoid problematic command lines, these definitions are now in an
9226 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
9227 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
9230 *) MIPS III/IV assembler module is reimplemented.
9233 *) More DES library cleanups: remove references to srand/rand and
9234 delete an unused file.
9237 *) Add support for the the free Netwide assembler (NASM) under Win32,
9238 since not many people have MASM (ml) and it can be hard to obtain.
9239 This is currently experimental but it seems to work OK and pass all
9240 the tests. Check out INSTALL.W32 for info.
9243 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
9244 without temporary keys kept an extra copy of the server key,
9245 and connections with temporary keys did not free everything in case
9249 *) New function RSA_check_key and new openssl rsa option -check
9250 for verifying the consistency of RSA keys.
9251 [Ulf Moeller, Bodo Moeller]
9253 *) Various changes to make Win32 compile work:
9254 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
9255 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
9256 comparison" warnings.
9257 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
9260 *) Add a debugging option to PKCS#5 v2 key generation function: when
9261 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
9262 derived keys are printed to stderr.
9265 *) Copy the flags in ASN1_STRING_dup().
9266 [Roman E. Pavlov <pre@mo.msk.ru>]
9268 *) The x509 application mishandled signing requests containing DSA
9269 keys when the signing key was also DSA and the parameters didn't match.
9271 It was supposed to omit the parameters when they matched the signing key:
9272 the verifying software was then supposed to automatically use the CA's
9273 parameters if they were absent from the end user certificate.
9275 Omitting parameters is no longer recommended. The test was also
9276 the wrong way round! This was probably due to unusual behaviour in
9277 EVP_cmp_parameters() which returns 1 if the parameters match.
9278 This meant that parameters were omitted when they *didn't* match and
9279 the certificate was useless. Certificates signed with 'ca' didn't have
9281 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
9283 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
9284 The interface is as follows:
9285 Applications can use
9286 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
9287 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
9288 "off" is now the default.
9289 The library internally uses
9290 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
9291 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
9292 to disable memory-checking temporarily.
9294 Some inconsistent states that previously were possible (and were
9295 even the default) are now avoided.
9297 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
9298 with each memory chunk allocated; this is occasionally more helpful
9299 than just having a counter.
9301 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
9303 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
9307 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
9308 which largely parallels "options", but is for changing API behaviour,
9309 whereas "options" are about protocol behaviour.
9310 Initial "mode" flags are:
9312 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
9313 a single record has been written.
9314 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
9315 retries use the same buffer location.
9316 (But all of the contents must be
9320 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
9323 *) Fix problems with no-hmac etc.
9324 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
9326 *) New functions RSA_get_default_method(), RSA_set_method() and
9327 RSA_get_method(). These allows replacement of RSA_METHODs without having
9328 to mess around with the internals of an RSA structure.
9331 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
9332 Also really enable memory leak checks in openssl.c and in some
9334 [Chad C. Mulligan, Bodo Moeller]
9336 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
9337 up the length of negative integers. This has now been simplified to just
9338 store the length when it is first determined and use it later, rather
9339 than trying to keep track of where data is copied and updating it to
9341 [Steve Henson, reported by Brien Wheeler
9342 <bwheeler@authentica-security.com>]
9344 *) Add a new function PKCS7_signatureVerify. This allows the verification
9345 of a PKCS#7 signature but with the signing certificate passed to the
9346 function itself. This contrasts with PKCS7_dataVerify which assumes the
9347 certificate is present in the PKCS#7 structure. This isn't always the
9348 case: certificates can be omitted from a PKCS#7 structure and be
9349 distributed by "out of band" means (such as a certificate database).
9352 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
9353 function prototypes in pem.h, also change util/mkdef.pl to add the
9354 necessary function names.
9357 *) mk1mf.pl (used by Windows builds) did not properly read the
9358 options set by Configure in the top level Makefile, and Configure
9359 was not even able to write more than one option correctly.
9360 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
9363 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
9364 file to be loaded from a BIO or FILE pointer. The BIO version will
9365 for example allow memory BIOs to contain config info.
9368 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
9369 Whoever hopes to achieve shared-library compatibility across versions
9370 must use this, not the compile-time macro.
9371 (Exercise 0.9.4: Which is the minimum library version required by
9373 Note: All this applies only to multi-threaded programs, others don't
9377 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
9378 through a BIO pair triggered the default case, i.e.
9379 SSLerr(...,SSL_R_UNKNOWN_STATE).
9382 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
9383 can use the SSL library even if none of the specific BIOs is
9387 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
9388 for the encoded length.
9389 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
9391 *) Add initial documentation of the X509V3 functions.
9394 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
9395 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
9396 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
9397 secure PKCS#8 private key format with a high iteration count.
9400 *) Fix determination of Perl interpreter: A perl or perl5
9401 _directory_ in $PATH was also accepted as the interpreter.
9402 [Ralf S. Engelschall]
9404 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
9405 wrong with it but it was very old and did things like calling
9406 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
9410 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
9411 to use the new extension code.
9414 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
9415 with macros. This should make it easier to change their form, add extra
9416 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
9420 *) Add to configuration table a new entry that can specify an alternative
9421 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
9422 according to Mark Crispin <MRC@Panda.COM>.
