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 EVP support for key wrapping algorithms, to avoid problems with
64 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
65 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
66 algorithms and include tests cases.
69 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
73 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
74 MGF1 digest and OAEP label.
77 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
78 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
79 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
82 *) Make openssl verify return errors.
83 [Chris Palmer <palmer@google.com> and Ben Laurie]
85 *) New function ASN1_TIME_diff to calculate the difference between two
86 ASN1_TIME structures or one structure and the current time.
89 *) Update fips_test_suite to support multiple command line options. New
90 test to induce all self test errors in sequence and check expected
94 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
95 sign or verify all in one operation.
98 *) Add fips_algvs: a multicall fips utility incorporaing all the algorithm
99 test programs and fips_test_suite. Includes functionality to parse
100 the minimal script output of fipsalgest.pl directly.
103 *) Add authorisation parameter to FIPS_module_mode_set().
106 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
109 *) Use separate DRBG fields for internal and external flags. New function
110 FIPS_drbg_health_check() to perform on demand health checking. Add
111 generation tests to fips_test_suite with reduced health check interval to
112 demonstrate periodic health checking. Add "nodh" option to
113 fips_test_suite to skip very slow DH test.
116 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
120 *) More extensive health check for DRBG checking many more failure modes.
121 New function FIPS_selftest_drbg_all() to handle every possible DRBG
122 combination: call this in fips_test_suite.
125 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
126 and POST to handle Dual EC cases.
129 *) Add support for canonical generation of DSA parameter 'g'. See
132 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
133 POST to handle HMAC cases.
136 *) Add functions FIPS_module_version() and FIPS_module_version_text()
137 to return numerical and string versions of the FIPS module number.
140 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
141 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
142 outside the validated module in the FIPS capable OpenSSL.
145 *) Minor change to DRBG entropy callback semantics. In some cases
146 there is no multiple of the block length between min_len and
147 max_len. Allow the callback to return more than max_len bytes
148 of entropy but discard any extra: it is the callback's responsibility
149 to ensure that the extra data discarded does not impact the
150 requested amount of entropy.
153 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
154 information in FIPS186-3, SP800-57 and SP800-131A.
157 *) CCM support via EVP. Interface is very similar to GCM case except we
158 must supply all data in one chunk (i.e. no update, final) and the
159 message length must be supplied if AAD is used. Add algorithm test
163 *) Initial version of POST overhaul. Add POST callback to allow the status
164 of POST to be monitored and/or failures induced. Modify fips_test_suite
165 to use callback. Always run all selftests even if one fails.
168 *) XTS support including algorithm test driver in the fips_gcmtest program.
169 Note: this does increase the maximum key length from 32 to 64 bytes but
170 there should be no binary compatibility issues as existing applications
171 will never use XTS mode.
174 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
175 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
176 performs algorithm blocking for unapproved PRNG types. Also do not
177 set PRNG type in FIPS_mode_set(): leave this to the application.
178 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
179 the standard OpenSSL PRNG: set additional data to a date time vector.
182 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
183 This shouldn't present any incompatibility problems because applications
184 shouldn't be using these directly and any that are will need to rethink
185 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
188 *) Extensive self tests and health checking required by SP800-90 DRBG.
189 Remove strength parameter from FIPS_drbg_instantiate and always
190 instantiate at maximum supported strength.
193 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
196 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
199 *) New function DH_compute_key_padded() to compute a DH key and pad with
200 leading zeroes if needed: this complies with SP800-56A et al.
203 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
204 anything, incomplete, subject to change and largely untested at present.
207 *) Modify fipscanisteronly build option to only build the necessary object
208 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
211 *) Add experimental option FIPSSYMS to give all symbols in
212 fipscanister.o and FIPS or fips prefix. This will avoid
213 conflicts with future versions of OpenSSL. Add perl script
214 util/fipsas.pl to preprocess assembly language source files
215 and rename any affected symbols.
218 *) Add selftest checks and algorithm block of non-fips algorithms in
219 FIPS mode. Remove DES2 from selftests.
222 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
223 return internal method without any ENGINE dependencies. Add new
224 tiny fips sign and verify functions.
227 *) New build option no-ec2m to disable characteristic 2 code.
230 *) New build option "fipscanisteronly". This only builds fipscanister.o
231 and (currently) associated fips utilities. Uses the file Makefile.fips
232 instead of Makefile.org as the prototype.
235 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
236 Update fips_gcmtest to use IV generator.
239 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
240 setting output buffer to NULL. The *Final function must be
241 called although it will not retrieve any additional data. The tag
242 can be set or retrieved with a ctrl. The IV length is by default 12
243 bytes (96 bits) but can be set to an alternative value. If the IV
244 length exceeds the maximum IV length (currently 16 bytes) it cannot be
248 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
249 underlying do_cipher function handles all cipher semantics itself
250 including padding and finalisation. This is useful if (for example)
251 an ENGINE cipher handles block padding itself. The behaviour of
252 do_cipher is subtly changed if this flag is set: the return value
253 is the number of characters written to the output buffer (zero is
254 no longer an error code) or a negative error code. Also if the
255 input buffer is NULL and length 0 finalisation should be performed.
258 *) If a candidate issuer certificate is already part of the constructed
259 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
262 *) Improve forward-security support: add functions
264 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
265 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
267 for use by SSL/TLS servers; the callback function will be called whenever a
268 new session is created, and gets to decide whether the session may be
269 cached to make it resumable (return 0) or not (return 1). (As by the
270 SSL/TLS protocol specifications, the session_id sent by the server will be
271 empty to indicate that the session is not resumable; also, the server will
272 not generate RFC 4507 (RFC 5077) session tickets.)
274 A simple reasonable callback implementation is to return is_forward_secure.
275 This parameter will be set to 1 or 0 depending on the ciphersuite selected
276 by the SSL/TLS server library, indicating whether it can provide forward
278 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
280 *) New -verify_name option in command line utilities to set verification
284 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
285 Add CMAC pkey methods.
288 *) Experimental regnegotiation in s_server -www mode. If the client
289 browses /reneg connection is renegotiated. If /renegcert it is
290 renegotiated requesting a certificate.
293 *) Add an "external" session cache for debugging purposes to s_server. This
294 should help trace issues which normally are only apparent in deployed
295 multi-process servers.
298 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
299 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
300 BIO_set_cipher() and some obscure PEM functions were changed so they
301 can now return an error. The RAND changes required a change to the
302 RAND_METHOD structure.
305 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
306 a gcc attribute to warn if the result of a function is ignored. This
307 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
308 whose return value is often ignored.
311 Changes between 1.0.1e and 1.0.2 [xx XXX xxxx]
313 *) Keep original DTLS digest and encryption contexts in retransmission
314 structures so we can use the previous session parameters if they need
315 to be resent. (CVE-2013-6450)
318 *) TLS pad extension: draft-agl-tls-padding-03
320 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
321 TLS client Hello record length value would otherwise be > 255 and
322 less that 512 pad with a dummy extension containing zeroes so it
323 is at least 512 bytes long.
325 [Adam Langley, Steve Henson]
327 *) Add functions to allocate and set the fields of an ECDSA_METHOD
329 [Douglas E. Engert, Steve Henson]
331 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
332 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
333 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
334 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
335 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
336 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
337 [Rob Stradling, Adam Langley]
339 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
340 difference in days and seconds between two tm or ASN1_TIME structures.
343 *) Add -rev test option to s_server to just reverse order of characters
344 received by client and send back to server. Also prints an abbreviated
345 summary of the connection parameters.
348 *) New option -brief for s_client and s_server to print out a brief summary
349 of connection parameters.
352 *) Add callbacks for arbitrary TLS extensions.
353 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
355 *) New option -crl_download in several openssl utilities to download CRLs
356 from CRLDP extension in certificates.
359 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
362 *) New function X509_CRL_diff to generate a delta CRL from the difference
363 of two full CRLs. Add support to "crl" utility.
366 *) New functions to set lookup_crls function and to retrieve
367 X509_STORE from X509_STORE_CTX.
370 *) Print out deprecated issuer and subject unique ID fields in
374 *) Extend OCSP I/O functions so they can be used for simple general purpose
375 HTTP as well as OCSP. New wrapper function which can be used to download
376 CRLs using the OCSP API.
379 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
382 *) SSL_CONF* functions. These provide a common framework for application
383 configuration using configuration files or command lines.
386 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
387 message callback and prints the results. Needs compile time option
388 "enable-ssl-trace". New options to s_client and s_server to enable
392 *) New ctrl and macro to retrieve supported points extensions.
393 Print out extension in s_server and s_client.
396 *) New functions to retrieve certificate signature and signature
400 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
404 *) New Suite B modes for TLS code. These use and enforce the requirements
405 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
406 only use Suite B curves. The Suite B modes can be set by using the
407 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
410 *) New chain verification flags for Suite B levels of security. Check
411 algorithms are acceptable when flags are set in X509_verify_cert.
414 *) Make tls1_check_chain return a set of flags indicating checks passed
415 by a certificate chain. Add additional tests to handle client
416 certificates: checks for matching certificate type and issuer name
420 *) If an attempt is made to use a signature algorithm not in the peer
421 preference list abort the handshake. If client has no suitable
422 signature algorithms in response to a certificate request do not
426 *) If server EC tmp key is not in client preference list abort handshake.
429 *) Add support for certificate stores in CERT structure. This makes it
430 possible to have different stores per SSL structure or one store in
431 the parent SSL_CTX. Include distint stores for certificate chain
432 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
433 to build and store a certificate chain in CERT structure: returing
434 an error if the chain cannot be built: this will allow applications
435 to test if a chain is correctly configured.
437 Note: if the CERT based stores are not set then the parent SSL_CTX
438 store is used to retain compatibility with existing behaviour.
442 *) New function ssl_set_client_disabled to set a ciphersuite disabled
443 mask based on the current session, check mask when sending client
444 hello and checking the requested ciphersuite.
447 *) New ctrls to retrieve and set certificate types in a certificate
448 request message. Print out received values in s_client. If certificate
449 types is not set with custom values set sensible values based on
450 supported signature algorithms.
453 *) Support for distinct client and server supported signature algorithms.
456 *) Add certificate callback. If set this is called whenever a certificate
457 is required by client or server. An application can decide which
458 certificate chain to present based on arbitrary criteria: for example
459 supported signature algorithms. Add very simple example to s_server.
460 This fixes many of the problems and restrictions of the existing client
461 certificate callback: for example you can now clear an existing
462 certificate and specify the whole chain.
465 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
466 the certificate can be used for (if anything). Set valid_flags field
467 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
468 to have similar checks in it.
470 Add new "cert_flags" field to CERT structure and include a "strict mode".
471 This enforces some TLS certificate requirements (such as only permitting
472 certificate signature algorithms contained in the supported algorithms
473 extension) which some implementations ignore: this option should be used
474 with caution as it could cause interoperability issues.
477 *) Update and tidy signature algorithm extension processing. Work out
478 shared signature algorithms based on preferences and peer algorithms
479 and print them out in s_client and s_server. Abort handshake if no
480 shared signature algorithms.
483 *) Add new functions to allow customised supported signature algorithms
484 for SSL and SSL_CTX structures. Add options to s_client and s_server
488 *) New function SSL_certs_clear() to delete all references to certificates
489 from an SSL structure. Before this once a certificate had been added
490 it couldn't be removed.
493 *) Integrate hostname, email address and IP address checking with certificate
494 verification. New verify options supporting checking in opensl utility.
497 *) Fixes and wildcard matching support to hostname and email checking
498 functions. Add manual page.
499 [Florian Weimer (Red Hat Product Security Team)]
501 *) New functions to check a hostname email or IP address against a
502 certificate. Add options x509 utility to print results of checks against
506 *) Fix OCSP checking.
507 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
509 *) Initial experimental support for explicitly trusted non-root CAs.
510 OpenSSL still tries to build a complete chain to a root but if an
511 intermediate CA has a trust setting included that is used. The first
512 setting is used: whether to trust (e.g., -addtrust option to the x509
516 *) Add -trusted_first option which attempts to find certificates in the
517 trusted store even if an untrusted chain is also supplied.
520 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
521 platform support for Linux and Android.
524 *) Support for linux-x32, ILP32 environment in x86_64 framework.
527 *) Experimental multi-implementation support for FIPS capable OpenSSL.
528 When in FIPS mode the approved implementations are used as normal,
529 when not in FIPS mode the internal unapproved versions are used instead.
530 This means that the FIPS capable OpenSSL isn't forced to use the
531 (often lower perfomance) FIPS implementations outside FIPS mode.
534 *) Transparently support X9.42 DH parameters when calling
535 PEM_read_bio_DHparameters. This means existing applications can handle
536 the new parameter format automatically.
539 *) Initial experimental support for X9.42 DH parameter format: mainly
540 to support use of 'q' parameter for RFC5114 parameters.
543 *) Add DH parameters from RFC5114 including test data to dhtest.
546 *) Support for automatic EC temporary key parameter selection. If enabled
547 the most preferred EC parameters are automatically used instead of
548 hardcoded fixed parameters. Now a server just has to call:
549 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
550 support ECDH and use the most appropriate parameters.
553 *) Enhance and tidy EC curve and point format TLS extension code. Use
554 static structures instead of allocation if default values are used.
555 New ctrls to set curves we wish to support and to retrieve shared curves.
556 Print out shared curves in s_server. New options to s_server and s_client
557 to set list of supported curves.
560 *) New ctrls to retrieve supported signature algorithms and
561 supported curve values as an array of NIDs. Extend openssl utility
562 to print out received values.
565 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
566 between NIDs and the more common NIST names such as "P-256". Enhance
567 ecparam utility and ECC method to recognise the NIST names for curves.
570 *) Enhance SSL/TLS certificate chain handling to support different
571 chains for each certificate instead of one chain in the parent SSL_CTX.
574 *) Support for fixed DH ciphersuite client authentication: where both
575 server and client use DH certificates with common parameters.
578 *) Support for fixed DH ciphersuites: those requiring DH server
582 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
584 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
585 supporting platforms or when small records were transferred.
586 [Andy Polyakov, Steve Henson]
588 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
590 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
592 This addresses the flaw in CBC record processing discovered by
593 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
594 at: http://www.isg.rhul.ac.uk/tls/
596 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
597 Security Group at Royal Holloway, University of London
598 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
599 Emilia Käsper for the initial patch.
601 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
603 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
604 ciphersuites which can be exploited in a denial of service attack.
605 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
606 and detecting this bug and to Wolfgang Ettlinger
607 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
611 *) Return an error when checking OCSP signatures when key is NULL.
612 This fixes a DoS attack. (CVE-2013-0166)
615 *) Make openssl verify return errors.
616 [Chris Palmer <palmer@google.com> and Ben Laurie]
618 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
619 the right response is stapled. Also change SSL_get_certificate()
620 so it returns the certificate actually sent.
621 See http://rt.openssl.org/Ticket/Display.html?id=2836.
622 [Rob Stradling <rob.stradling@comodo.com>]
624 *) Fix possible deadlock when decoding public keys.
627 *) Don't use TLS 1.0 record version number in initial client hello
631 Changes between 1.0.1b and 1.0.1c [10 May 2012]
633 *) Sanity check record length before skipping explicit IV in TLS
634 1.2, 1.1 and DTLS to fix DoS attack.
636 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
637 fuzzing as a service testing platform.
641 *) Initialise tkeylen properly when encrypting CMS messages.
642 Thanks to Solar Designer of Openwall for reporting this issue.
645 *) In FIPS mode don't try to use composite ciphers as they are not
649 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
651 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
652 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
653 mean any application compiled against OpenSSL 1.0.0 headers setting
654 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
655 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
656 0x10000000L Any application which was previously compiled against
657 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
658 will need to be recompiled as a result. Letting be results in
659 inability to disable specifically TLS 1.1 and in client context,
660 in unlike event, limit maximum offered version to TLS 1.0 [see below].
663 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
664 disable just protocol X, but all protocols above X *if* there are
665 protocols *below* X still enabled. In more practical terms it means
666 that if application wants to disable TLS1.0 in favor of TLS1.1 and
667 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
668 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
672 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
674 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
675 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
676 in CRYPTO_realloc_clean.
678 Thanks to Tavis Ormandy, Google Security Team, for discovering this
679 issue and to Adam Langley <agl@chromium.org> for fixing it.
681 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
683 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
686 *) Workarounds for some broken servers that "hang" if a client hello
687 record length exceeds 255 bytes.
689 1. Do not use record version number > TLS 1.0 in initial client
690 hello: some (but not all) hanging servers will now work.
691 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
692 the number of ciphers sent in the client hello. This should be
693 set to an even number, such as 50, for example by passing:
694 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
695 Most broken servers should now work.
696 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
697 TLS 1.2 client support entirely.
700 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
703 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
705 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
706 STRING form instead of a DigestInfo.
709 *) The format used for MDC2 RSA signatures is inconsistent between EVP
710 and the RSA_sign/RSA_verify functions. This was made more apparent when
711 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
712 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
713 the correct format in RSA_verify so both forms transparently work.
716 *) Some servers which support TLS 1.0 can choke if we initially indicate
717 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
718 encrypted premaster secret. As a workaround use the maximum pemitted
719 client version in client hello, this should keep such servers happy
720 and still work with previous versions of OpenSSL.
723 *) Add support for TLS/DTLS heartbeats.
724 [Robin Seggelmann <seggelmann@fh-muenster.de>]
726 *) Add support for SCTP.
727 [Robin Seggelmann <seggelmann@fh-muenster.de>]
729 *) Improved PRNG seeding for VOS.
730 [Paul Green <Paul.Green@stratus.com>]
732 *) Extensive assembler packs updates, most notably:
734 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
735 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
736 - x86_64: bit-sliced AES implementation;
737 - ARM: NEON support, contemporary platforms optimizations;
738 - s390x: z196 support;
739 - *: GHASH and GF(2^m) multiplication implementations;
743 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
744 (removal of unnecessary code)
745 [Peter Sylvester <peter.sylvester@edelweb.fr>]
747 *) Add TLS key material exporter from RFC 5705.
750 *) Add DTLS-SRTP negotiation from RFC 5764.
753 *) Add Next Protocol Negotiation,
754 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
755 disabled with a no-npn flag to config or Configure. Code donated
757 [Adam Langley <agl@google.com> and Ben Laurie]
759 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
760 NIST-P256, NIST-P521, with constant-time single point multiplication on
761 typical inputs. Compiler support for the nonstandard type __uint128_t is
762 required to use this (present in gcc 4.4 and later, for 64-bit builds).
763 Code made available under Apache License version 2.0.
765 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
766 line to include this in your build of OpenSSL, and run "make depend" (or
767 "make update"). This enables the following EC_METHODs:
769 EC_GFp_nistp224_method()
770 EC_GFp_nistp256_method()
771 EC_GFp_nistp521_method()
773 EC_GROUP_new_by_curve_name() will automatically use these (while
774 EC_GROUP_new_curve_GFp() currently prefers the more flexible
776 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
778 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
779 all platforms. Move ssize_t definition from e_os.h to the public
780 header file e_os2.h as it now appears in public header file cms.h
783 *) New -sigopt option to the ca, req and x509 utilities. Additional
784 signature parameters can be passed using this option and in
788 *) Add RSA PSS signing function. This will generate and set the
789 appropriate AlgorithmIdentifiers for PSS based on those in the
790 corresponding EVP_MD_CTX structure. No application support yet.
793 *) Support for companion algorithm specific ASN1 signing routines.
794 New function ASN1_item_sign_ctx() signs a pre-initialised
795 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
796 the appropriate parameters.
799 *) Add new algorithm specific ASN1 verification initialisation function
800 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
801 handling will be the same no matter what EVP_PKEY_METHOD is used.
802 Add a PSS handler to support verification of PSS signatures: checked
803 against a number of sample certificates.
806 *) Add signature printing for PSS. Add PSS OIDs.
807 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
809 *) Add algorithm specific signature printing. An individual ASN1 method
810 can now print out signatures instead of the standard hex dump.
812 More complex signatures (e.g. PSS) can print out more meaningful
813 information. Include DSA version that prints out the signature
817 *) Password based recipient info support for CMS library: implementing
821 *) Split password based encryption into PBES2 and PBKDF2 functions. This
822 neatly separates the code into cipher and PBE sections and is required
823 for some algorithms that split PBES2 into separate pieces (such as
827 *) Session-handling fixes:
828 - Fix handling of connections that are resuming with a session ID,
829 but also support Session Tickets.
830 - Fix a bug that suppressed issuing of a new ticket if the client
831 presented a ticket with an expired session.
832 - Try to set the ticket lifetime hint to something reasonable.
833 - Make tickets shorter by excluding irrelevant information.
834 - On the client side, don't ignore renewed tickets.
835 [Adam Langley, Bodo Moeller (Google)]
837 *) Fix PSK session representation.
840 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
842 This work was sponsored by Intel.
845 *) Add GCM support to TLS library. Some custom code is needed to split
846 the IV between the fixed (from PRF) and explicit (from TLS record)
847 portions. This adds all GCM ciphersuites supported by RFC5288 and
848 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
849 add a special AESGCM string for GCM only.
852 *) Expand range of ctrls for AES GCM. Permit setting invocation
853 field on decrypt and retrieval of invocation field only on encrypt.
856 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
857 As required by RFC5289 these ciphersuites cannot be used if for
858 versions of TLS earlier than 1.2.
861 *) For FIPS capable OpenSSL interpret a NULL default public key method
862 as unset and return the appopriate default but do *not* set the default.
863 This means we can return the appopriate method in applications that
864 swicth between FIPS and non-FIPS modes.
867 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
868 ENGINE is used then we cannot handle that in the FIPS module so we
869 keep original code iff non-FIPS operations are allowed.
872 *) Add -attime option to openssl utilities.
873 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
875 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
878 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
879 FIPS EC methods unconditionally for now.
882 *) New build option no-ec2m to disable characteristic 2 code.
885 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
886 all cases can be covered as some introduce binary incompatibilities.
889 *) Redirect RSA operations to FIPS module including keygen,
890 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
893 *) Add similar low level API blocking to ciphers.
896 *) Low level digest APIs are not approved in FIPS mode: any attempt
897 to use these will cause a fatal error. Applications that *really* want
898 to use them can use the private_* version instead.
901 *) Redirect cipher operations to FIPS module for FIPS builds.
904 *) Redirect digest operations to FIPS module for FIPS builds.
907 *) Update build system to add "fips" flag which will link in fipscanister.o
908 for static and shared library builds embedding a signature if needed.
911 *) Output TLS supported curves in preference order instead of numerical
912 order. This is currently hardcoded for the highest order curves first.
913 This should be configurable so applications can judge speed vs strength.
916 *) Add TLS v1.2 server support for client authentication.
919 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
923 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
924 FIPS modules versions.
927 *) Add TLS v1.2 client side support for client authentication. Keep cache
928 of handshake records longer as we don't know the hash algorithm to use
929 until after the certificate request message is received.
932 *) Initial TLS v1.2 client support. Add a default signature algorithms
933 extension including all the algorithms we support. Parse new signature
934 format in client key exchange. Relax some ECC signing restrictions for
935 TLS v1.2 as indicated in RFC5246.
938 *) Add server support for TLS v1.2 signature algorithms extension. Switch
939 to new signature format when needed using client digest preference.
940 All server ciphersuites should now work correctly in TLS v1.2. No client
941 support yet and no support for client certificates.
944 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
945 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
946 ciphersuites. At present only RSA key exchange ciphersuites work with
947 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
948 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
949 and version checking.
952 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
953 with this defined it will not be affected by any changes to ssl internal
954 structures. Add several utility functions to allow openssl application
955 to work with OPENSSL_NO_SSL_INTERN defined.
959 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
961 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
964 *) Permit abbreviated handshakes when renegotiating using the function
965 SSL_renegotiate_abbreviated().
966 [Robin Seggelmann <seggelmann@fh-muenster.de>]
968 *) Add call to ENGINE_register_all_complete() to
969 ENGINE_load_builtin_engines(), so some implementations get used
970 automatically instead of needing explicit application support.
973 *) Add support for TLS key exporter as described in RFC5705.
974 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
976 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
977 a few changes are required:
979 Add SSL_OP_NO_TLSv1_1 flag.
981 Update version checking logic to handle version 1.1.
982 Add explicit IV handling (ported from DTLS code).
983 Add command line options to s_client/s_server.
986 Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
988 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
990 This addresses the flaw in CBC record processing discovered by
991 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
992 at: http://www.isg.rhul.ac.uk/tls/
994 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
995 Security Group at Royal Holloway, University of London
996 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
997 Emilia Käsper for the initial patch.
999 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1001 *) Return an error when checking OCSP signatures when key is NULL.
1002 This fixes a DoS attack. (CVE-2013-0166)
1005 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1006 the right response is stapled. Also change SSL_get_certificate()
1007 so it returns the certificate actually sent.
1008 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1009 (This is a backport)
1010 [Rob Stradling <rob.stradling@comodo.com>]
1012 *) Fix possible deadlock when decoding public keys.
1015 Changes between 1.0.0i and 1.0.0j [10 May 2012]
1017 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
1020 *) Sanity check record length before skipping explicit IV in DTLS
1023 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1024 fuzzing as a service testing platform.
1028 *) Initialise tkeylen properly when encrypting CMS messages.
1029 Thanks to Solar Designer of Openwall for reporting this issue.
1032 Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
1034 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1035 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1036 in CRYPTO_realloc_clean.
1038 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1039 issue and to Adam Langley <agl@chromium.org> for fixing it.
1041 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1043 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
1045 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1046 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1047 content decryption and always return the same error. Note: this attack
1048 needs on average 2^20 messages so it only affects automated senders. The
1049 old behaviour can be reenabled in the CMS code by setting the
1050 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1051 an MMA defence is not necessary.
1052 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1053 this issue. (CVE-2012-0884)
1056 *) Fix CVE-2011-4619: make sure we really are receiving a
1057 client hello before rejecting multiple SGC restarts. Thanks to
1058 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1061 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
1063 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1064 Thanks to Antonio Martin, Enterprise Secure Access Research and
1065 Development, Cisco Systems, Inc. for discovering this bug and
1066 preparing a fix. (CVE-2012-0050)
1069 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
1071 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1072 of the Vaudenay padding oracle attack on CBC mode encryption
1073 which enables an efficient plaintext recovery attack against
1074 the OpenSSL implementation of DTLS. Their attack exploits timing
1075 differences arising during decryption processing. A research
1076 paper describing this attack can be found at:
1077 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1078 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1079 Security Group at Royal Holloway, University of London
1080 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1081 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1082 for preparing the fix. (CVE-2011-4108)
1083 [Robin Seggelmann, Michael Tuexen]
1085 *) Clear bytes used for block padding of SSL 3.0 records.
1087 [Adam Langley (Google)]
1089 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1090 Kadianakis <desnacked@gmail.com> for discovering this issue and
1091 Adam Langley for preparing the fix. (CVE-2011-4619)
1092 [Adam Langley (Google)]
1094 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
1095 [Andrey Kulikov <amdeich@gmail.com>]
1097 *) Prevent malformed RFC3779 data triggering an assertion failure.
1098 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1099 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1100 [Rob Austein <sra@hactrn.net>]
1102 *) Improved PRNG seeding for VOS.
1103 [Paul Green <Paul.Green@stratus.com>]
1105 *) Fix ssl_ciph.c set-up race.
1106 [Adam Langley (Google)]
1108 *) Fix spurious failures in ecdsatest.c.
1109 [Emilia Käsper (Google)]
1111 *) Fix the BIO_f_buffer() implementation (which was mixing different
1112 interpretations of the '..._len' fields).
1113 [Adam Langley (Google)]
1115 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1116 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1117 threads won't reuse the same blinding coefficients.
1119 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1120 lock to call BN_BLINDING_invert_ex, and avoids one use of
1121 BN_BLINDING_update for each BN_BLINDING structure (previously,
1122 the last update always remained unused).
1123 [Emilia Käsper (Google)]
1125 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
1126 [Bob Buckholz (Google)]
1128 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
1130 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
1131 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
1132 [Kaspar Brand <ossl@velox.ch>]
1134 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1135 for multi-threaded use of ECDH. (CVE-2011-3210)
1136 [Adam Langley (Google)]
1138 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1141 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
1142 signature public key algorithm by using OID xref utilities instead.
1143 Before this you could only use some ECC ciphersuites with SHA1 only.
1146 *) Add protection against ECDSA timing attacks as mentioned in the paper
1147 by Billy Bob Brumley and Nicola Tuveri, see:
1149 http://eprint.iacr.org/2011/232.pdf
1151 [Billy Bob Brumley and Nicola Tuveri]
1153 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
1155 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1156 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1158 *) Fix bug in string printing code: if *any* escaping is enabled we must
1159 escape the escape character (backslash) or the resulting string is
1163 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
1165 *) Disable code workaround for ancient and obsolete Netscape browsers
1166 and servers: an attacker can use it in a ciphersuite downgrade attack.
1167 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1170 *) Fixed J-PAKE implementation error, originally discovered by
1171 Sebastien Martini, further info and confirmation from Stefan
1172 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1175 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
1177 *) Fix extension code to avoid race conditions which can result in a buffer
1178 overrun vulnerability: resumed sessions must not be modified as they can
1179 be shared by multiple threads. CVE-2010-3864
1182 *) Fix WIN32 build system to correctly link an ENGINE directory into
1186 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
1188 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
1190 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
1192 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
1194 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
1195 context. The operation can be customised via the ctrl mechanism in
1196 case ENGINEs want to include additional functionality.
1199 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
1202 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
1203 output hashes compatible with older versions of OpenSSL.
1204 [Willy Weisz <weisz@vcpc.univie.ac.at>]
1206 *) Fix compression algorithm handling: if resuming a session use the
1207 compression algorithm of the resumed session instead of determining
1208 it from client hello again. Don't allow server to change algorithm.
1211 *) Add load_crls() function to apps tidying load_certs() too. Add option
1212 to verify utility to allow additional CRLs to be included.
1215 *) Update OCSP request code to permit adding custom headers to the request:
1216 some responders need this.
1219 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
1221 [Julia Lawall <julia@diku.dk>]
1223 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
1224 needlessly dereferenced structures, used obsolete functions and
1225 didn't handle all updated verify codes correctly.
1228 *) Disable MD2 in the default configuration.
1231 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
1232 indicate the initial BIO being pushed or popped. This makes it possible
1233 to determine whether the BIO is the one explicitly called or as a result
1234 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
1235 it handles reference counts correctly and doesn't zero out the I/O bio
1236 when it is not being explicitly popped. WARNING: applications which
1237 included workarounds for the old buggy behaviour will need to be modified
1238 or they could free up already freed BIOs.
1241 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
1242 renaming to all platforms (within the 0.9.8 branch, this was
1243 done conditionally on Netware platforms to avoid a name clash).
1244 [Guenter <lists@gknw.net>]
1246 *) Add ECDHE and PSK support to DTLS.
1247 [Michael Tuexen <tuexen@fh-muenster.de>]
1249 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
1253 *) Add "missing" function EVP_MD_flags() (without this the only way to
1254 retrieve a digest flags is by accessing the structure directly. Update
1255 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
1256 or cipher is registered as in the "from" argument. Print out all
1257 registered digests in the dgst usage message instead of manually
1258 attempting to work them out.
1261 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
1262 this allows the use of compression and extensions. Change default cipher
1263 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
1264 by default unless an application cipher string requests it.
1267 *) Alter match criteria in PKCS12_parse(). It used to try to use local
1268 key ids to find matching certificates and keys but some PKCS#12 files
1269 don't follow the (somewhat unwritten) rules and this strategy fails.
1270 Now just gather all certificates together and the first private key
1271 then look for the first certificate that matches the key.
1274 *) Support use of registered digest and cipher names for dgst and cipher
1275 commands instead of having to add each one as a special case. So now
1282 openssl dgst -sha256 foo
1284 and this works for ENGINE based algorithms too.
1288 *) Update Gost ENGINE to support parameter files.
1289 [Victor B. Wagner <vitus@cryptocom.ru>]
1291 *) Support GeneralizedTime in ca utility.
1292 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
1294 *) Enhance the hash format used for certificate directory links. The new
1295 form uses the canonical encoding (meaning equivalent names will work
1296 even if they aren't identical) and uses SHA1 instead of MD5. This form
1297 is incompatible with the older format and as a result c_rehash should
1298 be used to rebuild symbolic links.
1301 *) Make PKCS#8 the default write format for private keys, replacing the
1302 traditional format. This form is standardised, more secure and doesn't
1303 include an implicit MD5 dependency.
1306 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
1307 committed to OpenSSL should pass this lot as a minimum.
1310 *) Add session ticket override functionality for use by EAP-FAST.
1311 [Jouni Malinen <j@w1.fi>]
1313 *) Modify HMAC functions to return a value. Since these can be implemented
1314 in an ENGINE errors can occur.
1317 *) Type-checked OBJ_bsearch_ex.
1320 *) Type-checked OBJ_bsearch. Also some constification necessitated
1321 by type-checking. Still to come: TXT_DB, bsearch(?),
1322 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
1326 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
1327 seconds to a tm structure directly, instead of going through OS
1328 specific date routines. This avoids any issues with OS routines such
1329 as the year 2038 bug. New *_adj() functions for ASN1 time structures
1330 and X509_time_adj_ex() to cover the extended range. The existing
1331 X509_time_adj() is still usable and will no longer have any date issues.
1334 *) Delta CRL support. New use deltas option which will attempt to locate
1335 and search any appropriate delta CRLs available.
1337 This work was sponsored by Google.
1340 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
1341 code and add additional score elements. Validate alternate CRL paths
1342 as part of the CRL checking and indicate a new error "CRL path validation
1343 error" in this case. Applications wanting additional details can use
1344 the verify callback and check the new "parent" field. If this is not
1345 NULL CRL path validation is taking place. Existing applications wont
1346 see this because it requires extended CRL support which is off by
1349 This work was sponsored by Google.
1352 *) Support for freshest CRL extension.
1354 This work was sponsored by Google.
1357 *) Initial indirect CRL support. Currently only supported in the CRLs
1358 passed directly and not via lookup. Process certificate issuer
1359 CRL entry extension and lookup CRL entries by bother issuer name
1360 and serial number. Check and process CRL issuer entry in IDP extension.
1362 This work was sponsored by Google.
1365 *) Add support for distinct certificate and CRL paths. The CRL issuer
1366 certificate is validated separately in this case. Only enabled if
1367 an extended CRL support flag is set: this flag will enable additional
1368 CRL functionality in future.
1370 This work was sponsored by Google.
1373 *) Add support for policy mappings extension.
1375 This work was sponsored by Google.
1378 *) Fixes to pathlength constraint, self issued certificate handling,
1379 policy processing to align with RFC3280 and PKITS tests.
1381 This work was sponsored by Google.
1384 *) Support for name constraints certificate extension. DN, email, DNS
1385 and URI types are currently supported.
1387 This work was sponsored by Google.
1390 *) To cater for systems that provide a pointer-based thread ID rather
1391 than numeric, deprecate the current numeric thread ID mechanism and
1392 replace it with a structure and associated callback type. This
1393 mechanism allows a numeric "hash" to be extracted from a thread ID in
1394 either case, and on platforms where pointers are larger than 'long',
1395 mixing is done to help ensure the numeric 'hash' is usable even if it
1396 can't be guaranteed unique. The default mechanism is to use "&errno"
1397 as a pointer-based thread ID to distinguish between threads.
1399 Applications that want to provide their own thread IDs should now use
1400 CRYPTO_THREADID_set_callback() to register a callback that will call
1401 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
1403 Note that ERR_remove_state() is now deprecated, because it is tied
1404 to the assumption that thread IDs are numeric. ERR_remove_state(0)
1405 to free the current thread's error state should be replaced by
1406 ERR_remove_thread_state(NULL).
1408 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
1409 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
1410 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
1411 application was previously providing a numeric thread callback that
1412 was inappropriate for distinguishing threads, then uniqueness might
1413 have been obtained with &errno that happened immediately in the
1414 intermediate development versions of OpenSSL; this is no longer the
1415 case, the numeric thread callback will now override the automatic use
1417 [Geoff Thorpe, with help from Bodo Moeller]
1419 *) Initial support for different CRL issuing certificates. This covers a
1420 simple case where the self issued certificates in the chain exist and
1421 the real CRL issuer is higher in the existing chain.
1423 This work was sponsored by Google.
1426 *) Removed effectively defunct crypto/store from the build.
1429 *) Revamp of STACK to provide stronger type-checking. Still to come:
1430 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
1431 ASN1_STRING, CONF_VALUE.
1434 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
1435 RAM on SSL connections. This option can save about 34k per idle SSL.
1438 *) Revamp of LHASH to provide stronger type-checking. Still to come:
1439 STACK, TXT_DB, bsearch, qsort.
1442 *) Initial support for Cryptographic Message Syntax (aka CMS) based
1443 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
1444 support for data, signedData, compressedData, digestedData and
1445 encryptedData, envelopedData types included. Scripts to check against
1446 RFC4134 examples draft and interop and consistency checks of many
1447 content types and variants.
1450 *) Add options to enc utility to support use of zlib compression BIO.
1453 *) Extend mk1mf to support importing of options and assembly language
1454 files from Configure script, currently only included in VC-WIN32.
1455 The assembly language rules can now optionally generate the source
1456 files from the associated perl scripts.
1459 *) Implement remaining functionality needed to support GOST ciphersuites.
1460 Interop testing has been performed using CryptoPro implementations.
1461 [Victor B. Wagner <vitus@cryptocom.ru>]
1463 *) s390x assembler pack.
1466 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
1470 *) Implement Opaque PRF Input TLS extension as specified in
1471 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
1472 official specification yet and no extension type assignment by
1473 IANA exists, this extension (for now) will have to be explicitly
1474 enabled when building OpenSSL by providing the extension number
1475 to use. For example, specify an option
1477 -DTLSEXT_TYPE_opaque_prf_input=0x9527
1479 to the "config" or "Configure" script to enable the extension,
1480 assuming extension number 0x9527 (which is a completely arbitrary
1481 and unofficial assignment based on the MD5 hash of the Internet
1482 Draft). Note that by doing so, you potentially lose
1483 interoperability with other TLS implementations since these might
1484 be using the same extension number for other purposes.
1486 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
1487 opaque PRF input value to use in the handshake. This will create
1488 an interal copy of the length-'len' string at 'src', and will
1489 return non-zero for success.
1491 To get more control and flexibility, provide a callback function
1494 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
1495 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
1499 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
1502 Callback function 'cb' will be called in handshakes, and is
1503 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
1504 Argument 'arg' is for application purposes (the value as given to
1505 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
1506 be provided to the callback function). The callback function
1507 has to return non-zero to report success: usually 1 to use opaque
1508 PRF input just if possible, or 2 to enforce use of the opaque PRF
1509 input. In the latter case, the library will abort the handshake
1510 if opaque PRF input is not successfully negotiated.
1512 Arguments 'peerinput' and 'len' given to the callback function
1513 will always be NULL and 0 in the case of a client. A server will
1514 see the client's opaque PRF input through these variables if
1515 available (NULL and 0 otherwise). Note that if the server
1516 provides an opaque PRF input, the length must be the same as the
1517 length of the client's opaque PRF input.
1519 Note that the callback function will only be called when creating
1520 a new session (session resumption can resume whatever was
1521 previously negotiated), and will not be called in SSL 2.0
1522 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
1523 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
1524 for applications that need to enforce opaque PRF input.
1528 *) Update ssl code to support digests other than SHA1+MD5 for handshake
1531 [Victor B. Wagner <vitus@cryptocom.ru>]
1533 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1534 RFC4507bis. The encrypted ticket format is an encrypted encoded
1535 SSL_SESSION structure, that way new session features are automatically
1538 If a client application caches session in an SSL_SESSION structure
1539 support is transparent because tickets are now stored in the encoded
1542 The SSL_CTX structure automatically generates keys for ticket
1543 protection in servers so again support should be possible
1544 with no application modification.
1546 If a client or server wishes to disable RFC4507 support then the option
1547 SSL_OP_NO_TICKET can be set.
1549 Add a TLS extension debugging callback to allow the contents of any client
1550 or server extensions to be examined.
1552 This work was sponsored by Google.
1555 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1556 OpenSSL should now compile cleanly on gcc 4.2
1557 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1559 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1560 support including streaming MAC support: this is required for GOST
1561 ciphersuite support.
1562 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1564 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1565 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1566 to output in BER and PEM format.
1569 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1570 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1571 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1572 ENGINE support for HMAC keys which are unextractable. New -mac and
1573 -macopt options to dgst utility.
1576 *) New option -sigopt to dgst utility. Update dgst to use
1577 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1578 alternative signing paramaters such as X9.31 or PSS in the dgst
1582 *) Change ssl_cipher_apply_rule(), the internal function that does
1583 the work each time a ciphersuite string requests enabling
1584 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1585 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1586 the order of disabled ciphersuites such that those ciphersuites
1587 that most recently went from enabled to disabled not only stay
1588 in order with respect to each other, but also have higher priority
1589 than other disabled ciphersuites the next time ciphersuites are
1592 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1593 the same ciphersuites as with "HIGH" alone, but in a specific
1594 order where the PSK ciphersuites come first (since they are the
1595 most recently disabled ciphersuites when "HIGH" is parsed).
1597 Also, change ssl_create_cipher_list() (using this new
1598 funcionality) such that between otherwise identical
1599 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
1603 *) Change ssl_create_cipher_list() so that it automatically
1604 arranges the ciphersuites in reasonable order before starting
1605 to process the rule string. Thus, the definition for "DEFAULT"
1606 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
1607 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
1608 This makes it much easier to arrive at a reasonable default order
1609 in applications for which anonymous ciphers are OK (meaning
1610 that you can't actually use DEFAULT).
1611 [Bodo Moeller; suggested by Victor Duchovni]
1613 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
1614 processing) into multiple integers instead of setting
1615 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
1616 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
1617 (These masks as well as the individual bit definitions are hidden
1618 away into the non-exported interface ssl/ssl_locl.h, so this
1619 change to the definition of the SSL_CIPHER structure shouldn't
1620 affect applications.) This give us more bits for each of these
1621 categories, so there is no longer a need to coagulate AES128 and
1622 AES256 into a single algorithm bit, and to coagulate Camellia128
1623 and Camellia256 into a single algorithm bit, which has led to all
1626 Thus, among other things, the kludge introduced in 0.9.7m and
1627 0.9.8e for masking out AES256 independently of AES128 or masking
1628 out Camellia256 independently of AES256 is not needed here in 0.9.9.
1630 With the change, we also introduce new ciphersuite aliases that
1631 so far were missing: "AES128", "AES256", "CAMELLIA128", and
1635 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
1636 Use the leftmost N bytes of the signature input if the input is
1637 larger than the prime q (with N being the size in bytes of q).
1640 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1641 it yet and it is largely untested.
1644 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1647 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1648 some compilers (gcc 4.2 and later) reject their use. Safestack is
1649 reimplemented. Update ASN1 to avoid use of legacy functions.
1652 *) Win32/64 targets are linked with Winsock2.
1655 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1656 to external functions. This can be used to increase CRL handling
1657 efficiency especially when CRLs are very large by (for example) storing
1658 the CRL revoked certificates in a database.
1661 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1662 new CRLs added to a directory can be used. New command line option
1663 -verify_return_error to s_client and s_server. This causes real errors
1664 to be returned by the verify callback instead of carrying on no matter
1665 what. This reflects the way a "real world" verify callback would behave.
1668 *) GOST engine, supporting several GOST algorithms and public key formats.
1669 Kindly donated by Cryptocom.
1672 *) Partial support for Issuing Distribution Point CRL extension. CRLs
1673 partitioned by DP are handled but no indirect CRL or reason partitioning
1674 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
1675 selected via a scoring technique which handles IDP and AKID in CRLs.
1678 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1679 will ultimately be used for all verify operations: this will remove the
1680 X509_STORE dependency on certificate verification and allow alternative
1681 lookup methods. X509_STORE based implementations of these two callbacks.
1684 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1685 Modify get_crl() to find a valid (unexpired) CRL if possible.
1688 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1689 this would be called X509_CRL_cmp() but that name is already used by
1690 a function that just compares CRL issuer names. Cache several CRL
1691 extensions in X509_CRL structure and cache CRLDP in X509.
1694 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1695 this maps equivalent X509_NAME structures into a consistent structure.
1696 Name comparison can then be performed rapidly using memcmp().
1699 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1703 *) Allow digests to supply their own micalg string for S/MIME type using
1704 the ctrl EVP_MD_CTRL_MICALG.
1707 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1708 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1709 ctrl. It can then customise the structure before and/or after signing
1713 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1714 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1715 to free up any added signature OIDs.
1718 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1719 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1720 digest and cipher tables. New options added to openssl utility:
1721 list-message-digest-algorithms and list-cipher-algorithms.
1724 *) Change the array representation of binary polynomials: the list
1725 of degrees of non-zero coefficients is now terminated with -1.
1726 Previously it was terminated with 0, which was also part of the
1727 value; thus, the array representation was not applicable to
1728 polynomials where t^0 has coefficient zero. This change makes
1729 the array representation useful in a more general context.
1732 *) Various modifications and fixes to SSL/TLS cipher string
1733 handling. For ECC, the code now distinguishes between fixed ECDH
1734 with RSA certificates on the one hand and with ECDSA certificates
1735 on the other hand, since these are separate ciphersuites. The
1736 unused code for Fortezza ciphersuites has been removed.