9426 *) DES CBC did not update the IV. Weird.
9429 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
9430 Changing the behaviour of the former might break existing programs --
9431 where IV updating is needed, des_ncbc_encrypt can be used.
9434 *) When bntest is run from "make test" it drives bc to check its
9435 calculations, as well as internally checking them. If an internal check
9436 fails, it needs to cause bc to give a non-zero result or make test carries
9437 on without noticing the failure. Fixed.
9440 *) DES library cleanups.
9443 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
9444 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
9445 ciphers. NOTE: although the key derivation function has been verified
9446 against some published test vectors it has not been extensively tested
9447 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
9451 *) Instead of "mkdir -p", which is not fully portable, use new
9452 Perl script "util/mkdir-p.pl".
9455 *) Rewrite the way password based encryption (PBE) is handled. It used to
9456 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
9457 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
9458 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
9459 the 'parameter' field of the AlgorithmIdentifier is passed to the
9460 underlying key generation function so it must do its own ASN1 parsing.
9461 This has also changed the EVP_PBE_CipherInit() function which now has a
9462 'parameter' argument instead of literal salt and iteration count values
9463 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
9466 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
9467 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
9468 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
9469 KEY" because this clashed with PKCS#8 unencrypted string. Since this
9470 value was just used as a "magic string" and not used directly its
9471 value doesn't matter.
9474 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9478 *) "linux-sparc64" configuration (ultrapenguin).
9479 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9480 "linux-sparc" configuration.
9481 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9483 *) config now generates no-xxx options for missing ciphers.
9486 *) Support the EBCDIC character set (work in progress).
9487 File ebcdic.c not yet included because it has a different license.
9488 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9490 *) Support BS2000/OSD-POSIX.
9491 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9493 *) Make callbacks for key generation use void * instead of char *.
9496 *) Make S/MIME samples compile (not yet tested).
9499 *) Additional typesafe stacks.
9502 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9506 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9508 *) New configuration variant "sco5-gcc".
9510 *) Updated some demos.
9511 [Sean O Riordain, Wade Scholine]
9513 *) Add missing BIO_free at exit of pkcs12 application.
9516 *) Fix memory leak in conf.c.
9519 *) Updates for Win32 to assembler version of MD5.
9522 *) Set #! path to perl in apps/der_chop to where we found it
9523 instead of using a fixed path.
9526 *) SHA library changes for irix64-mips4-cc.
9529 *) Improvements for VMS support.
9533 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9535 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9536 This also avoids the problems with SC4.2 and unpatched SC5.
9537 [Andy Polyakov <appro@fy.chalmers.se>]
9539 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9540 These are required because of the typesafe stack would otherwise break
9541 existing code. If old code used a structure member which used to be STACK
9542 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9543 sk_num or sk_value it would produce an error because the num, data members
9544 are not present in STACK_OF. Now it just produces a warning. sk_set
9545 replaces the old method of assigning a value to sk_value
9546 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9547 that does this will no longer work (and should use sk_set instead) but
9548 this could be regarded as a "questionable" behaviour anyway.
9551 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9552 correctly handle encrypted S/MIME data.
9555 *) Change type of various DES function arguments from des_cblock
9556 (which means, in function argument declarations, pointer to char)
9557 to des_cblock * (meaning pointer to array with 8 char elements),
9558 which allows the compiler to do more typechecking; it was like
9559 that back in SSLeay, but with lots of ugly casts.
9561 Introduce new type const_des_cblock.
9564 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9565 problems: find RecipientInfo structure that matches recipient certificate
9566 and initialise the ASN1 structures properly based on passed cipher.
9569 *) Belatedly make the BN tests actually check the results.
9572 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9573 to and from BNs: it was completely broken. New compilation option
9574 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9575 key elements as negative integers.
9578 *) Reorganize and speed up MD5.
9579 [Andy Polyakov <appro@fy.chalmers.se>]
9582 [Richard Levitte <richard@levitte.org>]
9584 *) New option -out to asn1parse to allow the parsed structure to be
9585 output to a file. This is most useful when combined with the -strparse
9586 option to examine the output of things like OCTET STRINGS.
9589 *) Make SSL library a little more fool-proof by not requiring any longer
9590 that SSL_set_{accept,connect}_state be called before
9591 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9592 in many applications because usually everything *appeared* to work as
9593 intended anyway -- now it really works as intended).
9596 *) Move openssl.cnf out of lib/.
9599 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
9600 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
9601 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
9602 [Ralf S. Engelschall]
9604 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
9605 handle PKCS#7 enveloped data properly.
9606 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
9608 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
9609 copying pointers. The cert_st handling is changed by this in
9610 various ways (and thus what used to be known as ctx->default_cert
9611 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
9612 any longer when s->cert does not give us what we need).
9613 ssl_cert_instantiate becomes obsolete by this change.
9614 As soon as we've got the new code right (possibly it already is?),
9615 we have solved a couple of bugs of the earlier code where s->cert
9616 was used as if it could not have been shared with other SSL structures.
9618 Note that using the SSL API in certain dirty ways now will result
9619 in different behaviour than observed with earlier library versions:
9620 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
9621 does not influence s as it used to.