1738 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1739 (not "ECDHE"). For consistency with the code for DH
1740 certificates, use of ECDH certificates is now considered ECDH
1741 authentication, not RSA or ECDSA authentication (the latter is
1742 merely the CA's signing algorithm and not actively used in the
1745 The temporary ciphersuite alias "ECCdraft" is no longer
1746 available, and ECC ciphersuites are no longer excluded from "ALL"
1747 and "DEFAULT". The following aliases now exist for RFC 4492
1748 ciphersuites, most of these by analogy with the DH case:
1750 kECDHr - ECDH cert, signed with RSA
1751 kECDHe - ECDH cert, signed with ECDSA
1752 kECDH - ECDH cert (signed with either RSA or ECDSA)
1753 kEECDH - ephemeral ECDH
1754 ECDH - ECDH cert or ephemeral ECDH
1760 AECDH - anonymous ECDH
1761 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1765 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1766 Use correct micalg parameters depending on digest(s) in signed message.
1769 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1770 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1773 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1774 an engine to register a method. Add ENGINE lookups for methods and
1775 functional reference processing.
1778 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1779 EVP_{Sign,Verify}* which allow an application to customise the signature
1783 *) New -resign option to smime utility. This adds one or more signers
1784 to an existing PKCS#7 signedData structure. Also -md option to use an
1785 alternative message digest algorithm for signing.
1788 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1789 create PKCS7 structures containing multiple signers. Update smime
1790 application to support multiple signers.
1793 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1797 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1798 Reorganize PBE internals to lookup from a static table using NIDs,
1799 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1800 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1801 PRF which will be automatically used with PBES2.
1804 *) Replace the algorithm specific calls to generate keys in "req" with the
1808 *) Update PKCS#7 enveloped data routines to use new API. This is now
1809 supported by any public key method supporting the encrypt operation. A
1810 ctrl is added to allow the public key algorithm to examine or modify
1811 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1815 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1816 a default digest type to use. In most cases this will be SHA1 but some
1817 algorithms (such as GOST) need to specify an alternative digest. The
1818 return value indicates how strong the prefernce is 1 means optional and
1819 2 is mandatory (that is it is the only supported type). Modify
1820 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1821 use the default md. Update openssl utilities to use the default digest
1822 type for signing if it is not explicitly indicated.
1825 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1826 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1827 signing method from the key type. This effectively removes the link
1828 between digests and public key types.
1831 *) Add an OID cross reference table and utility functions. Its purpose is to
1832 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1833 rsaEncryption. This will allow some of the algorithm specific hackery
1834 needed to use the correct OID to be removed.
1837 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1838 structures for PKCS7_sign(). They are now set up by the relevant public
1842 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1845 *) Add support for key derivation (agreement) in the API, DH method and
1849 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1850 public and private key formats. As a side effect these add additional
1851 command line functionality not previously available: DSA signatures can be
1852 generated and verified using pkeyutl and DH key support and generation in
1857 [Oliver Tappe <zooey@hirschkaefer.de>]
1859 *) New make target "install_html_docs" installs HTML renditions of the
1861 [Oliver Tappe <zooey@hirschkaefer.de>]
1863 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1864 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1865 support key and parameter generation and add initial key generation
1866 functionality for RSA.
1869 *) Add functions for main EVP_PKEY_method operations. The undocumented
1870 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1871 EVP_PKEY_{encrypt,decrypt}_old.
1874 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1875 key API, doesn't do much yet.
1878 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1879 public key algorithms. New option to openssl utility:
1880 "list-public-key-algorithms" to print out info.
1883 *) Implement the Supported Elliptic Curves Extension for
1884 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1887 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1888 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1891 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1892 utilities such as rsa, dsa, dsaparam etc except they process any key
1896 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1897 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1898 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1902 *) Initial support for pluggable public key ASN1.
1903 De-spaghettify the public key ASN1 handling. Move public and private
1904 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1905 algorithm specific handling to a single module within the relevant
1906 algorithm directory. Add functions to allow (near) opaque processing
1907 of public and private key structures.
1910 *) Implement the Supported Point Formats Extension for
1911 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1914 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1915 for the psk identity [hint] and the psk callback functions to the
1916 SSL_SESSION, SSL and SSL_CTX structure.
1919 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1923 SSL_CTX_use_psk_identity_hint
1924 SSL_get_psk_identity_hint
1925 SSL_get_psk_identity
1926 SSL_use_psk_identity_hint
1928 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1930 *) Add RFC 3161 compliant time stamp request creation, response generation
1931 and response verification functionality.
1932 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1934 *) Add initial support for TLS extensions, specifically for the server_name
1935 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1936 have new members for a host name. The SSL data structure has an
1937 additional member SSL_CTX *initial_ctx so that new sessions can be
1938 stored in that context to allow for session resumption, even after the
1939 SSL has been switched to a new SSL_CTX in reaction to a client's
1940 server_name extension.
1942 New functions (subject to change):
1944 SSL_get_servername()
1945 SSL_get_servername_type()
1948 New CTRL codes and macros (subject to change):
1950 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1951 - SSL_CTX_set_tlsext_servername_callback()
1952 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1953 - SSL_CTX_set_tlsext_servername_arg()
1954 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1956 openssl s_client has a new '-servername ...' option.
1958 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1959 '-key2 ...', '-servername_fatal' (subject to change). This allows
1960 testing the HostName extension for a specific single host name ('-cert'
1961 and '-key' remain fallbacks for handshakes without HostName
1962 negotiation). If the unrecogninzed_name alert has to be sent, this by
1963 default is a warning; it becomes fatal with the '-servername_fatal'
1966 [Peter Sylvester, Remy Allais, Christophe Renou]
1968 *) Whirlpool hash implementation is added.
1971 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1972 bn(64,32). Because of instruction set limitations it doesn't have
1973 any negative impact on performance. This was done mostly in order
1974 to make it possible to share assembler modules, such as bn_mul_mont
1975 implementations, between 32- and 64-bit builds without hassle.
1978 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1979 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1983 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1984 dedicated Montgomery multiplication procedure, is introduced.
1985 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1986 "64-bit" performance on certain 32-bit targets.
1989 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1990 in SSL structures. New SSL ctrl to set maximum send fragment size.
1991 Save memory by seeting the I/O buffer sizes dynamically instead of
1992 using the maximum available value.
1995 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1996 in addition to the text details.
1999 *) Very, very preliminary EXPERIMENTAL support for printing of general
2000 ASN1 structures. This currently produces rather ugly output and doesn't
2001 handle several customised structures at all.
2004 *) Integrated support for PVK file format and some related formats such
2005 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
2006 these in the 'rsa' and 'dsa' utilities.
2009 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
2012 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
2013 place for the (very old) "NETSCAPE" format certificates which are now
2014 handled using new ASN1 code equivalents.
2017 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
2018 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
2019 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
2022 *) Modify CRL distribution points extension code to print out previously
2023 unsupported fields. Enhance extension setting code to allow setting of
2027 *) Add print and set support for Issuing Distribution Point CRL extension.
2030 *) Change 'Configure' script to enable Camellia by default.
2033 Changes between 0.9.8x and 0.9.8y [5 Feb 2013]
2035 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2037 This addresses the flaw in CBC record processing discovered by
2038 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2039 at: http://www.isg.rhul.ac.uk/tls/
2041 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2042 Security Group at Royal Holloway, University of London
2043 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2044 Emilia Käsper for the initial patch.
2046 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2048 *) Return an error when checking OCSP signatures when key is NULL.
2049 This fixes a DoS attack. (CVE-2013-0166)
2052 *) Return an error when checking OCSP signatures when key is NULL.
2053 This fixes a DoS attack. (CVE-2013-0166)
2056 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2057 the right response is stapled. Also change SSL_get_certificate()
2058 so it returns the certificate actually sent.
2059 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2060 (This is a backport)
2061 [Rob Stradling <rob.stradling@comodo.com>]
2063 *) Fix possible deadlock when decoding public keys.
2066 Changes between 0.9.8w and 0.9.8x [10 May 2012]
2068 *) Sanity check record length before skipping explicit IV in DTLS
2071 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2072 fuzzing as a service testing platform.
2076 *) Initialise tkeylen properly when encrypting CMS messages.
2077 Thanks to Solar Designer of Openwall for reporting this issue.
2080 Changes between 0.9.8v and 0.9.8w [23 Apr 2012]
2082 *) The fix for CVE-2012-2110 did not take into account that the
2083 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an
2084 int in OpenSSL 0.9.8, making it still vulnerable. Fix by
2085 rejecting negative len parameter. (CVE-2012-2131)
2086 [Tomas Hoger <thoger@redhat.com>]
2088 Changes between 0.9.8u and 0.9.8v [19 Apr 2012]
2090 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2091 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2092 in CRYPTO_realloc_clean.
2094 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2095 issue and to Adam Langley <agl@chromium.org> for fixing it.
2097 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2099 Changes between 0.9.8t and 0.9.8u [12 Mar 2012]
2101 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2102 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2103 content decryption and always return the same error. Note: this attack
2104 needs on average 2^20 messages so it only affects automated senders. The
2105 old behaviour can be reenabled in the CMS code by setting the
2106 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2107 an MMA defence is not necessary.
2108 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2109 this issue. (CVE-2012-0884)
2112 *) Fix CVE-2011-4619: make sure we really are receiving a
2113 client hello before rejecting multiple SGC restarts. Thanks to
2114 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2117 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
2119 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2120 Thanks to Antonio Martin, Enterprise Secure Access Research and
2121 Development, Cisco Systems, Inc. for discovering this bug and
2122 preparing a fix. (CVE-2012-0050)
2125 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
2127 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2128 of the Vaudenay padding oracle attack on CBC mode encryption
2129 which enables an efficient plaintext recovery attack against
2130 the OpenSSL implementation of DTLS. Their attack exploits timing
2131 differences arising during decryption processing. A research
2132 paper describing this attack can be found at:
2133 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2134 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2135 Security Group at Royal Holloway, University of London
2136 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2137 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2138 for preparing the fix. (CVE-2011-4108)
2139 [Robin Seggelmann, Michael Tuexen]
2141 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
2142 [Ben Laurie, Kasper <ekasper@google.com>]
2144 *) Clear bytes used for block padding of SSL 3.0 records.
2146 [Adam Langley (Google)]
2148 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2149 Kadianakis <desnacked@gmail.com> for discovering this issue and
2150 Adam Langley for preparing the fix. (CVE-2011-4619)
2151 [Adam Langley (Google)]
2153 *) Prevent malformed RFC3779 data triggering an assertion failure.
2154 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2155 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2156 [Rob Austein <sra@hactrn.net>]
2158 *) Fix ssl_ciph.c set-up race.
2159 [Adam Langley (Google)]
2161 *) Fix spurious failures in ecdsatest.c.
2162 [Emilia Käsper (Google)]
2164 *) Fix the BIO_f_buffer() implementation (which was mixing different
2165 interpretations of the '..._len' fields).
2166 [Adam Langley (Google)]
2168 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2169 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2170 threads won't reuse the same blinding coefficients.
2172 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2173 lock to call BN_BLINDING_invert_ex, and avoids one use of
2174 BN_BLINDING_update for each BN_BLINDING structure (previously,
2175 the last update always remained unused).
2176 [Emilia Käsper (Google)]
2178 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2179 for multi-threaded use of ECDH.
2180 [Adam Langley (Google)]
2182 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2185 *) Add protection against ECDSA timing attacks as mentioned in the paper
2186 by Billy Bob Brumley and Nicola Tuveri, see:
2188 http://eprint.iacr.org/2011/232.pdf
2190 [Billy Bob Brumley and Nicola Tuveri]
2192 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
2194 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2195 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2197 *) Fix bug in string printing code: if *any* escaping is enabled we must
2198 escape the escape character (backslash) or the resulting string is
2202 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
2204 *) Disable code workaround for ancient and obsolete Netscape browsers
2205 and servers: an attacker can use it in a ciphersuite downgrade attack.
2206 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2209 *) Fixed J-PAKE implementation error, originally discovered by
2210 Sebastien Martini, further info and confirmation from Stefan
2211 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2214 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
2216 *) Fix extension code to avoid race conditions which can result in a buffer
2217 overrun vulnerability: resumed sessions must not be modified as they can
2218 be shared by multiple threads. CVE-2010-3864
2221 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
2224 *) Don't reencode certificate when calculating signature: cache and use
2225 the original encoding instead. This makes signature verification of
2226 some broken encodings work correctly.
2229 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
2230 is also one of the inputs.
2231 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
2233 *) Don't repeatedly append PBE algorithms to table if they already exist.
2234 Sort table on each new add. This effectively makes the table read only
2235 after all algorithms are added and subsequent calls to PKCS12_pbe_add
2239 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
2241 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
2244 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
2245 access or freeing data twice (CVE-2010-0742)
2246 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
2248 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
2249 common in certificates and some applications which only call
2250 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
2254 Reduce copying into .apps and .test in makevms.com
2255 Don't try to use blank CA certificate in CA.com
2256 Allow use of C files from original directories in maketests.com
2257 [Steven M. Schweda" <sms@antinode.info>]
2259 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2261 *) When rejecting SSL/TLS records due to an incorrect version number, never
2262 update s->server with a new major version number. As of
2263 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2264 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2265 the previous behavior could result in a read attempt at NULL when
2266 receiving specific incorrect SSL/TLS records once record payload
2267 protection is active. (CVE-2010-0740)
2268 [Bodo Moeller, Adam Langley <agl@chromium.org>]
2270 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
2271 could be crashed if the relevant tables were not present (e.g. chrooted).
2272 [Tomas Hoger <thoger@redhat.com>]
2274 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
2276 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
2277 [Martin Olsson, Neel Mehta]
2279 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
2280 accommodate for stack sorting, always a write lock!).
2283 *) On some versions of WIN32 Heap32Next is very slow. This can cause
2284 excessive delays in the RAND_poll(): over a minute. As a workaround
2285 include a time check in the inner Heap32Next loop too.
2288 *) The code that handled flushing of data in SSL/TLS originally used the
2289 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
2290 the problem outlined in PR#1949. The fix suggested there however can
2291 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
2292 of Apache). So instead simplify the code to flush unconditionally.
2293 This should be fine since flushing with no data to flush is a no op.
2296 *) Handle TLS versions 2.0 and later properly and correctly use the
2297 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
2298 off ancient servers have a habit of sticking around for a while...
2301 *) Modify compression code so it frees up structures without using the
2302 ex_data callbacks. This works around a problem where some applications
2303 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
2304 restarting) then use compression (e.g. SSL with compression) later.
2305 This results in significant per-connection memory leaks and
2306 has caused some security issues including CVE-2008-1678 and
2310 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
2311 change when encrypting or decrypting.
2314 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
2315 connect and renegotiate with servers which do not support RI.
2316 Until RI is more widely deployed this option is enabled by default.
2319 *) Add "missing" ssl ctrls to clear options and mode.
2322 *) If client attempts to renegotiate and doesn't support RI respond with
2323 a no_renegotiation alert as required by RFC5746. Some renegotiating
2324 TLS clients will continue a connection gracefully when they receive
2325 the alert. Unfortunately OpenSSL mishandled this alert and would hang
2326 waiting for a server hello which it will never receive. Now we treat a
2327 received no_renegotiation alert as a fatal error. This is because
2328 applications requesting a renegotiation might well expect it to succeed
2329 and would have no code in place to handle the server denying it so the
2330 only safe thing to do is to terminate the connection.
2333 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
2334 peer supports secure renegotiation and 0 otherwise. Print out peer
2335 renegotiation support in s_client/s_server.
2338 *) Replace the highly broken and deprecated SPKAC certification method with
2339 the updated NID creation version. This should correctly handle UTF8.
2342 *) Implement RFC5746. Re-enable renegotiation but require the extension
2343 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2344 turns out to be a bad idea. It has been replaced by
2345 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2346 SSL_CTX_set_options(). This is really not recommended unless you
2347 know what you are doing.
2348 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2350 *) Fixes to stateless session resumption handling. Use initial_ctx when
2351 issuing and attempting to decrypt tickets in case it has changed during
2352 servername handling. Use a non-zero length session ID when attempting
2353 stateless session resumption: this makes it possible to determine if
2354 a resumption has occurred immediately after receiving server hello
2355 (several places in OpenSSL subtly assume this) instead of later in
2359 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2360 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
2361 fixes for a few places where the return code is not checked
2363 [Julia Lawall <julia@diku.dk>]
2365 *) Add --strict-warnings option to Configure script to include devteam
2366 warnings in other configurations.
2369 *) Add support for --libdir option and LIBDIR variable in makefiles. This
2370 makes it possible to install openssl libraries in locations which
2371 have names other than "lib", for example "/usr/lib64" which some
2373 [Steve Henson, based on patch from Jeremy Utley]
2375 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
2376 X690 8.9.12 and can produce some misleading textual output of OIDs.
2377 [Steve Henson, reported by Dan Kaminsky]
2379 *) Delete MD2 from algorithm tables. This follows the recommendation in
2380 several standards that it is not used in new applications due to
2381 several cryptographic weaknesses. For binary compatibility reasons
2382 the MD2 API is still compiled in by default.
2385 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
2389 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
2390 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
2392 [Guenter <lists@gknw.net>]
2394 *) Fix the server certificate chain building code to use X509_verify_cert(),
2395 it used to have an ad-hoc builder which was unable to cope with anything
2396 other than a simple chain.
2397 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
2399 *) Don't check self signed certificate signatures in X509_verify_cert()
2400 by default (a flag can override this): it just wastes time without
2401 adding any security. As a useful side effect self signed root CAs
2402 with non-FIPS digests are now usable in FIPS mode.
2405 *) In dtls1_process_out_of_seq_message() the check if the current message
2406 is already buffered was missing. For every new message was memory
2407 allocated, allowing an attacker to perform an denial of service attack
2408 with sending out of seq handshake messages until there is no memory
2409 left. Additionally every future messege was buffered, even if the
2410 sequence number made no sense and would be part of another handshake.
2411 So only messages with sequence numbers less than 10 in advance will be
2412 buffered. (CVE-2009-1378)
2413 [Robin Seggelmann, discovered by Daniel Mentz]
2415 *) Records are buffered if they arrive with a future epoch to be
2416 processed after finishing the corresponding handshake. There is
2417 currently no limitation to this buffer allowing an attacker to perform
2418 a DOS attack with sending records with future epochs until there is no
2419 memory left. This patch adds the pqueue_size() function to detemine
2420 the size of a buffer and limits the record buffer to 100 entries.
2422 [Robin Seggelmann, discovered by Daniel Mentz]
2424 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2425 parent structure is freed. (CVE-2009-1379)
2428 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2429 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2432 [Ilya O. <vrghost@gmail.com>]
2434 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2436 *) Disable renegotiation completely - this fixes a severe security
2437 problem (CVE-2009-3555) at the cost of breaking all
2438 renegotiation. Renegotiation can be re-enabled by setting
2439 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2440 run-time. This is really not recommended unless you know what
2444 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2446 *) Don't set val to NULL when freeing up structures, it is freed up by
2447 underlying code. If sizeof(void *) > sizeof(long) this can result in
2448 zeroing past the valid field. (CVE-2009-0789)
2449 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2451 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2452 checked correctly. This would allow some invalid signed attributes to
2453 appear to verify correctly. (CVE-2009-0591)
2454 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2456 *) Reject UniversalString and BMPString types with invalid lengths. This
2457 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2458 a legal length. (CVE-2009-0590)
2461 *) Set S/MIME signing as the default purpose rather than setting it
2462 unconditionally. This allows applications to override it at the store
2466 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2467 to handle some structures.
2470 *) Improve efficiency of mem_gets: don't search whole buffer each time
2472 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2474 *) New -hex option for openssl rand.
2477 *) Print out UTF8String and NumericString when parsing ASN1.
2480 *) Support NumericString type for name components.
2483 *) Allow CC in the environment to override the automatically chosen
2484 compiler. Note that nothing is done to ensure flags work with the
2488 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2490 *) Properly check EVP_VerifyFinal() and similar return values
2492 [Ben Laurie, Bodo Moeller, Google Security Team]
2494 *) Enable TLS extensions by default.
2497 *) Allow the CHIL engine to be loaded, whether the application is
2498 multithreaded or not. (This does not release the developer from the
2499 obligation to set up the dynamic locking callbacks.)
2500 [Sander Temme <sander@temme.net>]
2502 *) Use correct exit code if there is an error in dgst command.
2503 [Steve Henson; problem pointed out by Roland Dirlewanger]
2505 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2506 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2509 *) Add experimental JPAKE support, including demo authentication in
2510 s_client and s_server.
2513 *) Set the comparison function in v3_addr_canonize().
2514 [Rob Austein <sra@hactrn.net>]
2516 *) Add support for XMPP STARTTLS in s_client.
2517 [Philip Paeps <philip@freebsd.org>]
2519 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2520 to ensure that even with this option, only ciphersuites in the
2521 server's preference list will be accepted. (Note that the option
2522 applies only when resuming a session, so the earlier behavior was
2523 just about the algorithm choice for symmetric cryptography.)
2526 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2528 *) Fix NULL pointer dereference if a DTLS server received
2529 ChangeCipherSpec as first record (CVE-2009-1386).
2532 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
2533 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2536 *) The fix in 0.9.8c that supposedly got rid of unsafe
2537 double-checked locking was incomplete for RSA blinding,
2538 addressing just one layer of what turns out to have been
2539 doubly unsafe triple-checked locking.
2541 So now fix this for real by retiring the MONT_HELPER macro
2542 in crypto/rsa/rsa_eay.c.
2544 [Bodo Moeller; problem pointed out by Marius Schilder]
2546 *) Various precautionary measures:
2548 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2550 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2551 (NB: This would require knowledge of the secret session ticket key
2552 to exploit, in which case you'd be SOL either way.)
2554 - Change bn_nist.c so that it will properly handle input BIGNUMs
2555 outside the expected range.
2557 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2560 [Neel Mehta, Bodo Moeller]
2562 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2563 the load fails. Useful for distros.
2564 [Ben Laurie and the FreeBSD team]
2566 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2569 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2572 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2574 This work was sponsored by Logica.
2577 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2578 keystores. Support for SSL/TLS client authentication too.
2579 Not compiled unless enable-capieng specified to Configure.
2581 This work was sponsored by Logica.
2584 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2585 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2586 attribute creation routines such as certifcate requests and PKCS#12
2590 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2592 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2593 handshake which could lead to a cilent crash as found using the
2594 Codenomicon TLS test suite (CVE-2008-1672)
2595 [Steve Henson, Mark Cox]
2597 *) Fix double free in TLS server name extensions which could lead to
2598 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
2601 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
2603 Clear the error queue to ensure that error entries left from
2604 older function calls do not interfere with the correct operation.
2605 [Lutz Jaenicke, Erik de Castro Lopo]
2607 *) Remove root CA certificates of commercial CAs:
2609 The OpenSSL project does not recommend any specific CA and does not
2610 have any policy with respect to including or excluding any CA.
2611 Therefore it does not make any sense to ship an arbitrary selection
2612 of root CA certificates with the OpenSSL software.
2615 *) RSA OAEP patches to fix two separate invalid memory reads.
2616 The first one involves inputs when 'lzero' is greater than
2617 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
2618 before the beginning of from). The second one involves inputs where
2619 the 'db' section contains nothing but zeroes (there is a one-byte
2620 invalid read after the end of 'db').
2621 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2623 *) Partial backport from 0.9.9-dev:
2625 Introduce bn_mul_mont (dedicated Montgomery multiplication
2626 procedure) as a candidate for BIGNUM assembler implementation.
2627 While 0.9.9-dev uses assembler for various architectures, only
2628 x86_64 is available by default here in the 0.9.8 branch, and
2629 32-bit x86 is available through a compile-time setting.
2631 To try the 32-bit x86 assembler implementation, use Configure
2632 option "enable-montasm" (which exists only for this backport).
2634 As "enable-montasm" for 32-bit x86 disclaims code stability
2635 anyway, in this constellation we activate additional code
2636 backported from 0.9.9-dev for further performance improvements,
2637 namely BN_from_montgomery_word. (To enable this otherwise,
2638 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
2640 [Andy Polyakov (backport partially by Bodo Moeller)]
2642 *) Add TLS session ticket callback. This allows an application to set
2643 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
2644 values. This is useful for key rollover for example where several key
2645 sets may exist with different names.
2648 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
2649 This was broken until now in 0.9.8 releases, such that the only way
2650 a registered ENGINE could be used (assuming it initialises
2651 successfully on the host) was to explicitly set it as the default
2652 for the relevant algorithms. This is in contradiction with 0.9.7
2653 behaviour and the documentation. With this fix, when an ENGINE is
2654 registered into a given algorithm's table of implementations, the
2655 'uptodate' flag is reset so that auto-discovery will be used next
2656 time a new context for that algorithm attempts to select an
2658 [Ian Lister (tweaked by Geoff Thorpe)]
2660 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
2661 implemention in the following ways:
2663 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
2666 Lack of BER streaming support means one pass streaming processing is
2667 only supported if data is detached: setting the streaming flag is
2668 ignored for embedded content.
2670 CMS support is disabled by default and must be explicitly enabled
2671 with the enable-cms configuration option.
2674 *) Update the GMP engine glue to do direct copies between BIGNUM and
2675 mpz_t when openssl and GMP use the same limb size. Otherwise the
2676 existing "conversion via a text string export" trick is still used.
2677 [Paul Sheer <paulsheer@gmail.com>]
2679 *) Zlib compression BIO. This is a filter BIO which compressed and
2680 uncompresses any data passed through it.
2683 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
2684 RFC3394 compatible AES key wrapping.
2687 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
2688 sets string data without copying. X509_ALGOR_set0() and
2689 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
2690 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
2691 from an X509_ATTRIBUTE structure optionally checking it occurs only
2692 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
2696 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
2697 to get the expected BN_FLG_CONSTTIME behavior.
2698 [Bodo Moeller (Google)]
2702 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
2703 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
2704 - added some more tests to do_tests.pl
2705 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
2706 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
2707 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
2708 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
2709 - various changes to netware.pl to enable gcc-cross builds on Win32
2711 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
2712 - various changes to fix missing prototype warnings
2713 - fixed x86nasm.pl to create correct asm files for NASM COFF output
2714 - added AES, WHIRLPOOL and CPUID assembler code to build files
2715 - added missing AES assembler make rules to mk1mf.pl
2716 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
2717 [Guenter Knauf <eflash@gmx.net>]
2719 *) Implement certificate status request TLS extension defined in RFC3546.
2720 A client can set the appropriate parameters and receive the encoded
2721 OCSP response via a callback. A server can query the supplied parameters
2722 and set the encoded OCSP response in the callback. Add simplified examples
2723 to s_client and s_server.
2726 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
2728 *) Fix various bugs:
2729 + Binary incompatibility of ssl_ctx_st structure
2730 + DTLS interoperation with non-compliant servers
2731 + Don't call get_session_cb() without proposed session
2732 + Fix ia64 assembler code
2733 [Andy Polyakov, Steve Henson]
2735 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
2737 *) DTLS Handshake overhaul. There were longstanding issues with
2738 OpenSSL DTLS implementation, which were making it impossible for
2739 RFC 4347 compliant client to communicate with OpenSSL server.
2740 Unfortunately just fixing these incompatibilities would "cut off"
2741 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
2742 server keeps tolerating non RFC compliant syntax. The opposite is
2743 not true, 0.9.8f client can not communicate with earlier server.
2744 This update even addresses CVE-2007-4995.
2747 *) Changes to avoid need for function casts in OpenSSL: some compilers
2748 (gcc 4.2 and later) reject their use.
2749 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
2752 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2753 RFC4507bis. The encrypted ticket format is an encrypted encoded
2754 SSL_SESSION structure, that way new session features are automatically
2757 If a client application caches session in an SSL_SESSION structure
2758 support is transparent because tickets are now stored in the encoded
2761 The SSL_CTX structure automatically generates keys for ticket
2762 protection in servers so again support should be possible
2763 with no application modification.
2765 If a client or server wishes to disable RFC4507 support then the option
2766 SSL_OP_NO_TICKET can be set.
2768 Add a TLS extension debugging callback to allow the contents of any client
2769 or server extensions to be examined.
2771 This work was sponsored by Google.
2774 *) Add initial support for TLS extensions, specifically for the server_name
2775 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2776 have new members for a host name. The SSL data structure has an
2777 additional member SSL_CTX *initial_ctx so that new sessions can be
2778 stored in that context to allow for session resumption, even after the
2779 SSL has been switched to a new SSL_CTX in reaction to a client's
2780 server_name extension.
2782 New functions (subject to change):
2784 SSL_get_servername()
2785 SSL_get_servername_type()
2788 New CTRL codes and macros (subject to change):
2790 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2791 - SSL_CTX_set_tlsext_servername_callback()
2792 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2793 - SSL_CTX_set_tlsext_servername_arg()
2794 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2796 openssl s_client has a new '-servername ...' option.
2798 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2799 '-key2 ...', '-servername_fatal' (subject to change). This allows
2800 testing the HostName extension for a specific single host name ('-cert'
2801 and '-key' remain fallbacks for handshakes without HostName
2802 negotiation). If the unrecogninzed_name alert has to be sent, this by
2803 default is a warning; it becomes fatal with the '-servername_fatal'
2806 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2808 *) Add AES and SSE2 assembly language support to VC++ build.
2811 *) Mitigate attack on final subtraction in Montgomery reduction.
2814 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2815 (which previously caused an internal error).
2818 *) Squeeze another 10% out of IGE mode when in != out.
2821 *) AES IGE mode speedup.
2822 [Dean Gaudet (Google)]
2824 *) Add the Korean symmetric 128-bit cipher SEED (see
2825 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2826 add SEED ciphersuites from RFC 4162:
2828 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2829 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2830 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2831 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2833 To minimize changes between patchlevels in the OpenSSL 0.9.8
2834 series, SEED remains excluded from compilation unless OpenSSL
2835 is configured with 'enable-seed'.
2836 [KISA, Bodo Moeller]
2838 *) Mitigate branch prediction attacks, which can be practical if a
2839 single processor is shared, allowing a spy process to extract
2840 information. For detailed background information, see
2841 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2842 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2843 and Necessary Software Countermeasures"). The core of the change
2844 are new versions BN_div_no_branch() and
2845 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2846 respectively, which are slower, but avoid the security-relevant
2847 conditional branches. These are automatically called by BN_div()
2848 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2849 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2850 remove a conditional branch.
2852 BN_FLG_CONSTTIME is the new name for the previous
2853 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2854 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2855 in the exponent causes BN_mod_exp_mont() to use the alternative
2856 implementation in BN_mod_exp_mont_consttime().) The old name
2857 remains as a deprecated alias.
2859 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2860 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2861 constant-time implementations for more than just exponentiation.
2862 Here too the old name is kept as a deprecated alias.
2864 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2865 the BN_BLINDING structure gets an independent copy of the
2866 modulus. This means that the previous "BIGNUM *m" argument to
2867 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2868 essentially becomes "const BIGNUM *m", although we can't actually
2869 change this in the header file before 0.9.9. It allows
2870 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2871 enable BN_FLG_CONSTTIME.
2873 [Matthew D Wood (Intel Corp)]
2875 *) In the SSL/TLS server implementation, be strict about session ID
2876 context matching (which matters if an application uses a single
2877 external cache for different purposes). Previously,
2878 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2879 set. This did ensure strict client verification, but meant that,
2880 with applications using a single external cache for quite
2881 different requirements, clients could circumvent ciphersuite
2882 restrictions for a given session ID context by starting a session
2883 in a different context.
2886 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2887 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2888 authentication-only ciphersuites.
2891 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2892 not complete and could lead to a possible single byte overflow
2893 (CVE-2007-5135) [Ben Laurie]
2895 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2897 *) Since AES128 and AES256 (and similarly Camellia128 and
2898 Camellia256) share a single mask bit in the logic of
2899 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2900 kludge to work properly if AES128 is available and AES256 isn't
2901 (or if Camellia128 is available and Camellia256 isn't).
2904 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2905 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2906 When a point or a seed is encoded in a BIT STRING, we need to
2907 prevent the removal of trailing zero bits to get the proper DER
2908 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2909 of a NamedBitList, for which trailing 0 bits need to be removed.)
2912 *) Have SSL/TLS server implementation tolerate "mismatched" record
2913 protocol version while receiving ClientHello even if the
2914 ClientHello is fragmented. (The server can't insist on the
2915 particular protocol version it has chosen before the ServerHello
2916 message has informed the client about his choice.)
2919 *) Add RFC 3779 support.
2920 [Rob Austein for ARIN, Ben Laurie]
2922 *) Load error codes if they are not already present instead of using a
2923 static variable. This allows them to be cleanly unloaded and reloaded.
2924 Improve header file function name parsing.
2927 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2928 or CAPABILITY handshake as required by RFCs.
2931 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2933 *) Introduce limits to prevent malicious keys being able to
2934 cause a denial of service. (CVE-2006-2940)
2935 [Steve Henson, Bodo Moeller]
2937 *) Fix ASN.1 parsing of certain invalid structures that can result
2938 in a denial of service. (CVE-2006-2937) [Steve Henson]
2940 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2941 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2943 *) Fix SSL client code which could crash if connecting to a
2944 malicious SSLv2 server. (CVE-2006-4343)
2945 [Tavis Ormandy and Will Drewry, Google Security Team]
2947 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2948 match only those. Before that, "AES256-SHA" would be interpreted
2949 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2950 the same strength classification in 0.9.7h) as we currently only
2951 have a single AES bit in the ciphersuite description bitmap.
2952 That change, however, also applied to ciphersuite strings such as
2953 "RC4-MD5" that intentionally matched multiple ciphersuites --
2954 namely, SSL 2.0 ciphersuites in addition to the more common ones
2955 from SSL 3.0/TLS 1.0.
2957 So we change the selection algorithm again: Naming an explicit
2958 ciphersuite selects this one ciphersuite, and any other similar
2959 ciphersuite (same bitmap) from *other* protocol versions.
2960 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2961 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2963 Since SSL 2.0 does not have any ciphersuites for which the
2964 128/256 bit distinction would be relevant, this works for now.
2965 The proper fix will be to use different bits for AES128 and
2966 AES256, which would have avoided the problems from the beginning;
2967 however, bits are scarce, so we can only do this in a new release
2968 (not just a patchlevel) when we can change the SSL_CIPHER
2969 definition to split the single 'unsigned long mask' bitmap into
2970 multiple values to extend the available space.
2974 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2976 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2977 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2979 *) Add AES IGE and biIGE modes.
2982 *) Change the Unix randomness entropy gathering to use poll() when
2983 possible instead of select(), since the latter has some
2984 undesirable limitations.
2985 [Darryl Miles via Richard Levitte and Bodo Moeller]
2987 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2988 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2989 cannot be implicitly activated as part of, e.g., the "AES" alias.
2990 However, please upgrade to OpenSSL 0.9.9[-dev] for
2991 non-experimental use of the ECC ciphersuites to get TLS extension
2992 support, which is required for curve and point format negotiation
2993 to avoid potential handshake problems.
2996 *) Disable rogue ciphersuites:
2998 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2999 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3000 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3002 The latter two were purportedly from
3003 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3006 Also deactivate the remaining ciphersuites from
3007 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3008 unofficial, and the ID has long expired.
3011 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3012 dual-core machines) and other potential thread-safety issues.
3015 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3016 versions), which is now available for royalty-free use
3017 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3018 Also, add Camellia TLS ciphersuites from RFC 4132.
3020 To minimize changes between patchlevels in the OpenSSL 0.9.8
3021 series, Camellia remains excluded from compilation unless OpenSSL
3022 is configured with 'enable-camellia'.
3025 *) Disable the padding bug check when compression is in use. The padding
3026 bug check assumes the first packet is of even length, this is not
3027 necessarily true if compresssion is enabled and can result in false
3028 positives causing handshake failure. The actual bug test is ancient
3029 code so it is hoped that implementations will either have fixed it by
3030 now or any which still have the bug do not support compression.
3033 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3035 *) When applying a cipher rule check to see if string match is an explicit
3036 cipher suite and only match that one cipher suite if it is.
3039 *) Link in manifests for VC++ if needed.
3040 [Austin Ziegler <halostatue@gmail.com>]
3042 *) Update support for ECC-based TLS ciphersuites according to
3043 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3044 TLS extensions, which are supported starting with the 0.9.9
3045 branch, not in the OpenSSL 0.9.8 branch).
3048 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3049 opaque EVP_CIPHER_CTX handling.
3052 *) Fixes and enhancements to zlib compression code. We now only use
3053 "zlib1.dll" and use the default __cdecl calling convention on Win32
3054 to conform with the standards mentioned here:
3055 http://www.zlib.net/DLL_FAQ.txt
3056 Static zlib linking now works on Windows and the new --with-zlib-include
3057 --with-zlib-lib options to Configure can be used to supply the location
3058 of the headers and library. Gracefully handle case where zlib library
3062 *) Several fixes and enhancements to the OID generation code. The old code
3063 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3064 handle numbers larger than ULONG_MAX, truncated printing and had a
3065 non standard OBJ_obj2txt() behaviour.
3068 *) Add support for building of engines under engine/ as shared libraries
3069 under VC++ build system.
3072 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3073 Hopefully, we will not see any false combination of paths any more.
3076 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3078 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3079 (part of SSL_OP_ALL). This option used to disable the
3080 countermeasure against man-in-the-middle protocol-version
3081 rollback in the SSL 2.0 server implementation, which is a bad
3082 idea. (CVE-2005-2969)
3084 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3085 for Information Security, National Institute of Advanced Industrial
3086 Science and Technology [AIST], Japan)]
3088 *) Add two function to clear and return the verify parameter flags.
3091 *) Keep cipherlists sorted in the source instead of sorting them at
3092 runtime, thus removing the need for a lock.
3095 *) Avoid some small subgroup attacks in Diffie-Hellman.
3096 [Nick Mathewson and Ben Laurie]
3098 *) Add functions for well-known primes.
3101 *) Extended Windows CE support.
3102 [Satoshi Nakamura and Andy Polyakov]
3104 *) Initialize SSL_METHOD structures at compile time instead of during
3105 runtime, thus removing the need for a lock.
3108 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3109 attempting to decrypt each encrypted key in turn. Add support to
3113 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3115 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3118 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3121 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3122 key into the same file any more.
3125 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3128 *) Add -utf8 command line and config file option to 'ca'.
3129 [Stefan <stf@udoma.org]
3131 *) Removed the macro des_crypt(), as it seems to conflict with some
3132 libraries. Use DES_crypt().
3135 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3136 involves renaming the source and generated shared-libs for
3137 both. The engines will accept the corrected or legacy ids
3138 ('ncipher' and '4758_cca' respectively) when binding. NB,
3139 this only applies when building 'shared'.
3140 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3142 *) Add attribute functions to EVP_PKEY structure. Modify
3143 PKCS12_create() to recognize a CSP name attribute and
3144 use it. Make -CSP option work again in pkcs12 utility.
3147 *) Add new functionality to the bn blinding code:
3148 - automatic re-creation of the BN_BLINDING parameters after
3149 a fixed number of uses (currently 32)
3150 - add new function for parameter creation
3151 - introduce flags to control the update behaviour of the
3152 BN_BLINDING parameters
3153 - hide BN_BLINDING structure
3154 Add a second BN_BLINDING slot to the RSA structure to improve
3155 performance when a single RSA object is shared among several
3159 *) Add support for DTLS.
3160 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3162 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3163 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3166 *) Remove buggy and incompletet DH cert support from
3167 ssl/ssl_rsa.c and ssl/s3_both.c
3170 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3171 the apps/openssl applications.
3174 *) Compile clean with "-Wall -Wmissing-prototypes
3175 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
3176 DEBUG_SAFESTACK must also be set.
3179 *) Change ./Configure so that certain algorithms can be disabled by default.
3180 The new counterpiece to "no-xxx" is "enable-xxx".
3182 The patented RC5 and MDC2 algorithms will now be disabled unless
3183 "enable-rc5" and "enable-mdc2", respectively, are specified.
3185 (IDEA remains enabled despite being patented. This is because IDEA
3186 is frequently required for interoperability, and there is no license
3187 fee for non-commercial use. As before, "no-idea" can be used to
3188 avoid this algorithm.)
3192 *) Add processing of proxy certificates (see RFC 3820). This work was
3193 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
3194 EGEE (Enabling Grids for E-science in Europe).
3197 *) RC4 performance overhaul on modern architectures/implementations, such
3198 as Intel P4, IA-64 and AMD64.
3201 *) New utility extract-section.pl. This can be used specify an alternative
3202 section number in a pod file instead of having to treat each file as
3203 a separate case in Makefile. This can be done by adding two lines to the
3206 =for comment openssl_section:XXX
3208 The blank line is mandatory.
3212 *) New arguments -certform, -keyform and -pass for s_client and s_server
3213 to allow alternative format key and certificate files and passphrase
3217 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
3218 update associated structures and add various utility functions.
3220 Add new policy related verify parameters, include policy checking in
3221 standard verify code. Enhance 'smime' application with extra parameters
3222 to support policy checking and print out.
3225 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
3226 Nehemiah processors. These extensions support AES encryption in hardware
3227 as well as RNG (though RNG support is currently disabled).
3228 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
3230 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
3233 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
3234 [Andy Polyakov and a number of other people]
3236 *) Improved PowerPC platform support. Most notably BIGNUM assembler
3237 implementation contributed by IBM.
3238 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
3240 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
3241 exponent rather than 'unsigned long'. There is a corresponding change to
3242 the new 'rsa_keygen' element of the RSA_METHOD structure.
3243 [Jelte Jansen, Geoff Thorpe]
3245 *) Functionality for creating the initial serial number file is now
3246 moved from CA.pl to the 'ca' utility with a new option -create_serial.
3248 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
3249 number file to 1, which is bound to cause problems. To avoid
3250 the problems while respecting compatibility between different 0.9.7
3251 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
3252 CA.pl for serial number initialization. With the new release 0.9.8,
3253 we can fix the problem directly in the 'ca' utility.)
3256 *) Reduced header interdepencies by declaring more opaque objects in
3257 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
3258 give fewer recursive includes, which could break lazy source code - so
3259 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
3260 developers should define this symbol when building and using openssl to
3261 ensure they track the recommended behaviour, interfaces, [etc], but
3262 backwards-compatible behaviour prevails when this isn't defined.
3265 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
3268 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
3269 This will generate a random key of the appropriate length based on the
3270 cipher context. The EVP_CIPHER can provide its own random key generation
3271 routine to support keys of a specific form. This is used in the des and
3272 3des routines to generate a key of the correct parity. Update S/MIME
3273 code to use new functions and hence generate correct parity DES keys.
3274 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
3275 valid (weak or incorrect parity).
3278 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
3279 as looking them up. This is useful when the verified structure may contain
3280 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
3281 present unless the new PKCS7_NO_CRL flag is asserted.
3284 *) Extend ASN1 oid configuration module. It now additionally accepts the
3287 shortName = some long name, 1.2.3.4
3290 *) Reimplemented the BN_CTX implementation. There is now no more static
3291 limitation on the number of variables it can handle nor the depth of the
3292 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
3293 information can now expand as required, and rather than having a single
3294 static array of bignums, BN_CTX now uses a linked-list of such arrays
3295 allowing it to expand on demand whilst maintaining the usefulness of
3296 BN_CTX's "bundling".
3299 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
3300 to allow all RSA operations to function using a single BN_CTX.
3303 *) Preliminary support for certificate policy evaluation and checking. This
3304 is initially intended to pass the tests outlined in "Conformance Testing
3305 of Relying Party Client Certificate Path Processing Logic" v1.07.
3308 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
3309 remained unused and not that useful. A variety of other little bignum
3310 tweaks and fixes have also been made continuing on from the audit (see
3314 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
3315 associated ASN1, EVP and SSL functions and old ASN1 macros.
3318 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
3319 and this should never fail. So the return value from the use of
3320 BN_set_word() (which can fail due to needless expansion) is now deprecated;
3321 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
3324 *) BN_CTX_get() should return zero-valued bignums, providing the same
3325 initialised value as BN_new().
3326 [Geoff Thorpe, suggested by Ulf Möller]
3328 *) Support for inhibitAnyPolicy certificate extension.
3331 *) An audit of the BIGNUM code is underway, for which debugging code is
3332 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
3333 is considered valid when processing BIGNUMs, and causes execution to
3334 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
3335 further steps are taken to deliberately pollute unused data in BIGNUM
3336 structures to try and expose faulty code further on. For now, openssl will
3337 (in its default mode of operation) continue to tolerate the inconsistent
3338 forms that it has tolerated in the past, but authors and packagers should
3339 consider trying openssl and their own applications when compiled with
3340 these debugging symbols defined. It will help highlight potential bugs in
3341 their own code, and will improve the test coverage for OpenSSL itself. At
3342 some point, these tighter rules will become openssl's default to improve
3343 maintainability, though the assert()s and other overheads will remain only
3344 in debugging configurations. See bn.h for more details.
3345 [Geoff Thorpe, Nils Larsch, Ulf Möller]
3347 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
3348 that can only be obtained through BN_CTX_new() (which implicitly
3349 initialises it). The presence of this function only made it possible
3350 to overwrite an existing structure (and cause memory leaks).