9623 In order to clean up things more thoroughly, inside SSL_SESSION
9624 we don't use CERT any longer, but a new structure SESS_CERT
9625 that holds per-session data (if available); currently, this is
9626 the peer's certificate chain and, for clients, the server's certificate
9627 and temporary key. CERT holds only those values that can have
9628 meaningful defaults in an SSL_CTX.
9631 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
9632 from the internal representation. Various PKCS#7 fixes: remove some
9633 evil casts and set the enc_dig_alg field properly based on the signing
9637 *) Allow PKCS#12 password to be set from the command line or the
9638 environment. Let 'ca' get its config file name from the environment
9639 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
9643 *) Allow certificate policies extension to use an IA5STRING for the
9644 organization field. This is contrary to the PKIX definition but
9645 VeriSign uses it and IE5 only recognises this form. Document 'x509'
9649 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
9650 without disallowing inline assembler and the like for non-pedantic builds.
9653 *) Support Borland C++ builder.
9654 [Janez Jere <jj@void.si>, modified by Ulf Möller]
9659 *) SHA-1 cleanups and performance enhancements.
9660 [Andy Polyakov <appro@fy.chalmers.se>]
9662 *) Sparc v8plus assembler for the bignum library.
9663 [Andy Polyakov <appro@fy.chalmers.se>]
9665 *) Accept any -xxx and +xxx compiler options in Configure.
9668 *) Update HPUX configuration.
9671 *) Add missing sk_<type>_unshift() function to safestack.h
9672 [Ralf S. Engelschall]
9674 *) New function SSL_CTX_use_certificate_chain_file that sets the
9675 "extra_cert"s in addition to the certificate. (This makes sense
9676 only for "PEM" format files, as chains as a whole are not
9680 *) Support verify_depth from the SSL API.
9681 x509_vfy.c had what can be considered an off-by-one-error:
9682 Its depth (which was not part of the external interface)
9683 was actually counting the number of certificates in a chain;
9684 now it really counts the depth.
9687 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
9688 instead of X509err, which often resulted in confusing error
9689 messages since the error codes are not globally unique
9690 (e.g. an alleged error in ssl3_accept when a certificate
9691 didn't match the private key).
9693 *) New function SSL_CTX_set_session_id_context that allows to set a default
9694 value (so that you don't need SSL_set_session_id_context for each
9695 connection using the SSL_CTX).
9698 *) OAEP decoding bug fix.
9701 *) Support INSTALL_PREFIX for package builders, as proposed by
9705 *) New Configure options "threads" and "no-threads". For systems
9706 where the proper compiler options are known (currently Solaris
9707 and Linux), "threads" is the default.
9710 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
9713 *) Install various scripts to $(OPENSSLDIR)/misc, not to
9714 $(INSTALLTOP)/bin -- they shouldn't clutter directories
9715 such as /usr/local/bin.
9718 *) "make linux-shared" to build shared libraries.
9719 [Niels Poppe <niels@netbox.org>]
9721 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
9724 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
9725 extension adding in x509 utility.
9728 *) Remove NOPROTO sections and error code comments.
9731 *) Partial rewrite of the DEF file generator to now parse the ANSI
9735 *) New Configure options --prefix=DIR and --openssldir=DIR.
9738 *) Complete rewrite of the error code script(s). It is all now handled
9739 by one script at the top level which handles error code gathering,
9740 header rewriting and C source file generation. It should be much better
9741 than the old method: it now uses a modified version of Ulf's parser to
9742 read the ANSI prototypes in all header files (thus the old K&R definitions
9743 aren't needed for error creation any more) and do a better job of
9744 translating function codes into names. The old 'ASN1 error code imbedded
9745 in a comment' is no longer necessary and it doesn't use .err files which
9746 have now been deleted. Also the error code call doesn't have to appear all
9747 on one line (which resulted in some large lines...).
9750 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
9753 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
9754 0 (which usually indicates a closed connection), but continue reading.
9757 *) Fix some race conditions.
9760 *) Add support for CRL distribution points extension. Add Certificate
9761 Policies and CRL distribution points documentation.
9764 *) Move the autogenerated header file parts to crypto/opensslconf.h.
9767 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
9768 8 of keying material. Merlin has also confirmed interop with this fix
9769 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
9770 [Merlin Hughes <merlin@baltimore.ie>]
9772 *) Fix lots of warnings.
9773 [Richard Levitte <levitte@stacken.kth.se>]
9775 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
9776 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9777 [Richard Levitte <levitte@stacken.kth.se>]
9779 *) Fix problems with sizeof(long) == 8.
9780 [Andy Polyakov <appro@fy.chalmers.se>]
9782 *) Change functions to ANSI C.
9785 *) Fix typos in error codes.
9786 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9788 *) Remove defunct assembler files from Configure.
9791 *) SPARC v8 assembler BIGNUM implementation.
9792 [Andy Polyakov <appro@fy.chalmers.se>]
9794 *) Support for Certificate Policies extension: both print and set.
9795 Various additions to support the r2i method this uses.
9798 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9799 return a const string when you are expecting an allocated buffer.
9802 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9803 types DirectoryString and DisplayText.