3353 *) Because of the callback-based approach for implementing LHASH as a
3354 template type, lh_insert() adds opaque objects to hash-tables and
3355 lh_doall() or lh_doall_arg() are typically used with a destructor callback
3356 to clean up those corresponding objects before destroying the hash table
3357 (and losing the object pointers). So some over-zealous constifications in
3358 LHASH have been relaxed so that lh_insert() does not take (nor store) the
3359 objects as "const" and the lh_doall[_arg] callback wrappers are not
3360 prototyped to have "const" restrictions on the object pointers they are
3361 given (and so aren't required to cast them away any more).
3364 *) The tmdiff.h API was so ugly and minimal that our own timing utility
3365 (speed) prefers to use its own implementation. The two implementations
3366 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
3367 its object type properly exposed (MS_TM) instead of casting to/from "char
3368 *". This may still change yet if someone realises MS_TM and "ms_time_***"
3369 aren't necessarily the greatest nomenclatures - but this is what was used
3370 internally to the implementation so I've used that for now.
3373 *) Ensure that deprecated functions do not get compiled when
3374 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
3375 the self-tests were still using deprecated key-generation functions so
3376 these have been updated also.
3379 *) Reorganise PKCS#7 code to separate the digest location functionality
3380 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
3381 New function PKCS7_set_digest() to set the digest type for PKCS#7
3382 digestedData type. Add additional code to correctly generate the
3383 digestedData type and add support for this type in PKCS7 initialization
3387 *) New function PKCS7_set0_type_other() this initializes a PKCS7
3388 structure of type "other".
3391 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
3392 sure the loop does correctly stop and breaking ("division by zero")
3393 modulus operations are not performed. The (pre-generated) prime
3394 table crypto/bn/bn_prime.h was already correct, but it could not be
3395 re-generated on some platforms because of the "division by zero"
3396 situation in the script.
3397 [Ralf S. Engelschall]
3399 *) Update support for ECC-based TLS ciphersuites according to
3400 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
3401 SHA-1 now is only used for "small" curves (where the
3402 representation of a field element takes up to 24 bytes); for
3403 larger curves, the field element resulting from ECDH is directly
3404 used as premaster secret.
3405 [Douglas Stebila (Sun Microsystems Laboratories)]
3407 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
3408 curve secp160r1 to the tests.
3409 [Douglas Stebila (Sun Microsystems Laboratories)]
3411 *) Add the possibility to load symbols globally with DSO.
3412 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
3414 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
3415 control of the error stack.
3418 *) Add support for STORE in ENGINE.
3421 *) Add the STORE type. The intention is to provide a common interface
3422 to certificate and key stores, be they simple file-based stores, or
3423 HSM-type store, or LDAP stores, or...
3424 NOTE: The code is currently UNTESTED and isn't really used anywhere.
3427 *) Add a generic structure called OPENSSL_ITEM. This can be used to
3428 pass a list of arguments to any function as well as provide a way
3429 for a function to pass data back to the caller.
3432 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
3433 works like BUF_strdup() but can be used to duplicate a portion of
3434 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
3438 *) Add the function sk_find_ex() which works like sk_find(), but will
3439 return an index to an element even if an exact match couldn't be
3440 found. The index is guaranteed to point at the element where the
3441 searched-for key would be inserted to preserve sorting order.
3444 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
3445 takes an extra flags argument for optional functionality. Currently,
3446 the following flags are defined:
3448 OBJ_BSEARCH_VALUE_ON_NOMATCH
3449 This one gets OBJ_bsearch_ex() to return a pointer to the first
3450 element where the comparing function returns a negative or zero
3453 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
3454 This one gets OBJ_bsearch_ex() to return a pointer to the first
3455 element where the comparing function returns zero. This is useful
3456 if there are more than one element where the comparing function
3460 *) Make it possible to create self-signed certificates with 'openssl ca'
3461 in such a way that the self-signed certificate becomes part of the
3462 CA database and uses the same mechanisms for serial number generation
3463 as all other certificate signing. The new flag '-selfsign' enables
3464 this functionality. Adapt CA.sh and CA.pl.in.
3467 *) Add functionality to check the public key of a certificate request
3468 against a given private. This is useful to check that a certificate
3469 request can be signed by that key (self-signing).
3472 *) Make it possible to have multiple active certificates with the same
3473 subject in the CA index file. This is done only if the keyword
3474 'unique_subject' is set to 'no' in the main CA section (default
3475 if 'CA_default') of the configuration file. The value is saved
3476 with the database itself in a separate index attribute file,
3477 named like the index file with '.attr' appended to the name.
3480 *) Generate muti valued AVAs using '+' notation in config files for
3484 *) Support for nameConstraints certificate extension.
3487 *) Support for policyConstraints certificate extension.
3490 *) Support for policyMappings certificate extension.
3493 *) Make sure the default DSA_METHOD implementation only uses its
3494 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3495 and change its own handlers to be NULL so as to remove unnecessary
3496 indirection. This lets alternative implementations fallback to the
3497 default implementation more easily.
3500 *) Support for directoryName in GeneralName related extensions
3504 *) Make it possible to link applications using Makefile.shared.
3505 Make that possible even when linking against static libraries!
3508 *) Support for single pass processing for S/MIME signing. This now
3509 means that S/MIME signing can be done from a pipe, in addition
3510 cleartext signing (multipart/signed type) is effectively streaming
3511 and the signed data does not need to be all held in memory.
3513 This is done with a new flag PKCS7_STREAM. When this flag is set
3514 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3515 is done after the data is output (and digests calculated) in
3516 SMIME_write_PKCS7().
3519 *) Add full support for -rpath/-R, both in shared libraries and
3520 applications, at least on the platforms where it's known how
3524 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3525 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3526 will now compute a table of multiples of the generator that
3527 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3528 faster (notably in the case of a single point multiplication,
3529 scalar * generator).
3530 [Nils Larsch, Bodo Moeller]
3532 *) IPv6 support for certificate extensions. The various extensions
3533 which use the IP:a.b.c.d can now take IPv6 addresses using the
3534 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3538 *) Added an ENGINE that implements RSA by performing private key
3539 exponentiations with the GMP library. The conversions to and from
3540 GMP's mpz_t format aren't optimised nor are any montgomery forms
3541 cached, and on x86 it appears OpenSSL's own performance has caught up.
3542 However there are likely to be other architectures where GMP could
3543 provide a boost. This ENGINE is not built in by default, but it can be
3544 specified at Configure time and should be accompanied by the necessary
3545 linker additions, eg;
3546 ./config -DOPENSSL_USE_GMP -lgmp
3549 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3550 testing availability of engines with "-t" - the old behaviour is
3551 produced by increasing the feature's verbosity with "-tt".
3554 *) ECDSA routines: under certain error conditions uninitialized BN objects
3555 could be freed. Solution: make sure initialization is performed early
3556 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3560 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3561 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3562 software implementations. For DSA and DH, parameter generation can
3563 also be overriden by providing the appropriate method callbacks.
3566 *) Change the "progress" mechanism used in key-generation and
3567 primality testing to functions that take a new BN_GENCB pointer in
3568 place of callback/argument pairs. The new API functions have "_ex"
3569 postfixes and the older functions are reimplemented as wrappers for
3570 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3571 declarations of the old functions to help (graceful) attempts to
3572 migrate to the new functions. Also, the new key-generation API
3573 functions operate on a caller-supplied key-structure and return
3574 success/failure rather than returning a key or NULL - this is to
3575 help make "keygen" another member function of RSA_METHOD etc.
3577 Example for using the new callback interface:
3579 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3583 BN_GENCB_set(&my_cb, my_callback, my_arg);
3585 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3586 /* For the meaning of a, b in calls to my_callback(), see the
3587 * documentation of the function that calls the callback.
3588 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3589 * my_callback should return 1 if it wants BN_is_prime_ex()
3590 * to continue, or 0 to stop.
3595 *) Change the ZLIB compression method to be stateful, and make it
3596 available to TLS with the number defined in
3597 draft-ietf-tls-compression-04.txt.
3600 *) Add the ASN.1 structures and functions for CertificatePair, which
3601 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
3603 CertificatePair ::= SEQUENCE {
3604 forward [0] Certificate OPTIONAL,
3605 reverse [1] Certificate OPTIONAL,
3606 -- at least one of the pair shall be present -- }
3608 Also implement the PEM functions to read and write certificate
3609 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
3611 This needed to be defined, mostly for the sake of the LDAP
3612 attribute crossCertificatePair, but may prove useful elsewhere as
3616 *) Make it possible to inhibit symlinking of shared libraries in
3617 Makefile.shared, for Cygwin's sake.
3620 *) Extend the BIGNUM API by creating a function
3621 void BN_set_negative(BIGNUM *a, int neg);
3622 and a macro that behave like
3623 int BN_is_negative(const BIGNUM *a);
3625 to avoid the need to access 'a->neg' directly in applications.
3628 *) Implement fast modular reduction for pseudo-Mersenne primes
3629 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
3630 EC_GROUP_new_curve_GFp() will now automatically use this
3632 [Nils Larsch <nla@trustcenter.de>]
3634 *) Add new lock type (CRYPTO_LOCK_BN).
3637 *) Change the ENGINE framework to automatically load engines
3638 dynamically from specific directories unless they could be
3639 found to already be built in or loaded. Move all the
3640 current engines except for the cryptodev one to a new
3642 The engines in engines/ are built as shared libraries if
3643 the "shared" options was given to ./Configure or ./config.
3644 Otherwise, they are inserted in libcrypto.a.
3645 /usr/local/ssl/engines is the default directory for dynamic
3646 engines, but that can be overriden at configure time through
3647 the usual use of --prefix and/or --openssldir, and at run
3648 time with the environment variable OPENSSL_ENGINES.
3649 [Geoff Thorpe and Richard Levitte]
3651 *) Add Makefile.shared, a helper makefile to build shared
3652 libraries. Addapt Makefile.org.
3655 *) Add version info to Win32 DLLs.
3656 [Peter 'Luna' Runestig" <peter@runestig.com>]
3658 *) Add new 'medium level' PKCS#12 API. Certificates and keys
3659 can be added using this API to created arbitrary PKCS#12
3660 files while avoiding the low level API.
3662 New options to PKCS12_create(), key or cert can be NULL and
3663 will then be omitted from the output file. The encryption
3664 algorithm NIDs can be set to -1 for no encryption, the mac
3665 iteration count can be set to 0 to omit the mac.
3667 Enhance pkcs12 utility by making the -nokeys and -nocerts
3668 options work when creating a PKCS#12 file. New option -nomac
3669 to omit the mac, NONE can be set for an encryption algorithm.
3670 New code is modified to use the enhanced PKCS12_create()
3671 instead of the low level API.
3674 *) Extend ASN1 encoder to support indefinite length constructed
3675 encoding. This can output sequences tags and octet strings in
3676 this form. Modify pk7_asn1.c to support indefinite length
3677 encoding. This is experimental and needs additional code to
3678 be useful, such as an ASN1 bio and some enhanced streaming
3681 Extend template encode functionality so that tagging is passed
3682 down to the template encoder.
3685 *) Let 'openssl req' fail if an argument to '-newkey' is not
3686 recognized instead of using RSA as a default.
3689 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
3690 As these are not official, they are not included in "ALL";
3691 the "ECCdraft" ciphersuite group alias can be used to select them.
3692 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
3694 *) Add ECDH engine support.
3695 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
3697 *) Add ECDH in new directory crypto/ecdh/.
3698 [Douglas Stebila (Sun Microsystems Laboratories)]
3700 *) Let BN_rand_range() abort with an error after 100 iterations
3701 without success (which indicates a broken PRNG).
3704 *) Change BN_mod_sqrt() so that it verifies that the input value
3705 is really the square of the return value. (Previously,
3706 BN_mod_sqrt would show GIGO behaviour.)
3709 *) Add named elliptic curves over binary fields from X9.62, SECG,
3710 and WAP/WTLS; add OIDs that were still missing.
3712 [Sheueling Chang Shantz and Douglas Stebila
3713 (Sun Microsystems Laboratories)]
3715 *) Extend the EC library for elliptic curves over binary fields
3716 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
3719 EC_GF2m_simple_method
3723 EC_GROUP_new_curve_GF2m
3724 EC_GROUP_set_curve_GF2m
3725 EC_GROUP_get_curve_GF2m
3726 EC_POINT_set_affine_coordinates_GF2m
3727 EC_POINT_get_affine_coordinates_GF2m
3728 EC_POINT_set_compressed_coordinates_GF2m
3730 Point compression for binary fields is disabled by default for
3731 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
3734 As binary polynomials are represented as BIGNUMs, various members
3735 of the EC_GROUP and EC_POINT data structures can be shared
3736 between the implementations for prime fields and binary fields;
3737 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
3738 are essentially identical to their ..._GFp counterparts.
3739 (For simplicity, the '..._GFp' prefix has been dropped from
3740 various internal method names.)
3742 An internal 'field_div' method (similar to 'field_mul' and
3743 'field_sqr') has been added; this is used only for binary fields.
3745 [Sheueling Chang Shantz and Douglas Stebila
3746 (Sun Microsystems Laboratories)]
3748 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
3749 through methods ('mul', 'precompute_mult').
3751 The generic implementations (now internally called 'ec_wNAF_mul'
3752 and 'ec_wNAF_precomputed_mult') remain the default if these
3753 methods are undefined.
3755 [Sheueling Chang Shantz and Douglas Stebila
3756 (Sun Microsystems Laboratories)]
3758 *) New function EC_GROUP_get_degree, which is defined through
3759 EC_METHOD. For curves over prime fields, this returns the bit
3760 length of the modulus.
3762 [Sheueling Chang Shantz and Douglas Stebila
3763 (Sun Microsystems Laboratories)]
3765 *) New functions EC_GROUP_dup, EC_POINT_dup.
3766 (These simply call ..._new and ..._copy).
3768 [Sheueling Chang Shantz and Douglas Stebila
3769 (Sun Microsystems Laboratories)]
3771 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3772 Polynomials are represented as BIGNUMs (where the sign bit is not
3773 used) in the following functions [macros]:
3776 BN_GF2m_sub [= BN_GF2m_add]
3777 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3778 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3779 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3781 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3782 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3783 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3784 BN_GF2m_cmp [= BN_ucmp]
3786 (Note that only the 'mod' functions are actually for fields GF(2^m).
3787 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3789 For some functions, an the irreducible polynomial defining a
3790 field can be given as an 'unsigned int[]' with strictly
3791 decreasing elements giving the indices of those bits that are set;
3792 i.e., p[] represents the polynomial
3793 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3795 p[0] > p[1] > ... > p[k] = 0.
3796 This applies to the following functions:
3801 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3802 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3804 BN_GF2m_mod_sqrt_arr
3805 BN_GF2m_mod_solve_quad_arr
3809 Conversion can be performed by the following functions:
3814 bntest.c has additional tests for binary polynomial arithmetic.
3816 Two implementations for BN_GF2m_mod_div() are available.
3817 The default algorithm simply uses BN_GF2m_mod_inv() and
3818 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3819 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3820 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3822 [Sheueling Chang Shantz and Douglas Stebila
3823 (Sun Microsystems Laboratories)]
3825 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3826 functionality is disabled at compile-time.
3827 [Douglas Stebila <douglas.stebila@sun.com>]
3829 *) Change default behaviour of 'openssl asn1parse' so that more
3830 information is visible when viewing, e.g., a certificate:
3832 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3833 mode the content of non-printable OCTET STRINGs is output in a
3834 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3835 avoid the appearance of a printable string.
3836 [Nils Larsch <nla@trustcenter.de>]
3838 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3840 EC_GROUP_set_asn1_flag()
3841 EC_GROUP_get_asn1_flag()
3842 EC_GROUP_set_point_conversion_form()
3843 EC_GROUP_get_point_conversion_form()
3844 These control ASN1 encoding details:
3845 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3846 has been set to OPENSSL_EC_NAMED_CURVE.
3847 - Points are encoded in uncompressed form by default; options for
3848 asn1_for are as for point2oct, namely
3849 POINT_CONVERSION_COMPRESSED
3850 POINT_CONVERSION_UNCOMPRESSED
3851 POINT_CONVERSION_HYBRID
3853 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3856 EC_GROUP_get0_seed()
3857 EC_GROUP_get_seed_len()
3858 This is used only for ASN1 purposes (so far).
3859 [Nils Larsch <nla@trustcenter.de>]
3861 *) Add 'field_type' member to EC_METHOD, which holds the NID
3862 of the appropriate field type OID. The new function
3863 EC_METHOD_get_field_type() returns this value.
3864 [Nils Larsch <nla@trustcenter.de>]
3869 EC_POINT_point2hex()
3870 EC_POINT_hex2point()
3871 providing useful interfaces to EC_POINT_point2oct() and
3872 EC_POINT_oct2point().
3873 [Nils Larsch <nla@trustcenter.de>]
3875 *) Change internals of the EC library so that the functions
3876 EC_GROUP_set_generator()
3877 EC_GROUP_get_generator()
3878 EC_GROUP_get_order()
3879 EC_GROUP_get_cofactor()
3880 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3881 to methods, which would lead to unnecessary code duplication when
3882 adding different types of curves.
3883 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3885 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3886 arithmetic, and such that modified wNAFs are generated
3887 (which avoid length expansion in many cases).
3890 *) Add a function EC_GROUP_check_discriminant() (defined via
3891 EC_METHOD) that verifies that the curve discriminant is non-zero.
3893 Add a function EC_GROUP_check() that makes some sanity tests
3894 on a EC_GROUP, its generator and order. This includes
3895 EC_GROUP_check_discriminant().
3896 [Nils Larsch <nla@trustcenter.de>]
3898 *) Add ECDSA in new directory crypto/ecdsa/.
3900 Add applications 'openssl ecparam' and 'openssl ecdsa'
3901 (these are based on 'openssl dsaparam' and 'openssl dsa').
3903 ECDSA support is also included in various other files across the
3904 library. Most notably,
3905 - 'openssl req' now has a '-newkey ecdsa:file' option;
3906 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3907 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3908 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3909 them suitable for ECDSA where domain parameters must be
3910 extracted before the specific public key;
3911 - ECDSA engine support has been added.
3912 [Nils Larsch <nla@trustcenter.de>]
3914 *) Include some named elliptic curves, and add OIDs from X9.62,
3915 SECG, and WAP/WTLS. Each curve can be obtained from the new
3917 EC_GROUP_new_by_curve_name(),
3918 and the list of available named curves can be obtained with
3919 EC_get_builtin_curves().
3920 Also add a 'curve_name' member to EC_GROUP objects, which can be
3922 EC_GROUP_set_curve_name()
3923 EC_GROUP_get_curve_name()
3924 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3926 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3927 was actually never needed) and in BN_mul(). The removal in BN_mul()
3928 required a small change in bn_mul_part_recursive() and the addition
3929 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3930 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3931 bn_sub_words() and bn_add_words() except they take arrays with
3935 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3937 *) Cleanse PEM buffers before freeing them since they may contain
3939 [Benjamin Bennett <ben@psc.edu>]
3941 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3942 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3943 authentication-only ciphersuites.
3946 *) Since AES128 and AES256 share a single mask bit in the logic of
3947 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3948 kludge to work properly if AES128 is available and AES256 isn't.
3951 *) Expand security boundary to match 1.1.1 module.
3954 *) Remove redundant features: hash file source, editing of test vectors
3955 modify fipsld to use external fips_premain.c signature.
3958 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3959 run algorithm test programs.
3962 *) Make algorithm test programs more tolerant of whitespace.
3965 *) Have SSL/TLS server implementation tolerate "mismatched" record
3966 protocol version while receiving ClientHello even if the
3967 ClientHello is fragmented. (The server can't insist on the
3968 particular protocol version it has chosen before the ServerHello
3969 message has informed the client about his choice.)
3972 *) Load error codes if they are not already present instead of using a
3973 static variable. This allows them to be cleanly unloaded and reloaded.
3976 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3978 *) Introduce limits to prevent malicious keys being able to
3979 cause a denial of service. (CVE-2006-2940)
3980 [Steve Henson, Bodo Moeller]
3982 *) Fix ASN.1 parsing of certain invalid structures that can result
3983 in a denial of service. (CVE-2006-2937) [Steve Henson]
3985 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3986 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3988 *) Fix SSL client code which could crash if connecting to a
3989 malicious SSLv2 server. (CVE-2006-4343)
3990 [Tavis Ormandy and Will Drewry, Google Security Team]
3992 *) Change ciphersuite string processing so that an explicit
3993 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3994 will no longer include "AES128-SHA"), and any other similar
3995 ciphersuite (same bitmap) from *other* protocol versions (so that
3996 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3997 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3998 changes from 0.9.8b and 0.9.8d.
4001 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4003 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4004 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4006 *) Change the Unix randomness entropy gathering to use poll() when
4007 possible instead of select(), since the latter has some
4008 undesirable limitations.
4009 [Darryl Miles via Richard Levitte and Bodo Moeller]
4011 *) Disable rogue ciphersuites:
4013 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4014 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4015 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4017 The latter two were purportedly from
4018 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4021 Also deactive the remaining ciphersuites from
4022 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4023 unofficial, and the ID has long expired.
4026 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4027 dual-core machines) and other potential thread-safety issues.
4030 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4032 *) Adapt fipsld and the build system to link against the validated FIPS
4033 module in FIPS mode.
4036 *) Fixes for VC++ 2005 build under Windows.
4039 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4040 from a Windows bash shell such as MSYS. It is autodetected from the
4041 "config" script when run from a VC++ environment. Modify standard VC++
4042 build to use fipscanister.o from the GNU make build.
4045 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4047 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4048 The value now differs depending on if you build for FIPS or not.
4049 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4050 safely run with a non-FIPSed libcrypto, as it may crash because of
4051 the difference induced by this change.
4054 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4056 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4057 (part of SSL_OP_ALL). This option used to disable the
4058 countermeasure against man-in-the-middle protocol-version
4059 rollback in the SSL 2.0 server implementation, which is a bad
4060 idea. (CVE-2005-2969)
4062 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4063 for Information Security, National Institute of Advanced Industrial
4064 Science and Technology [AIST], Japan)]
4066 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4067 mainly for FIPS compliance and not fully integrated at this stage.
4070 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4071 the exponentiation using a fixed-length exponent. (Otherwise,
4072 the information leaked through timing could expose the secret key
4073 after many signatures; cf. Bleichenbacher's attack on DSA with
4077 *) Make a new fixed-window mod_exp implementation the default for
4078 RSA, DSA, and DH private-key operations so that the sequence of
4079 squares and multiplies and the memory access pattern are
4080 independent of the particular secret key. This will mitigate
4081 cache-timing and potential related attacks.
4083 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4084 and this is automatically used by BN_mod_exp_mont() if the new flag
4085 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4086 will use this BN flag for private exponents unless the flag
4087 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4088 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4090 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4092 *) Change the client implementation for SSLv23_method() and
4093 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4094 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4095 (Previously, the SSL 2.0 backwards compatible Client Hello
4096 message format would be used even with SSL_OP_NO_SSLv2.)
4099 *) Add support for smime-type MIME parameter in S/MIME messages which some
4103 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4104 a threadsafe manner. Modify rsa code to use new function and add calls
4105 to dsa and dh code (which had race conditions before).
4108 *) Include the fixed error library code in the C error file definitions
4109 instead of fixing them up at runtime. This keeps the error code
4110 structures constant.
4113 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4115 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4118 *) Fixes for newer kerberos headers. NB: the casts are needed because
4119 the 'length' field is signed on one version and unsigned on another
4120 with no (?) obvious way to tell the difference, without these VC++
4121 complains. Also the "definition" of FAR (blank) is no longer included
4122 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4123 some needed definitions.
4126 *) Undo Cygwin change.
4129 *) Added support for proxy certificates according to RFC 3820.
4130 Because they may be a security thread to unaware applications,
4131 they must be explicitely allowed in run-time. See
4132 docs/HOWTO/proxy_certificates.txt for further information.
4135 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4137 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4138 server and client random values. Previously
4139 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4140 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4142 This change has negligible security impact because:
4144 1. Server and client random values still have 24 bytes of pseudo random
4147 2. Server and client random values are sent in the clear in the initial
4150 3. The master secret is derived using the premaster secret (48 bytes in
4151 size for static RSA ciphersuites) as well as client server and random
4154 The OpenSSL team would like to thank the UK NISCC for bringing this issue
4157 [Stephen Henson, reported by UK NISCC]
4159 *) Use Windows randomness collection on Cygwin.
4162 *) Fix hang in EGD/PRNGD query when communication socket is closed
4163 prematurely by EGD/PRNGD.
4164 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
4166 *) Prompt for pass phrases when appropriate for PKCS12 input format.
4169 *) Back-port of selected performance improvements from development
4170 branch, as well as improved support for PowerPC platforms.
4173 *) Add lots of checks for memory allocation failure, error codes to indicate
4174 failure and freeing up memory if a failure occurs.
4175 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
4177 *) Add new -passin argument to dgst.
4180 *) Perform some character comparisons of different types in X509_NAME_cmp:
4181 this is needed for some certificates that reencode DNs into UTF8Strings
4182 (in violation of RFC3280) and can't or wont issue name rollover
4186 *) Make an explicit check during certificate validation to see that
4187 the CA setting in each certificate on the chain is correct. As a
4188 side effect always do the following basic checks on extensions,
4189 not just when there's an associated purpose to the check:
4191 - if there is an unhandled critical extension (unless the user
4192 has chosen to ignore this fault)
4193 - if the path length has been exceeded (if one is set at all)
4194 - that certain extensions fit the associated purpose (if one has
4198 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
4200 *) Avoid a race condition when CRLs are checked in a multi threaded
4201 environment. This would happen due to the reordering of the revoked
4202 entries during signature checking and serial number lookup. Now the
4203 encoding is cached and the serial number sort performed under a lock.
4204 Add new STACK function sk_is_sorted().
4207 *) Add Delta CRL to the extension code.
4210 *) Various fixes to s3_pkt.c so alerts are sent properly.
4211 [David Holmes <d.holmes@f5.com>]
4213 *) Reduce the chances of duplicate issuer name and serial numbers (in
4214 violation of RFC3280) using the OpenSSL certificate creation utilities.
4215 This is done by creating a random 64 bit value for the initial serial
4216 number when a serial number file is created or when a self signed
4217 certificate is created using 'openssl req -x509'. The initial serial
4218 number file is created using 'openssl x509 -next_serial' in CA.pl
4219 rather than being initialized to 1.
4222 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
4224 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4225 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4226 [Joe Orton, Steve Henson]
4228 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
4230 [Joe Orton, Steve Henson]
4232 *) Make it possible to have multiple active certificates with the same
4233 subject in the CA index file. This is done only if the keyword
4234 'unique_subject' is set to 'no' in the main CA section (default
4235 if 'CA_default') of the configuration file. The value is saved
4236 with the database itself in a separate index attribute file,
4237 named like the index file with '.attr' appended to the name.
4240 *) X509 verify fixes. Disable broken certificate workarounds when
4241 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
4242 keyUsage extension present. Don't accept CRLs with unhandled critical
4243 extensions: since verify currently doesn't process CRL extensions this
4244 rejects a CRL with *any* critical extensions. Add new verify error codes
4248 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
4249 A clarification of RFC2560 will require the use of OCTET STRINGs and
4250 some implementations cannot handle the current raw format. Since OpenSSL
4251 copies and compares OCSP nonces as opaque blobs without any attempt at
4252 parsing them this should not create any compatibility issues.
4255 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
4256 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
4257 this HMAC (and other) operations are several times slower than OpenSSL
4261 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
4262 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4264 *) Use the correct content when signing type "other".
4267 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
4269 *) Fix various bugs revealed by running the NISCC test suite:
4271 Stop out of bounds reads in the ASN1 code when presented with
4272 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4274 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
4276 If verify callback ignores invalid public key errors don't try to check
4277 certificate signature with the NULL public key.
4281 *) New -ignore_err option in ocsp application to stop the server
4282 exiting on the first error in a request.
4285 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4286 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4290 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4291 extra data after the compression methods not only for TLS 1.0
4292 but also for SSL 3.0 (as required by the specification).
4293 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4295 *) Change X509_certificate_type() to mark the key as exported/exportable
4296 when it's 512 *bits* long, not 512 bytes.
4299 *) Change AES_cbc_encrypt() so it outputs exact multiple of
4300 blocks during encryption.
4303 *) Various fixes to base64 BIO and non blocking I/O. On write
4304 flushes were not handled properly if the BIO retried. On read
4305 data was not being buffered properly and had various logic bugs.
4306 This also affects blocking I/O when the data being decoded is a
4310 *) Various S/MIME bugfixes and compatibility changes:
4311 output correct application/pkcs7 MIME type if
4312 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
4313 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
4314 of files as .eml work). Correctly handle very long lines in MIME
4318 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
4320 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4321 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4322 a protocol version number mismatch like a decryption error
4323 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4326 *) Turn on RSA blinding by default in the default implementation
4327 to avoid a timing attack. Applications that don't want it can call
4328 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4329 They would be ill-advised to do so in most cases.
4330 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4332 *) Change RSA blinding code so that it works when the PRNG is not
4333 seeded (in this case, the secret RSA exponent is abused as
4334 an unpredictable seed -- if it is not unpredictable, there
4335 is no point in blinding anyway). Make RSA blinding thread-safe
4336 by remembering the creator's thread ID in rsa->blinding and
4337 having all other threads use local one-time blinding factors
4338 (this requires more computation than sharing rsa->blinding, but
4339 avoids excessive locking; and if an RSA object is not shared
4340 between threads, blinding will still be very fast).
4343 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
4344 ENGINE as defaults for all supported algorithms irrespective of
4345 the 'flags' parameter. 'flags' is now honoured, so applications
4346 should make sure they are passing it correctly.
4349 *) Target "mingw" now allows native Windows code to be generated in
4350 the Cygwin environment as well as with the MinGW compiler.
4353 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
4355 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4356 via timing by performing a MAC computation even if incorrrect
4357 block cipher padding has been found. This is a countermeasure
4358 against active attacks where the attacker has to distinguish
4359 between bad padding and a MAC verification error. (CVE-2003-0078)
4361 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4362 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4363 Martin Vuagnoux (EPFL, Ilion)]
4365 *) Make the no-err option work as intended. The intention with no-err
4366 is not to have the whole error stack handling routines removed from
4367 libcrypto, it's only intended to remove all the function name and
4368 reason texts, thereby removing some of the footprint that may not
4369 be interesting if those errors aren't displayed anyway.
4371 NOTE: it's still possible for any application or module to have it's
4372 own set of error texts inserted. The routines are there, just not
4373 used by default when no-err is given.
4376 *) Add support for FreeBSD on IA64.
4377 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
4379 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
4380 Kerberos function mit_des_cbc_cksum(). Before this change,
4381 the value returned by DES_cbc_cksum() was like the one from
4382 mit_des_cbc_cksum(), except the bytes were swapped.
4383 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
4385 *) Allow an application to disable the automatic SSL chain building.
4386 Before this a rather primitive chain build was always performed in
4387 ssl3_output_cert_chain(): an application had no way to send the
4388 correct chain if the automatic operation produced an incorrect result.
4390 Now the chain builder is disabled if either:
4392 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
4394 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
4396 The reasoning behind this is that an application would not want the
4397 auto chain building to take place if extra chain certificates are
4398 present and it might also want a means of sending no additional
4399 certificates (for example the chain has two certificates and the
4403 *) Add the possibility to build without the ENGINE framework.
4404 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4406 *) Under Win32 gmtime() can return NULL: check return value in
4407 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
4410 *) DSA routines: under certain error conditions uninitialized BN objects
4411 could be freed. Solution: make sure initialization is performed early
4412 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
4413 Nils Larsch <nla@trustcenter.de> via PR#459)
4416 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
4417 checked on reconnect on the client side, therefore session resumption
4418 could still fail with a "ssl session id is different" error. This
4419 behaviour is masked when SSL_OP_ALL is used due to
4420 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
4421 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4422 followup to PR #377.
4425 *) IA-32 assembler support enhancements: unified ELF targets, support
4426 for SCO/Caldera platforms, fix for Cygwin shared build.
4429 *) Add support for FreeBSD on sparc64. As a consequence, support for
4430 FreeBSD on non-x86 processors is separate from x86 processors on
4431 the config script, much like the NetBSD support.
4432 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
4434 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
4436 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
4439 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
4440 code (06) was taken as the first octet of the session ID and the last
4441 octet was ignored consequently. As a result SSLv2 client side session
4442 caching could not have worked due to the session ID mismatch between
4444 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4448 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
4449 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
4453 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
4454 seems that in spite of existing for more than a year, many application
4455 author have done nothing to provide the necessary callbacks, which
4456 means that this particular engine will not work properly anywhere.
4457 This is a very unfortunate situation which forces us, in the name
4458 of usability, to give the hw_ncipher.c a static lock, which is part
4460 NOTE: This is for the 0.9.7 series ONLY. This hack will never
4461 appear in 0.9.8 or later. We EXPECT application authors to have
4462 dealt properly with this when 0.9.8 is released (unless we actually
4463 make such changes in the libcrypto locking code that changes will
4464 have to be made anyway).
4467 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
4468 octets have been read, EOF or an error occurs. Without this change
4469 some truncated ASN1 structures will not produce an error.
4472 *) Disable Heimdal support, since it hasn't been fully implemented.
4473 Still give the possibility to force the use of Heimdal, but with
4474 warnings and a request that patches get sent to openssl-dev.
4477 *) Add the VC-CE target, introduce the WINCE sysname, and add
4478 INSTALL.WCE and appropriate conditionals to make it build.
4479 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4481 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4482 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4483 edit numbers of the version.
4484 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4486 *) Introduce safe string copy and catenation functions
4487 (BUF_strlcpy() and BUF_strlcat()).
4488 [Ben Laurie (CHATS) and Richard Levitte]
4490 *) Avoid using fixed-size buffers for one-line DNs.
4491 [Ben Laurie (CHATS)]
4493 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4494 resizing buffers containing secrets, and use where appropriate.
4495 [Ben Laurie (CHATS)]
4497 *) Avoid using fixed size buffers for configuration file location.
4498 [Ben Laurie (CHATS)]
4500 *) Avoid filename truncation for various CA files.
4501 [Ben Laurie (CHATS)]
4503 *) Use sizeof in preference to magic numbers.
4504 [Ben Laurie (CHATS)]
4506 *) Avoid filename truncation in cert requests.
4507 [Ben Laurie (CHATS)]
4509 *) Add assertions to check for (supposedly impossible) buffer
4511 [Ben Laurie (CHATS)]
4513 *) Don't cache truncated DNS entries in the local cache (this could
4514 potentially lead to a spoofing attack).
4515 [Ben Laurie (CHATS)]
4517 *) Fix various buffers to be large enough for hex/decimal
4518 representations in a platform independent manner.
4519 [Ben Laurie (CHATS)]
4521 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4522 resizing buffers containing secrets, and use where appropriate.
4523 [Ben Laurie (CHATS)]
4525 *) Add BIO_indent() to avoid much slightly worrying code to do
4527 [Ben Laurie (CHATS)]
4529 *) Convert sprintf()/BIO_puts() to BIO_printf().
4530 [Ben Laurie (CHATS)]
4532 *) buffer_gets() could terminate with the buffer only half
4534 [Ben Laurie (CHATS)]
4536 *) Add assertions to prevent user-supplied crypto functions from
4537 overflowing internal buffers by having large block sizes, etc.
4538 [Ben Laurie (CHATS)]
4540 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4542 [Ben Laurie (CHATS)]
4544 *) Eliminate unused copy of key in RC4.
4545 [Ben Laurie (CHATS)]
4547 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4548 [Ben Laurie (CHATS)]
4550 *) Fix off-by-one error in EGD path.
4551 [Ben Laurie (CHATS)]
4553 *) If RANDFILE path is too long, ignore instead of truncating.
4554 [Ben Laurie (CHATS)]
4556 *) Eliminate unused and incorrectly sized X.509 structure
4558 [Ben Laurie (CHATS)]
4560 *) Eliminate unused and dangerous function knumber().
4561 [Ben Laurie (CHATS)]
4563 *) Eliminate unused and dangerous structure, KSSL_ERR.
4564 [Ben Laurie (CHATS)]
4566 *) Protect against overlong session ID context length in an encoded
4567 session object. Since these are local, this does not appear to be
4569 [Ben Laurie (CHATS)]
4571 *) Change from security patch (see 0.9.6e below) that did not affect
4572 the 0.9.6 release series:
4574 Remote buffer overflow in SSL3 protocol - an attacker could
4575 supply an oversized master key in Kerberos-enabled versions.
4577 [Ben Laurie (CHATS)]
4579 *) Change the SSL kerb5 codes to match RFC 2712.
4582 *) Make -nameopt work fully for req and add -reqopt switch.
4583 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4585 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4586 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4588 *) Make sure tests can be performed even if the corresponding algorithms
4589 have been removed entirely. This was also the last step to make
4590 OpenSSL compilable with DJGPP under all reasonable conditions.
4591 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
4593 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
4594 to allow version independent disabling of normally unselected ciphers,
4595 which may be activated as a side-effect of selecting a single cipher.
4597 (E.g., cipher list string "RSA" enables ciphersuites that are left
4598 out of "ALL" because they do not provide symmetric encryption.
4599 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
4600 [Lutz Jaenicke, Bodo Moeller]
4602 *) Add appropriate support for separate platform-dependent build
4603 directories. The recommended way to make a platform-dependent
4604 build directory is the following (tested on Linux), maybe with
4607 # Place yourself outside of the OpenSSL source tree. In
4608 # this example, the environment variable OPENSSL_SOURCE
4609 # is assumed to contain the absolute OpenSSL source directory.
4610 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
4611 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
4612 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
4613 mkdir -p `dirname $F`
4614 ln -s $OPENSSL_SOURCE/$F $F
4617 To be absolutely sure not to disturb the source tree, a "make clean"
4618 is a good thing. If it isn't successfull, don't worry about it,
4619 it probably means the source directory is very clean.
4622 *) Make sure any ENGINE control commands make local copies of string
4623 pointers passed to them whenever necessary. Otherwise it is possible
4624 the caller may have overwritten (or deallocated) the original string
4625 data when a later ENGINE operation tries to use the stored values.
4626 [Götz Babin-Ebell <babinebell@trustcenter.de>]
4628 *) Improve diagnostics in file reading and command-line digests.
4629 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
4631 *) Add AES modes CFB and OFB to the object database. Correct an
4632 error in AES-CFB decryption.
4635 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
4636 allows existing EVP_CIPHER_CTX structures to be reused after
4637 calling EVP_*Final(). This behaviour is used by encryption
4638 BIOs and some applications. This has the side effect that
4639 applications must explicitly clean up cipher contexts with
4640 EVP_CIPHER_CTX_cleanup() or they will leak memory.
4643 *) Check the values of dna and dnb in bn_mul_recursive before calling
4644 bn_mul_comba (a non zero value means the a or b arrays do not contain
4645 n2 elements) and fallback to bn_mul_normal if either is not zero.
4648 *) Fix escaping of non-ASCII characters when using the -subj option
4649 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
4652 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
4653 form for "surname", serialNumber has no short form.
4654 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
4655 therefore remove "mail" short name for "internet 7".
4656 The OID for unique identifiers in X509 certificates is
4657 x500UniqueIdentifier, not uniqueIdentifier.
4658 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
4661 *) Add an "init" command to the ENGINE config module and auto initialize
4662 ENGINEs. Without any "init" command the ENGINE will be initialized
4663 after all ctrl commands have been executed on it. If init=1 the
4664 ENGINE is initailized at that point (ctrls before that point are run
4665 on the uninitialized ENGINE and after on the initialized one). If
4666 init=0 then the ENGINE will not be iniatialized at all.
4669 *) Fix the 'app_verify_callback' interface so that the user-defined
4670 argument is actually passed to the callback: In the
4671 SSL_CTX_set_cert_verify_callback() prototype, the callback
4672 declaration has been changed from
4675 int (*cb)(X509_STORE_CTX *,void *);
4676 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
4677 i=s->ctx->app_verify_callback(&ctx)
4678 has been changed into
4679 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
4681 To update applications using SSL_CTX_set_cert_verify_callback(),
4682 a dummy argument can be added to their callback functions.
4683 [D. K. Smetters <smetters@parc.xerox.com>]
4685 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
4686 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
4688 *) Add and OPENSSL_LOAD_CONF define which will cause
4689 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
4690 This allows older applications to transparently support certain
4691 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
4692 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
4693 load the config file and OPENSSL_add_all_algorithms_conf() which will
4694 always load it have also been added.
4697 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
4698 Adjust NIDs and EVP layer.
4699 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4701 *) Config modules support in openssl utility.
4703 Most commands now load modules from the config file,
4704 though in a few (such as version) this isn't done
4705 because it couldn't be used for anything.
4707 In the case of ca and req the config file used is
4708 the same as the utility itself: that is the -config
4709 command line option can be used to specify an
4713 *) Move default behaviour from OPENSSL_config(). If appname is NULL
4714 use "openssl_conf" if filename is NULL use default openssl config file.
4717 *) Add an argument to OPENSSL_config() to allow the use of an alternative
4718 config section name. Add a new flag to tolerate a missing config file
4719 and move code to CONF_modules_load_file().
4722 *) Support for crypto accelerator cards from Accelerated Encryption
4723 Processing, www.aep.ie. (Use engine 'aep')
4724 The support was copied from 0.9.6c [engine] and adapted/corrected
4725 to work with the new engine framework.
4726 [AEP Inc. and Richard Levitte]
4728 *) Support for SureWare crypto accelerator cards from Baltimore
4729 Technologies. (Use engine 'sureware')
4730 The support was copied from 0.9.6c [engine] and adapted
4731 to work with the new engine framework.
4734 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
4735 make the newer ENGINE framework commands for the CHIL engine work.
4736 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
4738 *) Make it possible to produce shared libraries on ReliantUNIX.
4739 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
4741 *) Add the configuration target debug-linux-ppro.
4742 Make 'openssl rsa' use the general key loading routines
4743 implemented in apps.c, and make those routines able to
4744 handle the key format FORMAT_NETSCAPE and the variant
4746 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4748 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4749 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4751 *) Add -keyform to rsautl, and document -engine.
4752 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
4754 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
4755 BIO_R_NO_SUCH_FILE error code rather than the generic
4756 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
4759 *) Add new functions
4761 ERR_peek_last_error_line
4762 ERR_peek_last_error_line_data.
4763 These are similar to
4766 ERR_peek_error_line_data,
4767 but report on the latest error recorded rather than the first one
4768 still in the error queue.
4769 [Ben Laurie, Bodo Moeller]
4771 *) default_algorithms option in ENGINE config module. This allows things
4773 default_algorithms = ALL
4774 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4777 *) Prelminary ENGINE config module.
4780 *) New experimental application configuration code.
4783 *) Change the AES code to follow the same name structure as all other
4784 symmetric ciphers, and behave the same way. Move everything to
4785 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4786 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4788 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4789 [Ben Laurie and Theo de Raadt]
4791 *) Add option to output public keys in req command.
4792 [Massimiliano Pala madwolf@openca.org]
4794 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4795 (up to about 10% better than before for P-192 and P-224).
4798 *) New functions/macros
4800 SSL_CTX_set_msg_callback(ctx, cb)
4801 SSL_CTX_set_msg_callback_arg(ctx, arg)
4802 SSL_set_msg_callback(ssl, cb)
4803 SSL_set_msg_callback_arg(ssl, arg)
4805 to request calling a callback function
4807 void cb(int write_p, int version, int content_type,
4808 const void *buf, size_t len, SSL *ssl, void *arg)
4810 whenever a protocol message has been completely received
4811 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4812 protocol version according to which the SSL library interprets
4813 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4814 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4815 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4816 specification (change_cipher_spec(20), alert(21), handshake(22)).
4817 'buf' and 'len' point to the actual message, 'ssl' to the
4818 SSL object, and 'arg' is the application-defined value set by
4819 SSL[_CTX]_set_msg_callback_arg().
4821 'openssl s_client' and 'openssl s_server' have new '-msg' options
4822 to enable a callback that displays all protocol messages.
4825 *) Change the shared library support so shared libraries are built as
4826 soon as the corresponding static library is finished, and thereby get
4827 openssl and the test programs linked against the shared library.
4828 This still only happens when the keyword "shard" has been given to
4829 the configuration scripts.
4831 NOTE: shared library support is still an experimental thing, and
4832 backward binary compatibility is still not guaranteed.
4833 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4835 *) Add support for Subject Information Access extension.
4836 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4838 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4839 additional bytes when new memory had to be allocated, not just
4840 when reusing an existing buffer.
4843 *) New command line and configuration option 'utf8' for the req command.
4844 This allows field values to be specified as UTF8 strings.
4847 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4848 runs for the former and machine-readable output for the latter.
4851 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4852 of the e-mail address in the DN (i.e., it will go into a certificate
4853 extension only). The new configuration file option 'email_in_dn = no'
4854 has the same effect.
4855 [Massimiliano Pala madwolf@openca.org]
4857 *) Change all functions with names starting with des_ to be starting
4858 with DES_ instead. Add wrappers that are compatible with libdes,
4859 but are named _ossl_old_des_*. Finally, add macros that map the
4860 des_* symbols to the corresponding _ossl_old_des_* if libdes
4861 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4862 desired, the des_* symbols will be mapped to DES_*, with one
4865 Since we provide two compatibility mappings, the user needs to
4866 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4867 compatibility is desired. The default (i.e., when that macro
4868 isn't defined) is OpenSSL 0.9.6c compatibility.