9806 *) Add code to allow r2i extensions to access the configuration database,
9807 add an LHASH database driver and add several ctx helper functions.
9810 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9811 fail when they extended the size of a BIGNUM.
9814 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9815 support typesafe stack.
9818 *) Fix typo in SSL_[gs]et_options().
9819 [Nils Frostberg <nils@medcom.se>]
9821 *) Delete various functions and files that belonged to the (now obsolete)
9822 old X509V3 handling code.
9825 *) New Configure option "rsaref".
9828 *) Don't auto-generate pem.h.
9831 *) Introduce type-safe ASN.1 SETs.
9834 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9835 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9837 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9838 that links with OpenSSL (well at least cause lots of warnings), but fear
9839 not: the conversion is trivial, and it eliminates loads of evil casts. A
9840 few STACKed things have been converted already. Feel free to convert more.
9841 In the fullness of time, I'll do away with the STACK type altogether.
9844 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9845 specified in <certfile> by updating the entry in the index.txt file.
9846 This way one no longer has to edit the index.txt file manually for
9847 revoking a certificate. The -revoke option does the gory details now.
9848 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9850 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9851 `-text' option at all and this way the `-noout -text' combination was
9852 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9853 [Ralf S. Engelschall]
9855 *) Make sure a corresponding plain text error message exists for the
9856 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9857 verify callback function determined that a certificate was revoked.
9858 [Ralf S. Engelschall]
9860 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9861 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9862 all available cipers including rc5, which was forgotten until now.
9863 In order to let the testing shell script know which algorithms
9864 are available, a new (up to now undocumented) command
9865 "openssl list-cipher-commands" is used.
9868 *) Bugfix: s_client occasionally would sleep in select() when
9869 it should have checked SSL_pending() first.
9872 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9873 the raw DSA values prior to ASN.1 encoding.
9876 *) Tweaks to Configure
9877 [Niels Poppe <niels@netbox.org>]
9879 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9883 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9886 *) New config option to avoid instructions that are illegal on the 80386.
9887 The default code is faster, but requires at least a 486.
9890 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9891 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9892 same as SSL2_VERSION anyway.
9895 *) New "-showcerts" option for s_client.
9898 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9899 application. Various cleanups and fixes.
9902 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9903 modify error routines to work internally. Add error codes and PBE init
9904 to library startup routines.
9907 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9908 packing functions to asn1 and evp. Changed function names and error
9909 codes along the way.
9912 *) PKCS12 integration: and so it begins... First of several patches to
9913 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9914 objects to objects.h
9917 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9918 and display support for Thawte strong extranet extension.
9921 *) Add LinuxPPC support.
9922 [Jeff Dubrule <igor@pobox.org>]
9924 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9925 bn_div_words in alpha.s.
9926 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9928 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9929 OAEP isn't supported when OpenSSL is built with RSAref.
9930 [Ulf Moeller <ulf@fitug.de>]
9932 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9933 so they no longer are missing under -DNOPROTO.
9934 [Soren S. Jorvang <soren@t.dk>]
9937 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9939 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9940 doesn't work when the session is reused. Coming soon!
9943 *) Fix a security hole, that allows sessions to be reused in the wrong
9944 context thus bypassing client cert protection! All software that uses
9945 client certs and session caches in multiple contexts NEEDS PATCHING to
9946 allow session reuse! A fuller solution is in the works.
9947 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9949 *) Some more source tree cleanups (removed obsolete files
9950 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9951 permission on "config" script to be executable) and a fix for the INSTALL
9953 [Ulf Moeller <ulf@fitug.de>]
9955 *) Remove some legacy and erroneous uses of malloc, free instead of
9957 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9959 *) Make rsa_oaep_test return non-zero on error.
9960 [Ulf Moeller <ulf@fitug.de>]
9962 *) Add support for native Solaris shared libraries. Configure
9963 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9964 if someone would make that last step automatic.
9965 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9967 *) ctx_size was not built with the right compiler during "make links". Fixed.
9970 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9971 except NULL ciphers". This means the default cipher list will no longer
9972 enable NULL ciphers. They need to be specifically enabled e.g. with
9973 the string "DEFAULT:eNULL".
9976 *) Fix to RSA private encryption routines: if p < q then it would
9977 occasionally produce an invalid result. This will only happen with
9978 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9981 *) Be less restrictive and allow also `perl util/perlpath.pl
9982 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9983 because this way one can also use an interpreter named `perl5' (which is
9984 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9985 installed as `perl').
9986 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9988 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9989 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9991 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9992 advapi32.lib to Win32 build and change the pem test comparision
9993 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9994 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9995 and crypto/des/ede_cbcm_enc.c.
9998 *) DES quad checksum was broken on big-endian architectures. Fixed.
10001 *) Comment out two functions in bio.h that aren't implemented. Fix up the
10002 Win32 test batch file so it (might) work again. The Win32 test batch file
10003 is horrible: I feel ill....
10006 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
10007 in e_os.h. Audit of header files to check ANSI and non ANSI
10008 sections: 10 functions were absent from non ANSI section and not exported
10009 from Windows DLLs. Fixed up libeay.num for new functions.
10012 *) Make `openssl version' output lines consistent.