4870 There are also macros that enable and disable the support of old
4871 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4872 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4873 are defined, the default will apply: to support the old des routines.
4875 In either case, one must include openssl/des.h to get the correct
4876 definitions. Do not try to just include openssl/des_old.h, that
4879 NOTE: This is a major break of an old API into a new one. Software
4880 authors are encouraged to switch to the DES_ style functions. Some
4881 time in the future, des_old.h and the libdes compatibility functions
4882 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4883 default), and then completely removed.
4886 *) Test for certificates which contain unsupported critical extensions.
4887 If such a certificate is found during a verify operation it is
4888 rejected by default: this behaviour can be overridden by either
4889 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4890 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4891 X509_supported_extension() has also been added which returns 1 if a
4892 particular extension is supported.
4895 *) Modify the behaviour of EVP cipher functions in similar way to digests
4896 to retain compatibility with existing code.
4899 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4900 compatibility with existing code. In particular the 'ctx' parameter does
4901 not have to be to be initialized before the call to EVP_DigestInit() and
4902 it is tidied up after a call to EVP_DigestFinal(). New function
4903 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4904 EVP_MD_CTX_copy() changed to not require the destination to be
4905 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4906 requires the destination to be valid.
4908 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4909 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4912 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4913 so that complete 'Handshake' protocol structures are kept in memory
4914 instead of overwriting 'msg_type' and 'length' with 'body' data.
4917 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4918 [Massimo Santin via Richard Levitte]
4920 *) Major restructuring to the underlying ENGINE code. This includes
4921 reduction of linker bloat, separation of pure "ENGINE" manipulation
4922 (initialisation, etc) from functionality dealing with implementations
4923 of specific crypto iterfaces. This change also introduces integrated
4924 support for symmetric ciphers and digest implementations - so ENGINEs
4925 can now accelerate these by providing EVP_CIPHER and EVP_MD
4926 implementations of their own. This is detailed in crypto/engine/README
4927 as it couldn't be adequately described here. However, there are a few
4928 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4929 were changed in the original introduction of ENGINE code have now
4930 reverted back - the hooking from this code to ENGINE is now a good
4931 deal more passive and at run-time, operations deal directly with
4932 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4933 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4934 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4935 they were not being used by the framework as there is no concept of a
4936 BIGNUM_METHOD and they could not be generalised to the new
4937 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4938 ENGINE_cpy() has been removed as it cannot be consistently defined in
4942 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4945 *) Change mkdef.pl to sort symbols that get the same entry number,
4946 and make sure the automatically generated functions ERR_load_*
4947 become part of libeay.num as well.
4950 *) New function SSL_renegotiate_pending(). This returns true once
4951 renegotiation has been requested (either SSL_renegotiate() call
4952 or HelloRequest/ClientHello receveived from the peer) and becomes
4953 false once a handshake has been completed.
4954 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4955 sends a HelloRequest, but does not ensure that a handshake takes
4956 place. SSL_renegotiate_pending() is useful for checking if the
4957 client has followed the request.)
4960 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4961 By default, clients may request session resumption even during
4962 renegotiation (if session ID contexts permit); with this option,
4963 session resumption is possible only in the first handshake.
4965 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4966 more bits available for options that should not be part of
4967 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4970 *) Add some demos for certificate and certificate request creation.
4973 *) Make maximum certificate chain size accepted from the peer application
4974 settable (SSL*_get/set_max_cert_list()), as proposed by
4975 "Douglas E. Engert" <deengert@anl.gov>.
4978 *) Add support for shared libraries for Unixware-7
4979 (Boyd Lynn Gerber <gerberb@zenez.com>).
4982 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4983 be done prior to destruction. Use this to unload error strings from
4984 ENGINEs that load their own error strings. NB: This adds two new API
4985 functions to "get" and "set" this destroy handler in an ENGINE.
4988 *) Alter all existing ENGINE implementations (except "openssl" and
4989 "openbsd") to dynamically instantiate their own error strings. This
4990 makes them more flexible to be built both as statically-linked ENGINEs
4991 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4992 Also, add stub code to each that makes building them as self-contained
4993 shared-libraries easier (see README.ENGINE).
4996 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4997 implementations into applications that are completely implemented in
4998 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4999 commands that can be used to configure what shared-library to load and
5000 to control aspects of the way it is handled. Also, made an update to
5001 the README.ENGINE file that brings its information up-to-date and
5002 provides some information and instructions on the "dynamic" ENGINE
5003 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5006 *) Make it possible to unload ranges of ERR strings with a new
5007 "ERR_unload_strings" function.
5010 *) Add a copy() function to EVP_MD.
5013 *) Make EVP_MD routines take a context pointer instead of just the
5014 md_data void pointer.
5017 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5018 that the digest can only process a single chunk of data
5019 (typically because it is provided by a piece of
5020 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5021 is only going to provide a single chunk of data, and hence the
5022 framework needn't accumulate the data for oneshot drivers.
5025 *) As with "ERR", make it possible to replace the underlying "ex_data"
5026 functions. This change also alters the storage and management of global
5027 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5028 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5029 index counters. The API functions that use this state have been changed
5030 to take a "class_index" rather than pointers to the class's local STACK
5031 and counter, and there is now an API function to dynamically create new
5032 classes. This centralisation allows us to (a) plug a lot of the
5033 thread-safety problems that existed, and (b) makes it possible to clean
5034 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5035 such data would previously have always leaked in application code and
5036 workarounds were in place to make the memory debugging turn a blind eye
5037 to it. Application code that doesn't use this new function will still
5038 leak as before, but their memory debugging output will announce it now
5039 rather than letting it slide.
5041 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5042 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5043 has a return value to indicate success or failure.
5046 *) Make it possible to replace the underlying "ERR" functions such that the
5047 global state (2 LHASH tables and 2 locks) is only used by the "default"
5048 implementation. This change also adds two functions to "get" and "set"
5049 the implementation prior to it being automatically set the first time
5050 any other ERR function takes place. Ie. an application can call "get",
5051 pass the return value to a module it has just loaded, and that module
5052 can call its own "set" function using that value. This means the
5053 module's "ERR" operations will use (and modify) the error state in the
5054 application and not in its own statically linked copy of OpenSSL code.
5057 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5058 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5059 the operation, and provides a more encapsulated way for external code
5060 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5061 to use these functions rather than manually incrementing the counts.
5063 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5066 *) Add EVP test program.
5069 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5072 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5073 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5074 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5075 These allow a CRL to be built without having to access X509_CRL fields
5076 directly. Modify 'ca' application to use new functions.
5079 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5080 bug workarounds. Rollback attack detection is a security feature.
5081 The problem will only arise on OpenSSL servers when TLSv1 is not
5082 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5083 Software authors not wanting to support TLSv1 will have special reasons
5084 for their choice and can explicitly enable this option.
5085 [Bodo Moeller, Lutz Jaenicke]
5087 *) Rationalise EVP so it can be extended: don't include a union of
5088 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5089 (similar to those existing for EVP_CIPHER_CTX).
5094 EVP_MD_CTX_init(&md); /* new function call */
5095 EVP_DigestInit(&md, EVP_sha1());
5096 EVP_DigestUpdate(&md, in, len);
5097 EVP_DigestFinal(&md, out, NULL);
5098 EVP_MD_CTX_cleanup(&md); /* new function call */
5102 *) Make DES key schedule conform to the usual scheme, as well as
5103 correcting its structure. This means that calls to DES functions
5104 now have to pass a pointer to a des_key_schedule instead of a
5105 plain des_key_schedule (which was actually always a pointer
5108 des_key_schedule ks;
5110 des_set_key_checked(..., &ks);
5111 des_ncbc_encrypt(..., &ks, ...);
5113 (Note that a later change renames 'des_...' into 'DES_...'.)
5116 *) Initial reduction of linker bloat: the use of some functions, such as
5117 PEM causes large amounts of unused functions to be linked in due to
5118 poor organisation. For example pem_all.c contains every PEM function
5119 which has a knock on effect of linking in large amounts of (unused)
5120 ASN1 code. Grouping together similar functions and splitting unrelated
5121 functions prevents this.
5124 *) Cleanup of EVP macros.
5127 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5128 correct _ecb suffix.
5131 *) Add initial OCSP responder support to ocsp application. The
5132 revocation information is handled using the text based index
5133 use by the ca application. The responder can either handle
5134 requests generated internally, supplied in files (for example
5135 via a CGI script) or using an internal minimal server.
5138 *) Add configuration choices to get zlib compression for TLS.
5141 *) Changes to Kerberos SSL for RFC 2712 compliance:
5142 1. Implemented real KerberosWrapper, instead of just using
5143 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5144 2. Implemented optional authenticator field of KerberosWrapper.
5146 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5147 and authenticator structs; see crypto/krb5/.
5149 Generalized Kerberos calls to support multiple Kerberos libraries.
5150 [Vern Staats <staatsvr@asc.hpc.mil>,
5151 Jeffrey Altman <jaltman@columbia.edu>
5152 via Richard Levitte]
5154 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
5155 already does with RSA. testdsa.h now has 'priv_key/pub_key'
5156 values for each of the key sizes rather than having just
5157 parameters (and 'speed' generating keys each time).
5160 *) Speed up EVP routines.
5163 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
5164 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
5165 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
5166 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
5168 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
5169 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
5170 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
5173 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
5175 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
5178 *) Added the OS2-EMX target.
5179 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
5181 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
5182 to support NCONF routines in extension code. New function CONF_set_nconf()
5183 to allow functions which take an NCONF to also handle the old LHASH
5184 structure: this means that the old CONF compatible routines can be
5185 retained (in particular wrt extensions) without having to duplicate the
5186 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
5189 *) Enhance the general user interface with mechanisms for inner control
5190 and with possibilities to have yes/no kind of prompts.
5193 *) Change all calls to low level digest routines in the library and
5194 applications to use EVP. Add missing calls to HMAC_cleanup() and
5195 don't assume HMAC_CTX can be copied using memcpy().
5196 [Verdon Walker <VWalker@novell.com>, Steve Henson]
5198 *) Add the possibility to control engines through control names but with
5199 arbitrary arguments instead of just a string.
5200 Change the key loaders to take a UI_METHOD instead of a callback
5201 function pointer. NOTE: this breaks binary compatibility with earlier
5202 versions of OpenSSL [engine].
5203 Adapt the nCipher code for these new conditions and add a card insertion
5207 *) Enhance the general user interface with mechanisms to better support
5208 dialog box interfaces, application-defined prompts, the possibility
5209 to use defaults (for example default passwords from somewhere else)
5210 and interrupts/cancellations.
5213 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
5214 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
5217 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
5218 tidy up some unnecessarily weird code in 'sk_new()').
5219 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
5221 *) Change the key loading routines for ENGINEs to use the same kind
5222 callback (pem_password_cb) as all other routines that need this
5226 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
5227 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
5228 than this minimum value is recommended.
5231 *) New random seeder for OpenVMS, using the system process statistics
5232 that are easily reachable.
5235 *) Windows apparently can't transparently handle global
5236 variables defined in DLLs. Initialisations such as:
5238 const ASN1_ITEM *it = &ASN1_INTEGER_it;
5240 wont compile. This is used by the any applications that need to
5241 declare their own ASN1 modules. This was fixed by adding the option
5242 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
5243 needed for static libraries under Win32.
5246 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
5247 setting of purpose and trust fields. New X509_STORE trust and
5248 purpose functions and tidy up setting in other SSL functions.
5251 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
5252 structure. These are inherited by X509_STORE_CTX when it is
5253 initialised. This allows various defaults to be set in the
5254 X509_STORE structure (such as flags for CRL checking and custom
5255 purpose or trust settings) for functions which only use X509_STORE_CTX
5256 internally such as S/MIME.
5258 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
5259 trust settings if they are not set in X509_STORE. This allows X509_STORE
5260 purposes and trust (in S/MIME for example) to override any set by default.
5262 Add command line options for CRL checking to smime, s_client and s_server
5266 *) Initial CRL based revocation checking. If the CRL checking flag(s)
5267 are set then the CRL is looked up in the X509_STORE structure and
5268 its validity and signature checked, then if the certificate is found
5269 in the CRL the verify fails with a revoked error.
5271 Various new CRL related callbacks added to X509_STORE_CTX structure.
5273 Command line options added to 'verify' application to support this.
5275 This needs some additional work, such as being able to handle multiple
5276 CRLs with different times, extension based lookup (rather than just
5277 by subject name) and ultimately more complete V2 CRL extension
5281 *) Add a general user interface API (crypto/ui/). This is designed
5282 to replace things like des_read_password and friends (backward
5283 compatibility functions using this new API are provided).
5284 The purpose is to remove prompting functions from the DES code
5285 section as well as provide for prompting through dialog boxes in
5286 a window system and the like.
5289 *) Add "ex_data" support to ENGINE so implementations can add state at a
5290 per-structure level rather than having to store it globally.
5293 *) Make it possible for ENGINE structures to be copied when retrieved by
5294 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
5295 This causes the "original" ENGINE structure to act like a template,
5296 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
5297 operational state can be localised to each ENGINE structure, despite the
5298 fact they all share the same "methods". New ENGINE structures returned in
5299 this case have no functional references and the return value is the single
5300 structural reference. This matches the single structural reference returned
5301 by ENGINE_by_id() normally, when it is incremented on the pre-existing
5305 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
5306 needs to match any other type at all we need to manually clear the
5310 *) Changes to the "openssl engine" utility to include;
5311 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
5312 about an ENGINE's available control commands.
5313 - executing control commands from command line arguments using the
5314 '-pre' and '-post' switches. '-post' is only used if '-t' is
5315 specified and the ENGINE is successfully initialised. The syntax for
5316 the individual commands are colon-separated, for example;
5317 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
5320 *) New dynamic control command support for ENGINEs. ENGINEs can now
5321 declare their own commands (numbers), names (strings), descriptions,
5322 and input types for run-time discovery by calling applications. A
5323 subset of these commands are implicitly classed as "executable"
5324 depending on their input type, and only these can be invoked through
5325 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
5326 can be based on user input, config files, etc). The distinction is
5327 that "executable" commands cannot return anything other than a boolean
5328 result and can only support numeric or string input, whereas some
5329 discoverable commands may only be for direct use through
5330 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
5331 pointers, or other custom uses. The "executable" commands are to
5332 support parameterisations of ENGINE behaviour that can be
5333 unambiguously defined by ENGINEs and used consistently across any
5334 OpenSSL-based application. Commands have been added to all the
5335 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
5336 control over shared-library paths without source code alterations.
5339 *) Changed all ENGINE implementations to dynamically allocate their
5340 ENGINEs rather than declaring them statically. Apart from this being
5341 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
5342 this also allows the implementations to compile without using the
5343 internal engine_int.h header.
5346 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
5347 'const' value. Any code that should be able to modify a RAND_METHOD
5348 should already have non-const pointers to it (ie. they should only
5349 modify their own ones).
5352 *) Made a variety of little tweaks to the ENGINE code.
5353 - "atalla" and "ubsec" string definitions were moved from header files
5354 to C code. "nuron" string definitions were placed in variables
5355 rather than hard-coded - allowing parameterisation of these values
5356 later on via ctrl() commands.
5357 - Removed unused "#if 0"'d code.
5358 - Fixed engine list iteration code so it uses ENGINE_free() to release
5359 structural references.
5360 - Constified the RAND_METHOD element of ENGINE structures.
5361 - Constified various get/set functions as appropriate and added
5362 missing functions (including a catch-all ENGINE_cpy that duplicates
5363 all ENGINE values onto a new ENGINE except reference counts/state).
5364 - Removed NULL parameter checks in get/set functions. Setting a method
5365 or function to NULL is a way of cancelling out a previously set
5366 value. Passing a NULL ENGINE parameter is just plain stupid anyway
5367 and doesn't justify the extra error symbols and code.
5368 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
5369 flags from engine_int.h to engine.h.
5370 - Changed prototypes for ENGINE handler functions (init(), finish(),
5371 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
5374 *) Implement binary inversion algorithm for BN_mod_inverse in addition
5375 to the algorithm using long division. The binary algorithm can be
5376 used only if the modulus is odd. On 32-bit systems, it is faster
5377 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
5378 roughly 5-15% for 256-bit moduli), so we use it only for moduli
5379 up to 450 bits. In 64-bit environments, the binary algorithm
5380 appears to be advantageous for much longer moduli; here we use it
5381 for moduli up to 2048 bits.
5384 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
5385 could not support the combine flag in choice fields.
5388 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
5389 extensions from a certificate request to the certificate.
5392 *) Allow multiple 'certopt' and 'nameopt' options to be separated
5393 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
5394 file: this allows the display of the certificate about to be
5395 signed to be customised, to allow certain fields to be included
5396 or excluded and extension details. The old system didn't display
5397 multicharacter strings properly, omitted fields not in the policy
5398 and couldn't display additional details such as extensions.
5401 *) Function EC_POINTs_mul for multiple scalar multiplication
5402 of an arbitrary number of elliptic curve points
5403 \sum scalars[i]*points[i],
5404 optionally including the generator defined for the EC_GROUP:
5405 scalar*generator + \sum scalars[i]*points[i].
5407 EC_POINT_mul is a simple wrapper function for the typical case
5408 that the point list has just one item (besides the optional
5412 *) First EC_METHODs for curves over GF(p):
5414 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
5415 operations and provides various method functions that can also
5416 operate with faster implementations of modular arithmetic.
5418 EC_GFp_mont_method() reuses most functions that are part of
5419 EC_GFp_simple_method, but uses Montgomery arithmetic.
5421 [Bodo Moeller; point addition and point doubling
5422 implementation directly derived from source code provided by
5423 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
5425 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
5426 crypto/ec/ec_lib.c):
5428 Curves are EC_GROUP objects (with an optional group generator)
5429 based on EC_METHODs that are built into the library.
5431 Points are EC_POINT objects based on EC_GROUP objects.
5433 Most of the framework would be able to handle curves over arbitrary
5434 finite fields, but as there are no obvious types for fields other
5435 than GF(p), some functions are limited to that for now.
5438 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
5439 that the file contains a complete HTTP response.
5442 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
5443 change the def and num file printf format specifier from "%-40sXXX"
5444 to "%-39s XXX". The latter will always guarantee a space after the
5445 field while the former will cause them to run together if the field
5446 is 40 of more characters long.
5449 *) Constify the cipher and digest 'method' functions and structures
5450 and modify related functions to take constant EVP_MD and EVP_CIPHER
5454 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
5455 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
5458 *) Modify EVP_Digest*() routines so they now return values. Although the
5459 internal software routines can never fail additional hardware versions
5463 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
5465 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
5466 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
5469 ERR_R_NESTED_ASN1_ERROR
5471 ERR_R_MISSING_ASN1_EOS
5472 were 4 .. 9, conflicting with
5473 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5475 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5476 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5478 Add new error code 'ERR_R_INTERNAL_ERROR'.
5481 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5485 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5486 sets the subject name for a new request or supersedes the
5487 subject name in a given request. Formats that can be parsed are
5488 'CN=Some Name, OU=myOU, C=IT'
5490 'CN=Some Name/OU=myOU/C=IT'.
5492 Add options '-batch' and '-verbose' to 'openssl req'.
5493 [Massimiliano Pala <madwolf@hackmasters.net>]
5495 *) Introduce the possibility to access global variables through
5496 functions on platform were that's the best way to handle exporting
5497 global variables in shared libraries. To enable this functionality,
5498 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5499 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5500 is normally done by Configure or something similar).
5502 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5503 in the source file (foo.c) like this:
5505 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5506 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5508 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5509 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5511 OPENSSL_DECLARE_GLOBAL(int,foo);
5512 #define foo OPENSSL_GLOBAL_REF(foo)
5513 OPENSSL_DECLARE_GLOBAL(double,bar);
5514 #define bar OPENSSL_GLOBAL_REF(bar)
5516 The #defines are very important, and therefore so is including the
5517 header file everywhere where the defined globals are used.
5519 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5520 of ASN.1 items, but that structure is a bit different.
5522 The largest change is in util/mkdef.pl which has been enhanced with
5523 better and easier to understand logic to choose which symbols should
5524 go into the Windows .def files as well as a number of fixes and code
5525 cleanup (among others, algorithm keywords are now sorted
5526 lexicographically to avoid constant rewrites).
5529 *) In BN_div() keep a copy of the sign of 'num' before writing the
5530 result to 'rm' because if rm==num the value will be overwritten
5531 and produce the wrong result if 'num' is negative: this caused
5532 problems with BN_mod() and BN_nnmod().
5535 *) Function OCSP_request_verify(). This checks the signature on an
5536 OCSP request and verifies the signer certificate. The signer
5537 certificate is just checked for a generic purpose and OCSP request
5541 *) Add OCSP_check_validity() function to check the validity of OCSP
5542 responses. OCSP responses are prepared in real time and may only
5543 be a few seconds old. Simply checking that the current time lies
5544 between thisUpdate and nextUpdate max reject otherwise valid responses
5545 caused by either OCSP responder or client clock inaccuracy. Instead
5546 we allow thisUpdate and nextUpdate to fall within a certain period of
5547 the current time. The age of the response can also optionally be
5548 checked. Two new options -validity_period and -status_age added to
5552 *) If signature or public key algorithm is unrecognized print out its
5553 OID rather that just UNKNOWN.
5556 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5557 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5558 ID to be generated from the issuer certificate alone which can then be
5559 passed to OCSP_id_issuer_cmp().
5562 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5563 ASN1 modules to export functions returning ASN1_ITEM pointers
5564 instead of the ASN1_ITEM structures themselves. This adds several
5565 new macros which allow the underlying ASN1 function/structure to
5566 be accessed transparently. As a result code should not use ASN1_ITEM
5567 references directly (such as &X509_it) but instead use the relevant
5568 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5569 use of the new ASN1 code on platforms where exporting structures
5570 is problematical (for example in shared libraries) but exporting
5571 functions returning pointers to structures is not.
5574 *) Add support for overriding the generation of SSL/TLS session IDs.
5575 These callbacks can be registered either in an SSL_CTX or per SSL.
5576 The purpose of this is to allow applications to control, if they wish,
5577 the arbitrary values chosen for use as session IDs, particularly as it
5578 can be useful for session caching in multiple-server environments. A
5579 command-line switch for testing this (and any client code that wishes
5580 to use such a feature) has been added to "s_server".
5581 [Geoff Thorpe, Lutz Jaenicke]
5583 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5584 of the form '#if defined(...) || defined(...) || ...' and
5585 '#if !defined(...) && !defined(...) && ...'. This also avoids
5586 the growing number of special cases it was previously handling.