10013 [Ralf S. Engelschall]
10015 *) Fix Win32 symbol export lists for BIO functions: Added
10016 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10017 to ms/libeay{16,32}.def.
10018 [Ralf S. Engelschall]
10020 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10021 fine under Unix and passes some trivial tests I've now added. But the
10022 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10023 added to make sure no one expects that this stuff really works in the
10024 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10025 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10027 [Ralf S. Engelschall]
10029 *) Fix the generation of two part addresses in perl.
10030 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10032 *) Add config entry for Linux on MIPS.
10033 [John Tobey <jtobey@channel1.com>]
10035 *) Make links whenever Configure is run, unless we are on Windoze.
10038 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10039 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10043 *) Add a useful kludge to allow package maintainers to specify compiler and
10044 other platforms details on the command line without having to patch the
10045 Configure script everytime: One now can use ``perl Configure
10046 <id>:<details>'', i.e. platform ids are allowed to have details appended
10047 to them (seperated by colons). This is treated as there would be a static
10048 pre-configured entry in Configure's %table under key <id> with value
10049 <details> and ``perl Configure <id>'' is called. So, when you want to
10050 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10051 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10052 now, which overrides the FreeBSD-elf entry on-the-fly.
10053 [Ralf S. Engelschall]
10055 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10058 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10059 on the `perl Configure ...' command line. This way one can compile
10060 OpenSSL libraries with Position Independent Code (PIC) which is needed
10061 for linking it into DSOs.
10062 [Ralf S. Engelschall]
10064 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10068 *) Cleaned up the LICENSE document: The official contact for any license
10069 questions now is the OpenSSL core team under openssl-core@openssl.org.
10070 And add a paragraph about the dual-license situation to make sure people
10071 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10072 to the OpenSSL toolkit.
10073 [Ralf S. Engelschall]
10075 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10076 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10077 Additonally cleaned up the `make links' target: Remove unnecessary
10078 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10079 to speed processing and no longer clutter the display with confusing
10080 stuff. Instead only the actually done links are displayed.
10081 [Ralf S. Engelschall]
10083 *) Permit null encryption ciphersuites, used for authentication only. It used
10084 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10085 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10089 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10090 signed attributes when verifying signatures (this would break them),
10091 the detached data encoding was wrong and public keys obtained using
10092 X509_get_pubkey() weren't freed.
10095 *) Add text documentation for the BUFFER functions. Also added a work around
10096 to a Win95 console bug. This was triggered by the password read stuff: the
10097 last character typed gets carried over to the next fread(). If you were
10098 generating a new cert request using 'req' for example then the last
10099 character of the passphrase would be CR which would then enter the first
10103 *) Added the new `Includes OpenSSL Cryptography Software' button as
10104 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10105 button and can be used by applications based on OpenSSL to show the
10106 relationship to the OpenSSL project.
10107 [Ralf S. Engelschall]
10109 *) Remove confusing variables in function signatures in files
10110 ssl/ssl_lib.c and ssl/ssl.h.
10111 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10113 *) Don't install bss_file.c under PREFIX/include/
10114 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10116 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10117 functions that return function pointers and has support for NT specific
10118 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10119 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10120 unsigned to signed types: this was killing the Win32 compile.
10123 *) Add new certificate file to stack functions,
10124 SSL_add_dir_cert_subjects_to_stack() and
10125 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10126 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10127 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10128 This means that Apache-SSL and similar packages don't have to mess around
10129 to add as many CAs as they want to the preferred list.
10132 *) Experiment with doxygen documentation. Currently only partially applied to
10134 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10135 openssl.doxy as the configuration file.
10138 *) Get rid of remaining C++-style comments which strict C compilers hate.
10139 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10141 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10142 compiled in by default: it has problems with large keys.
10145 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10146 DH private keys and/or callback functions which directly correspond to
10147 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10148 is needed for applications which have to configure certificates on a
10149 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10151 For the RSA certificate situation is makes no difference, but
10152 for the DSA certificate situation this fixes the "no shared cipher"
10153 problem where the OpenSSL cipher selection procedure failed because the
10154 temporary keys were not overtaken from the context and the API provided
10155 no way to reconfigure them.
10156 The new functions now let applications reconfigure the stuff and they
10157 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
10158 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
10159 non-public-API function ssl_cert_instantiate() is used as a helper
10160 function and also to reduce code redundancy inside ssl_rsa.c.
10161 [Ralf S. Engelschall]
10163 *) Move s_server -dcert and -dkey options out of the undocumented feature
10164 area because they are useful for the DSA situation and should be
10165 recognized by the users.
10166 [Ralf S. Engelschall]
10168 *) Fix the cipher decision scheme for export ciphers: the export bits are
10169 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
10170 SSL_EXP_MASK. So, the original variable has to be used instead of the
10171 already masked variable.
10172 [Richard Levitte <levitte@stacken.kth.se>]
10174 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
10175 [Richard Levitte <levitte@stacken.kth.se>]
10177 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
10178 from `int' to `unsigned int' because it's a length and initialized by
10179 EVP_DigestFinal() which expects an `unsigned int *'.
10180 [Richard Levitte <levitte@stacken.kth.se>]
10182 *) Don't hard-code path to Perl interpreter on shebang line of Configure
10183 script. Instead use the usual Shell->Perl transition trick.
10184 [Ralf S. Engelschall]
10186 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
10187 (in addition to RSA certificates) to match the behaviour of `openssl dsa
10188 -noout -modulus' as it's already the case for `openssl rsa -noout
10189 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
10190 currently the public key is printed (a decision which was already done by
10191 `openssl dsa -modulus' in the past) which serves a similar purpose.
10192 Additionally the NO_RSA no longer completely removes the whole -modulus
10193 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
10195 [Ralf S. Engelschall]
10197 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
10198 BIO. See the source (crypto/evp/bio_ok.c) for more info.
10199 [Arne Ansper <arne@ats.cyber.ee>]
10201 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
10202 to be added. Now both 'req' and 'ca' can use new objects defined in the
10206 *) Add cool BIO that does syslog (or event log on NT).
10207 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
10209 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
10210 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
10211 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
10212 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
10215 *) Add preliminary config info for new extension code.
10218 *) Make RSA_NO_PADDING really use no padding.
10219 [Ulf Moeller <ulf@fitug.de>]
10221 *) Generate errors when private/public key check is done.
10224 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
10225 for some CRL extensions and new objects added.
10228 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
10229 key usage extension and fuller support for authority key id.
10232 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
10233 padding method for RSA, which is recommended for new applications in PKCS
10234 #1 v2.0 (RFC 2437, October 1998).
10235 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
10236 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
10237 against Bleichbacher's attack on RSA.
10238 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
10241 *) Updates to the new SSL compression code
10242 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10244 *) Fix so that the version number in the master secret, when passed
10245 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
10246 (because the server will not accept higher), that the version number
10247 is 0x03,0x01, not 0x03,0x00
10248 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10250 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
10251 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
10252 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
10255 *) Support for RAW extensions where an arbitrary extension can be
10256 created by including its DER encoding. See apps/openssl.cnf for
10260 *) Make sure latest Perl versions don't interpret some generated C array
10261 code as Perl array code in the crypto/err/err_genc.pl script.
10262 [Lars Weber <3weber@informatik.uni-hamburg.de>]
10264 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
10265 not many people have the assembler. Various Win32 compilation fixes and
10266 update to the INSTALL.W32 file with (hopefully) more accurate Win32
10267 build instructions.
10270 *) Modify configure script 'Configure' to automatically create crypto/date.h
10271 file under Win32 and also build pem.h from pem.org. New script
10272 util/mkfiles.pl to create the MINFO file on environments that can't do a
10273 'make files': perl util/mkfiles.pl >MINFO should work.
10276 *) Major rework of DES function declarations, in the pursuit of correctness
10277 and purity. As a result, many evil casts evaporated, and some weirdness,
10278 too. You may find this causes warnings in your code. Zapping your evil
10279 casts will probably fix them. Mostly.
10282 *) Fix for a typo in asn1.h. Bug fix to object creation script
10283 obj_dat.pl. It considered a zero in an object definition to mean
10284 "end of object": none of the objects in objects.h have any zeros
10285 so it wasn't spotted.
10286 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
10288 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
10289 Masking (CBCM). In the absence of test vectors, the best I have been able
10290 to do is check that the decrypt undoes the encrypt, so far. Send me test
10291 vectors if you have them.
10294 *) Correct calculation of key length for export ciphers (too much space was
10295 allocated for null ciphers). This has not been tested!
10298 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
10299 message is now correct (it understands "crypto" and "ssl" on its
10300 command line). There is also now an "update" option. This will update
10301 the util/ssleay.num and util/libeay.num files with any new functions.
10303 perl util/mkdef.pl crypto ssl update
10304 it will update them.
10307 *) Overhauled the Perl interface (perl/*):
10308 - ported BN stuff to OpenSSL's different BN library
10309 - made the perl/ source tree CVS-aware
10310 - renamed the package from SSLeay to OpenSSL (the files still contain
10311 their history because I've copied them in the repository)
10312 - removed obsolete files (the test scripts will be replaced
10313 by better Test::Harness variants in the future)
10314 [Ralf S. Engelschall]
10316 *) First cut for a very conservative source tree cleanup:
10317 1. merge various obsolete readme texts into doc/ssleay.txt
10318 where we collect the old documents and readme texts.
10319 2. remove the first part of files where I'm already sure that we no
10320 longer need them because of three reasons: either they are just temporary
10321 files which were left by Eric or they are preserved original files where
10322 I've verified that the diff is also available in the CVS via "cvs diff
10323 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
10324 the crypto/md/ stuff).
10325 [Ralf S. Engelschall]
10327 *) More extension code. Incomplete support for subject and issuer alt
10328 name, issuer and authority key id. Change the i2v function parameters
10329 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
10330 what that's for :-) Fix to ASN1 macro which messed up
10331 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
10334 *) Preliminary support for ENUMERATED type. This is largely copied from the
10338 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
10339 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10341 *) Make sure `make rehash' target really finds the `openssl' program.
10342 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10344 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
10345 like to hear about it if this slows down other processors.
10348 *) Add CygWin32 platform information to Configure script.
10349 [Alan Batie <batie@aahz.jf.intel.com>]
10351 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
10352 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
10354 *) New program nseq to manipulate netscape certificate sequences
10357 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
10361 *) Fixes to BN code. Previously the default was to define BN_RECURSION
10362 but the BN code had some problems that would cause failures when
10363 doing certificate verification and some other functions.
10364 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10366 *) Add ASN1 and PEM code to support netscape certificate sequences.
10369 *) Add ASN1 and PEM code to support netscape certificate sequences.
10372 *) Add several PKIX and private extended key usage OIDs.
10375 *) Modify the 'ca' program to handle the new extension code. Modify
10376 openssl.cnf for new extension format, add comments.
10379 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
10380 and add a sample to openssl.cnf so req -x509 now adds appropriate
10384 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
10385 error code, add initial support to X509_print() and x509 application.
10388 *) Takes a deep breath and start addding X509 V3 extension support code. Add
10389 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
10390 stuff is currently isolated and isn't even compiled yet.
10393 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
10394 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
10395 Removed the versions check from X509 routines when loading extensions:
10396 this allows certain broken certificates that don't set the version
10397 properly to be processed.
10400 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
10401 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
10402 can still be regenerated with "make depend".
10405 *) Spelling mistake in C version of CAST-128.
10406 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
10408 *) Changes to the error generation code. The perl script err-code.pl
10409 now reads in the old error codes and retains the old numbers, only
10410 adding new ones if necessary. It also only changes the .err files if new
10411 codes are added. The makefiles have been modified to only insert errors
10412 when needed (to avoid needlessly modifying header files). This is done
10413 by only inserting errors if the .err file is newer than the auto generated
10414 C file. To rebuild all the error codes from scratch (the old behaviour)
10415 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
10416 or delete all the .err files.
10419 *) CAST-128 was incorrectly implemented for short keys. The C version has
10420 been fixed, but is untested. The assembler versions are also fixed, but
10421 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
10422 to regenerate it if needed.
10423 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
10424 Hagino <itojun@kame.net>]
10426 *) File was opened incorrectly in randfile.c.
10427 [Ulf Möller <ulf@fitug.de>]
10429 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
10430 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
10431 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
10432 al: it's just almost always a UTCTime. Note this patch adds new error
10433 codes so do a "make errors" if there are problems.
10436 *) Correct Linux 1 recognition in config.
10437 [Ulf Möller <ulf@fitug.de>]
10439 *) Remove pointless MD5 hash when using DSA keys in ca.
10440 [Anonymous <nobody@replay.com>]
10442 *) Generate an error if given an empty string as a cert directory. Also
10443 generate an error if handed NULL (previously returned 0 to indicate an
10444 error, but didn't set one).
10445 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
10447 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
10450 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
10451 parameters. This was causing a warning which killed off the Win32 compile.
10454 *) Remove C++ style comments from crypto/bn/bn_local.h.
10455 [Neil Costigan <neil.costigan@celocom.com>]
10457 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
10458 based on a text string, looking up short and long names and finally
10459 "dot" format. The "dot" format stuff didn't work. Added new function
10460 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
10461 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
10462 OID is not part of the table.
10465 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
10466 X509_LOOKUP_by_alias().
10469 *) Sort openssl functions by name.
10472 *) Get the gendsa program working (hopefully) and add it to app list. Remove
10473 encryption from sample DSA keys (in case anyone is interested the password
10477 *) Make _all_ *_free functions accept a NULL pointer.
10478 [Frans Heymans <fheymans@isaserver.be>]
10480 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10482 [Anonymous <nobody@replay.com>]
10484 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10485 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10487 *) Don't blow it for numeric -newkey arguments to apps/req.
10488 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10490 *) Temp key "for export" tests were wrong in s3_srvr.c.
10491 [Anonymous <nobody@replay.com>]
10493 *) Add prototype for temp key callback functions
10494 SSL_CTX_set_tmp_{rsa,dh}_callback().
10497 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10498 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10501 *) X509_name_add_entry() freed the wrong thing after an error.
10502 [Arne Ansper <arne@ats.cyber.ee>]
10504 *) rsa_eay.c would attempt to free a NULL context.
10505 [Arne Ansper <arne@ats.cyber.ee>]
10507 *) BIO_s_socket() had a broken should_retry() on Windoze.
10508 [Arne Ansper <arne@ats.cyber.ee>]
10510 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10511 [Arne Ansper <arne@ats.cyber.ee>]
10513 *) Make sure the already existing X509_STORE->depth variable is initialized
10514 in X509_STORE_new(), but document the fact that this variable is still
10515 unused in the certificate verification process.
10516 [Ralf S. Engelschall]
10518 *) Fix the various library and apps files to free up pkeys obtained from
10519 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10522 *) Fix reference counting in X509_PUBKEY_get(). This makes
10523 demos/maurice/example2.c work, amongst others, probably.
10524 [Steve Henson and Ben Laurie]
10526 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10527 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10528 are no longer created. This way we have a single and consistent command
10529 line interface `openssl <command>', similar to `cvs <command>'.
10530 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10532 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10533 BIT STRING wrapper always have zero unused bits.
10536 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10539 *) Make the top-level INSTALL documentation easier to understand.
10542 *) Makefiles updated to exit if an error occurs in a sub-directory
10543 make (including if user presses ^C) [Paul Sutton]
10545 *) Make Montgomery context stuff explicit in RSA data structure.
10548 *) Fix build order of pem and err to allow for generated pem.h.
10551 *) Fix renumbering bug in X509_NAME_delete_entry().
10554 *) Enhanced the err-ins.pl script so it makes the error library number
10555 global and can add a library name. This is needed for external ASN1 and
10556 other error libraries.
10559 *) Fixed sk_insert which never worked properly.
10562 *) Fix ASN1 macros so they can handle indefinite length construted
10563 EXPLICIT tags. Some non standard certificates use these: they can now
10567 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10568 into a single doc/ssleay.txt bundle. This way the information is still
10569 preserved but no longer messes up this directory. Now it's new room for
10570 the new set of documenation files.
10571 [Ralf S. Engelschall]
10573 *) SETs were incorrectly DER encoded. This was a major pain, because they
10574 shared code with SEQUENCEs, which aren't coded the same. This means that
10575 almost everything to do with SETs or SEQUENCEs has either changed name or
10576 number of arguments.
10577 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10579 *) Fix test data to work with the above.
10582 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10583 was already fixed by Eric for 0.9.1 it seems.
10584 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10586 *) Autodetect FreeBSD3.
10589 *) Fix various bugs in Configure. This affects the following platforms:
10593 unixware-2.0-pentium
10597 *) Eliminate generated files from CVS. Reorder tests to regenerate files
10598 before they are needed.
10601 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
10605 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
10607 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
10608 changed SSLeay to OpenSSL in version strings.
10609 [Ralf S. Engelschall]
10611 *) Some fixups to the top-level documents.
10614 *) Fixed the nasty bug where rsaref.h was not found under compile-time
10615 because the symlink to include/ was missing.
10616 [Ralf S. Engelschall]
10618 *) Incorporated the popular no-RSA/DSA-only patches
10619 which allow to compile a RSA-free SSLeay.
10620 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
10622 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
10623 when "ssleay" is still not found.
10624 [Ralf S. Engelschall]
10626 *) Added more platforms to Configure: Cray T3E, HPUX 11,
10627 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
10629 *) Updated the README file.
10630 [Ralf S. Engelschall]
10632 *) Added various .cvsignore files in the CVS repository subdirs
10633 to make a "cvs update" really silent.
10634 [Ralf S. Engelschall]
10636 *) Recompiled the error-definition header files and added
10637 missing symbols to the Win32 linker tables.
10638 [Ralf S. Engelschall]
10640 *) Cleaned up the top-level documents;
10641 o new files: CHANGES and LICENSE
10642 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
10643 o merged COPYRIGHT into LICENSE
10644 o removed obsolete TODO file
10645 o renamed MICROSOFT to INSTALL.W32
10646 [Ralf S. Engelschall]
10648 *) Removed dummy files from the 0.9.1b source tree:
10649 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
10650 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
10651 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
10652 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
10653 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
10654 [Ralf S. Engelschall]
10656 *) Added various platform portability fixes.
10659 *) The Genesis of the OpenSSL rpject:
10660 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
10661 Young and Tim J. Hudson created while they were working for C2Net until
10663 [The OpenSSL Project]
10666 Changes between 0.9.0b and 0.9.1b [not released]
10668 *) Updated a few CA certificates under certs/
10671 *) Changed some BIGNUM api stuff.
10674 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
10675 DGUX x86, Linux Alpha, etc.
10678 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
10679 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
10683 *) Add -strparse option to asn1pars program which parses nested
10685 [Dr Stephen Henson <shenson@bigfoot.com>]
10687 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
10690 *) DSA fix for "ca" program.
10693 *) Added "-genkey" option to "dsaparam" program.
10696 *) Added RIPE MD160 (rmd160) message digest.
10699 *) Added -a (all) option to "ssleay version" command.
10702 *) Added PLATFORM define which is the id given to Configure.
10705 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
10708 *) Extended the ASN.1 parser routines.
10711 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
10714 *) Added a BN_CTX to the BN library.
10717 *) Fixed the weak key values in DES library
10720 *) Changed API in EVP library for cipher aliases.
10723 *) Added support for RC2/64bit cipher.
10726 *) Converted the lhash library to the crypto/mem.c functions.
10729 *) Added more recognized ASN.1 object ids.
10732 *) Added more RSA padding checks for SSL/TLS.
10735 *) Added BIO proxy/filter functionality.
10738 *) Added extra_certs to SSL_CTX which can be used
10739 send extra CA certificates to the client in the CA cert chain sending
10740 process. It can be configured with SSL_CTX_add_extra_chain_cert().
10743 *) Now Fortezza is denied in the authentication phase because
10744 this is key exchange mechanism is not supported by SSLeay at all.
10747 *) Additional PKCS1 checks.
10750 *) Support the string "TLSv1" for all TLS v1 ciphers.
10753 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
10754 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
10757 *) Fixed a few memory leaks.
10760 *) Fixed various code and comment typos.
10763 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
10764 bytes sent in the client random.
10765 [Edward Bishop <ebishop@spyglass.com>]