5589 *) Make all configuration macros available for application by making
5590 sure they are available in opensslconf.h, by giving them names starting
5591 with "OPENSSL_" to avoid conflicts with other packages and by making
5592 sure e_os2.h will cover all platform-specific cases together with
5594 Additionally, it is now possible to define configuration/platform-
5595 specific names (called "system identities"). In the C code, these
5596 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
5597 macro with the name beginning with "OPENSSL_SYS_", which is determined
5598 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
5602 *) New option -set_serial to 'req' and 'x509' this allows the serial
5603 number to use to be specified on the command line. Previously self
5604 signed certificates were hard coded with serial number 0 and the
5605 CA options of 'x509' had to use a serial number in a file which was
5609 *) New options to 'ca' utility to support V2 CRL entry extensions.
5610 Currently CRL reason, invalidity date and hold instruction are
5611 supported. Add new CRL extensions to V3 code and some new objects.
5614 *) New function EVP_CIPHER_CTX_set_padding() this is used to
5615 disable standard block padding (aka PKCS#5 padding) in the EVP
5616 API, which was previously mandatory. This means that the data is
5617 not padded in any way and so the total length much be a multiple
5618 of the block size, otherwise an error occurs.
5621 *) Initial (incomplete) OCSP SSL support.
5624 *) New function OCSP_parse_url(). This splits up a URL into its host,
5625 port and path components: primarily to parse OCSP URLs. New -url
5626 option to ocsp utility.
5629 *) New nonce behavior. The return value of OCSP_check_nonce() now
5630 reflects the various checks performed. Applications can decide
5631 whether to tolerate certain situations such as an absent nonce
5632 in a response when one was present in a request: the ocsp application
5633 just prints out a warning. New function OCSP_add1_basic_nonce()
5634 this is to allow responders to include a nonce in a response even if
5635 the request is nonce-less.
5638 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
5639 skipped when using openssl x509 multiple times on a single input file,
5640 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
5643 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
5644 set string type: to handle setting ASN1_TIME structures. Fix ca
5645 utility to correctly initialize revocation date of CRLs.
5648 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
5649 the clients preferred ciphersuites and rather use its own preferences.
5650 Should help to work around M$ SGC (Server Gated Cryptography) bug in
5651 Internet Explorer by ensuring unchanged hash method during stepup.
5652 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
5655 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
5656 to aes and add a new 'exist' option to print out symbols that don't
5660 *) Additional options to ocsp utility to allow flags to be set and
5661 additional certificates supplied.
5664 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
5665 OCSP client a number of certificate to only verify the response
5669 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
5670 handle the new API. Currently only ECB, CBC modes supported. Add new
5673 Add TLS AES ciphersuites as described in RFC3268, "Advanced
5674 Encryption Standard (AES) Ciphersuites for Transport Layer
5675 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
5676 not enabled by default and were not part of the "ALL" ciphersuite
5677 alias because they were not yet official; they could be
5678 explicitly requested by specifying the "AESdraft" ciphersuite
5679 group alias. In the final release of OpenSSL 0.9.7, the group
5680 alias is called "AES" and is part of "ALL".)
5681 [Ben Laurie, Steve Henson, Bodo Moeller]
5683 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
5684 request to response.
5687 *) Functions for OCSP responders. OCSP_request_onereq_count(),
5688 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
5689 extract information from a certificate request. OCSP_response_create()
5690 creates a response and optionally adds a basic response structure.
5691 OCSP_basic_add1_status() adds a complete single response to a basic
5692 response and returns the OCSP_SINGLERESP structure just added (to allow
5693 extensions to be included for example). OCSP_basic_add1_cert() adds a
5694 certificate to a basic response and OCSP_basic_sign() signs a basic
5695 response with various flags. New helper functions ASN1_TIME_check()
5696 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
5697 (converts ASN1_TIME to GeneralizedTime).
5700 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
5701 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
5702 structure from a certificate. X509_pubkey_digest() digests the public_key
5703 contents: this is used in various key identifiers.
5706 *) Make sk_sort() tolerate a NULL argument.
5707 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
5709 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
5710 passed by the function are trusted implicitly. If any of them signed the
5711 response then it is assumed to be valid and is not verified.
5714 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
5715 to data. This was previously part of the PKCS7 ASN1 code. This
5716 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
5717 [Steve Henson, reported by Kenneth R. Robinette
5718 <support@securenetterm.com>]
5720 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
5721 routines: without these tracing memory leaks is very painful.
5722 Fix leaks in PKCS12 and PKCS7 routines.
5725 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
5726 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
5727 effectively meant GeneralizedTime would never be used. Now it
5728 is initialised to -1 but X509_time_adj() now has to check the value
5729 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
5730 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
5731 [Steve Henson, reported by Kenneth R. Robinette
5732 <support@securenetterm.com>]
5734 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
5735 result in a zero length in the ASN1_INTEGER structure which was
5736 not consistent with the structure when d2i_ASN1_INTEGER() was used
5737 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
5738 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
5739 where it did not print out a minus for negative ASN1_INTEGER.
5742 *) Add summary printout to ocsp utility. The various functions which
5743 convert status values to strings have been renamed to:
5744 OCSP_response_status_str(), OCSP_cert_status_str() and
5745 OCSP_crl_reason_str() and are no longer static. New options
5746 to verify nonce values and to disable verification. OCSP response
5747 printout format cleaned up.
5750 *) Add additional OCSP certificate checks. These are those specified
5751 in RFC2560. This consists of two separate checks: the CA of the
5752 certificate being checked must either be the OCSP signer certificate
5753 or the issuer of the OCSP signer certificate. In the latter case the
5754 OCSP signer certificate must contain the OCSP signing extended key
5755 usage. This check is performed by attempting to match the OCSP
5756 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
5757 in the OCSP_CERTID structures of the response.
5760 *) Initial OCSP certificate verification added to OCSP_basic_verify()
5761 and related routines. This uses the standard OpenSSL certificate
5762 verify routines to perform initial checks (just CA validity) and
5763 to obtain the certificate chain. Then additional checks will be
5764 performed on the chain. Currently the root CA is checked to see
5765 if it is explicitly trusted for OCSP signing. This is used to set
5766 a root CA as a global signing root: that is any certificate that
5767 chains to that CA is an acceptable OCSP signing certificate.
5770 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5771 extensions from a separate configuration file.
5772 As when reading extensions from the main configuration file,
5773 the '-extensions ...' option may be used for specifying the
5775 [Massimiliano Pala <madwolf@comune.modena.it>]
5777 *) New OCSP utility. Allows OCSP requests to be generated or
5778 read. The request can be sent to a responder and the output
5779 parsed, outputed or printed in text form. Not complete yet:
5780 still needs to check the OCSP response validity.
5783 *) New subcommands for 'openssl ca':
5784 'openssl ca -status <serial>' prints the status of the cert with
5785 the given serial number (according to the index file).
5786 'openssl ca -updatedb' updates the expiry status of certificates
5788 [Massimiliano Pala <madwolf@comune.modena.it>]
5790 *) New '-newreq-nodes' command option to CA.pl. This is like
5791 '-newreq', but calls 'openssl req' with the '-nodes' option
5792 so that the resulting key is not encrypted.
5793 [Damien Miller <djm@mindrot.org>]
5795 *) New configuration for the GNU Hurd.
5796 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5798 *) Initial code to implement OCSP basic response verify. This
5799 is currently incomplete. Currently just finds the signer's
5800 certificate and verifies the signature on the response.
5803 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5804 value of OPENSSLDIR. This is available via the new '-d' option
5805 to 'openssl version', and is also included in 'openssl version -a'.
5808 *) Allowing defining memory allocation callbacks that will be given
5809 file name and line number information in additional arguments
5810 (a const char* and an int). The basic functionality remains, as
5811 well as the original possibility to just replace malloc(),
5812 realloc() and free() by functions that do not know about these
5813 additional arguments. To register and find out the current
5814 settings for extended allocation functions, the following
5815 functions are provided:
5817 CRYPTO_set_mem_ex_functions
5818 CRYPTO_set_locked_mem_ex_functions
5819 CRYPTO_get_mem_ex_functions
5820 CRYPTO_get_locked_mem_ex_functions
5822 These work the same way as CRYPTO_set_mem_functions and friends.
5823 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5824 extended allocation function is enabled.
5825 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5826 a conventional allocation function is enabled.
5827 [Richard Levitte, Bodo Moeller]
5829 *) Finish off removing the remaining LHASH function pointer casts.
5830 There should no longer be any prototype-casting required when using
5831 the LHASH abstraction, and any casts that remain are "bugs". See
5832 the callback types and macros at the head of lhash.h for details
5833 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5836 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5837 If /dev/[u]random devices are not available or do not return enough
5838 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5840 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5841 /etc/entropy will be queried once each in this sequence, quering stops
5842 when enough entropy was collected without querying more sockets.
5845 *) Change the Unix RAND_poll() variant to be able to poll several
5846 random devices, as specified by DEVRANDOM, until a sufficient amount
5847 of data has been collected. We spend at most 10 ms on each file
5848 (select timeout) and read in non-blocking mode. DEVRANDOM now
5849 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5850 (previously it was just the string "/dev/urandom"), so on typical
5851 platforms the 10 ms delay will never occur.
5852 Also separate out the Unix variant to its own file, rand_unix.c.
5853 For VMS, there's a currently-empty rand_vms.c.
5856 *) Move OCSP client related routines to ocsp_cl.c. These
5857 provide utility functions which an application needing
5858 to issue a request to an OCSP responder and analyse the
5859 response will typically need: as opposed to those which an
5860 OCSP responder itself would need which will be added later.
5862 OCSP_request_sign() signs an OCSP request with an API similar
5863 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5864 response. OCSP_response_get1_basic() extracts basic response
5865 from response. OCSP_resp_find_status(): finds and extracts status
5866 information from an OCSP_CERTID structure (which will be created
5867 when the request structure is built). These are built from lower
5868 level functions which work on OCSP_SINGLERESP structures but
5869 wont normally be used unless the application wishes to examine
5870 extensions in the OCSP response for example.
5872 Replace nonce routines with a pair of functions.
5873 OCSP_request_add1_nonce() adds a nonce value and optionally
5874 generates a random value. OCSP_check_nonce() checks the
5875 validity of the nonce in an OCSP response.
5878 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5879 This doesn't copy the supplied OCSP_CERTID and avoids the
5880 need to free up the newly created id. Change return type
5881 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5882 This can then be used to add extensions to the request.
5883 Deleted OCSP_request_new(), since most of its functionality
5884 is now in OCSP_REQUEST_new() (and the case insensitive name
5885 clash) apart from the ability to set the request name which
5886 will be added elsewhere.
5889 *) Update OCSP API. Remove obsolete extensions argument from
5890 various functions. Extensions are now handled using the new
5891 OCSP extension code. New simple OCSP HTTP function which
5892 can be used to send requests and parse the response.
5895 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5896 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5897 uses the special reorder version of SET OF to sort the attributes
5898 and reorder them to match the encoded order. This resolves a long
5899 standing problem: a verify on a PKCS7 structure just after signing
5900 it used to fail because the attribute order did not match the
5901 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5902 it uses the received order. This is necessary to tolerate some broken
5903 software that does not order SET OF. This is handled by encoding
5904 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5905 to produce the required SET OF.
5908 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5909 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5910 files to get correct declarations of the ASN.1 item variables.
5913 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5914 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5915 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5916 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5917 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5918 ASN1_ITEM and no wrapper functions.
5921 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5922 replace the old function pointer based I/O routines. Change most of
5923 the *_d2i_bio() and *_d2i_fp() functions to use these.
5926 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5927 lines, recognice more "algorithms" that can be deselected, and make
5928 it complain about algorithm deselection that isn't recognised.
5931 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5932 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5933 to use new functions. Add NO_ASN1_OLD which can be set to remove
5934 some old style ASN1 functions: this can be used to determine if old
5935 code will still work when these eventually go away.
5938 *) New extension functions for OCSP structures, these follow the
5939 same conventions as certificates and CRLs.
5942 *) New function X509V3_add1_i2d(). This automatically encodes and
5943 adds an extension. Its behaviour can be customised with various
5944 flags to append, replace or delete. Various wrappers added for
5945 certifcates and CRLs.
5948 *) Fix to avoid calling the underlying ASN1 print routine when
5949 an extension cannot be parsed. Correct a typo in the
5950 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5953 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5954 entries for variables.
5957 *) Add functionality to apps/openssl.c for detecting locking
5958 problems: As the program is single-threaded, all we have
5959 to do is register a locking callback using an array for
5960 storing which locks are currently held by the program.
5963 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5964 SSL_get_ex_data_X509_STORE_idx(), which is used in
5965 ssl_verify_cert_chain() and thus can be called at any time
5966 during TLS/SSL handshakes so that thread-safety is essential.
5967 Unfortunately, the ex_data design is not at all suited
5968 for multi-threaded use, so it probably should be abolished.
5971 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5972 [Broadcom, tweaked and integrated by Geoff Thorpe]
5974 *) Move common extension printing code to new function
5975 X509V3_print_extensions(). Reorganise OCSP print routines and
5976 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5979 *) New function X509_signature_print() to remove duplication in some
5983 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5984 set (this was treated exactly the same as SET OF previously). This
5985 is used to reorder the STACK representing the structure to match the
5986 encoding. This will be used to get round a problem where a PKCS7
5987 structure which was signed could not be verified because the STACK
5988 order did not reflect the encoded order.
5991 *) Reimplement the OCSP ASN1 module using the new code.
5994 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5995 for its ASN1 operations. The old style function pointers still exist
5996 for now but they will eventually go away.
5999 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
6000 completely replaces the old ASN1 functionality with a table driven
6001 encoder and decoder which interprets an ASN1_ITEM structure describing
6002 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6003 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6004 has also been converted to the new form.
6007 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6008 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6009 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6010 for negative moduli.
6013 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6014 of not touching the result's sign bit.
6017 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6021 *) Changed the LHASH code to use prototypes for callbacks, and created
6022 macros to declare and implement thin (optionally static) functions
6023 that provide type-safety and avoid function pointer casting for the
6024 type-specific callbacks.
6027 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6029 [Veers Staats <staatsvr@asc.hpc.mil>,
6030 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6032 *) Reformat the FAQ so the different questions and answers can be divided
6033 in sections depending on the subject.
6036 *) Have the zlib compression code load ZLIB.DLL dynamically under
6040 *) New function BN_mod_sqrt for computing square roots modulo a prime
6041 (using the probabilistic Tonelli-Shanks algorithm unless
6042 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6043 be handled deterministically).
6044 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6046 *) Make BN_mod_inverse faster by explicitly handling small quotients
6047 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6048 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6051 *) New function BN_kronecker.
6054 *) Fix BN_gcd so that it works on negative inputs; the result is
6055 positive unless both parameters are zero.
6056 Previously something reasonably close to an infinite loop was
6057 possible because numbers could be growing instead of shrinking
6058 in the implementation of Euclid's algorithm.
6061 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6062 sign of the number in question.
6064 Fix BN_is_word(a,w) to work correctly for w == 0.
6066 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6067 because its test if the absolute value of 'a' equals 'w'.
6068 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6069 it exists mostly for use in the implementations of BN_is_zero(),
6070 BN_is_one(), and BN_is_word().
6073 *) New function BN_swap.
6076 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6077 the exponentiation functions are more likely to produce reasonable
6078 results on negative inputs.
6081 *) Change BN_mod_mul so that the result is always non-negative.
6082 Previously, it could be negative if one of the factors was negative;
6083 I don't think anyone really wanted that behaviour.
6086 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6087 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6088 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6089 and add new functions:
6098 BN_mod_lshift1_quick
6102 These functions always generate non-negative results.
6104 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6105 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6107 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6108 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6109 be reduced modulo m.
6110 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6113 The following entry accidentily appeared in the CHANGES file
6114 distributed with OpenSSL 0.9.7. The modifications described in
6115 it do *not* apply to OpenSSL 0.9.7.
6117 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6118 was actually never needed) and in BN_mul(). The removal in BN_mul()
6119 required a small change in bn_mul_part_recursive() and the addition
6120 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6121 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6122 bn_sub_words() and bn_add_words() except they take arrays with
6127 *) In 'openssl passwd', verify passwords read from the terminal
6128 unless the '-salt' option is used (which usually means that
6129 verification would just waste user's time since the resulting
6130 hash is going to be compared with some given password hash)
6131 or the new '-noverify' option is used.
6133 This is an incompatible change, but it does not affect
6134 non-interactive use of 'openssl passwd' (passwords on the command
6135 line, '-stdin' option, '-in ...' option) and thus should not
6139 *) Remove all references to RSAref, since there's no more need for it.
6142 *) Make DSO load along a path given through an environment variable
6143 (SHLIB_PATH) with shl_load().
6146 *) Constify the ENGINE code as a result of BIGNUM constification.
6147 Also constify the RSA code and most things related to it. In a
6148 few places, most notable in the depth of the ASN.1 code, ugly
6149 casts back to non-const were required (to be solved at a later
6153 *) Make it so the openssl application has all engines loaded by default.
6156 *) Constify the BIGNUM routines a little more.
6159 *) Add the following functions:
6161 ENGINE_load_cswift()
6163 ENGINE_load_atalla()
6165 ENGINE_load_builtin_engines()
6167 That way, an application can itself choose if external engines that
6168 are built-in in OpenSSL shall ever be used or not. The benefit is
6169 that applications won't have to be linked with libdl or other dso
6170 libraries unless it's really needed.
6172 Changed 'openssl engine' to load all engines on demand.
6173 Changed the engine header files to avoid the duplication of some
6174 declarations (they differed!).
6177 *) 'openssl engine' can now list capabilities.
6180 *) Better error reporting in 'openssl engine'.
6183 *) Never call load_dh_param(NULL) in s_server.
6186 *) Add engine application. It can currently list engines by name and
6187 identity, and test if they are actually available.
6190 *) Improve RPM specification file by forcing symbolic linking and making
6191 sure the installed documentation is also owned by root.root.
6192 [Damien Miller <djm@mindrot.org>]
6194 *) Give the OpenSSL applications more possibilities to make use of
6195 keys (public as well as private) handled by engines.
6198 *) Add OCSP code that comes from CertCo.
6201 *) Add VMS support for the Rijndael code.
6204 *) Added untested support for Nuron crypto accelerator.
6207 *) Add support for external cryptographic devices. This code was
6208 previously distributed separately as the "engine" branch.
6209 [Geoff Thorpe, Richard Levitte]
6211 *) Rework the filename-translation in the DSO code. It is now possible to
6212 have far greater control over how a "name" is turned into a filename
6213 depending on the operating environment and any oddities about the
6214 different shared library filenames on each system.
6217 *) Support threads on FreeBSD-elf in Configure.
6220 *) Fix for SHA1 assembly problem with MASM: it produces
6221 warnings about corrupt line number information when assembling
6222 with debugging information. This is caused by the overlapping
6224 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
6227 NCONF_get_number() has no error checking at all. As a replacement,
6228 NCONF_get_number_e() is defined (_e for "error checking") and is
6229 promoted strongly. The old NCONF_get_number is kept around for
6230 binary backward compatibility.
6231 Make it possible for methods to load from something other than a BIO,
6232 by providing a function pointer that is given a name instead of a BIO.
6233 For example, this could be used to load configuration data from an
6237 *) Fix for non blocking accept BIOs. Added new I/O special reason
6238 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
6239 with non blocking I/O was not possible because no retry code was
6240 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
6244 *) Added the beginnings of Rijndael support.
6247 *) Fix for bug in DirectoryString mask setting. Add support for
6248 X509_NAME_print_ex() in 'req' and X509_print_ex() function
6249 to allow certificate printing to more controllable, additional
6250 'certopt' option to 'x509' to allow new printing options to be
6254 *) Clean old EAY MD5 hack from e_os.h.
6257 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
6259 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6260 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6261 [Joe Orton, Steve Henson]
6263 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
6265 *) Fix additional bug revealed by the NISCC test suite:
6267 Stop bug triggering large recursion when presented with
6268 certain ASN.1 tags (CVE-2003-0851)
6271 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
6273 *) Fix various bugs revealed by running the NISCC test suite:
6275 Stop out of bounds reads in the ASN1 code when presented with
6276 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6278 If verify callback ignores invalid public key errors don't try to check
6279 certificate signature with the NULL public key.
6283 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6284 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6288 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6289 extra data after the compression methods not only for TLS 1.0
6290 but also for SSL 3.0 (as required by the specification).
6291 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6293 *) Change X509_certificate_type() to mark the key as exported/exportable
6294 when it's 512 *bits* long, not 512 bytes.
6297 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
6299 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6300 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6301 a protocol version number mismatch like a decryption error
6302 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6305 *) Turn on RSA blinding by default in the default implementation
6306 to avoid a timing attack. Applications that don't want it can call
6307 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6308 They would be ill-advised to do so in most cases.
6309 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6311 *) Change RSA blinding code so that it works when the PRNG is not
6312 seeded (in this case, the secret RSA exponent is abused as
6313 an unpredictable seed -- if it is not unpredictable, there
6314 is no point in blinding anyway). Make RSA blinding thread-safe
6315 by remembering the creator's thread ID in rsa->blinding and
6316 having all other threads use local one-time blinding factors
6317 (this requires more computation than sharing rsa->blinding, but
6318 avoids excessive locking; and if an RSA object is not shared
6319 between threads, blinding will still be very fast).
6322 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
6324 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6325 via timing by performing a MAC computation even if incorrrect
6326 block cipher padding has been found. This is a countermeasure
6327 against active attacks where the attacker has to distinguish
6328 between bad padding and a MAC verification error. (CVE-2003-0078)
6330 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6331 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6332 Martin Vuagnoux (EPFL, Ilion)]
6334 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
6336 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
6337 memory from it's contents. This is done with a counter that will
6338 place alternating values in each byte. This can be used to solve
6339 two issues: 1) the removal of calls to memset() by highly optimizing
6340 compilers, and 2) cleansing with other values than 0, since those can
6341 be read through on certain media, for example a swap space on disk.
6344 *) Bugfix: client side session caching did not work with external caching,
6345 because the session->cipher setting was not restored when reloading
6346 from the external cache. This problem was masked, when
6347 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
6348 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
6351 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
6352 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
6353 [Zeev Lieber <zeev-l@yahoo.com>]
6355 *) Undo an undocumented change introduced in 0.9.6e which caused
6356 repeated calls to OpenSSL_add_all_ciphers() and
6357 OpenSSL_add_all_digests() to be ignored, even after calling
6361 *) Change the default configuration reader to deal with last line not
6362 being properly terminated.
6365 *) Change X509_NAME_cmp() so it applies the special rules on handling
6366 DN values that are of type PrintableString, as well as RDNs of type
6367 emailAddress where the value has the type ia5String.
6368 [stefank@valicert.com via Richard Levitte]
6370 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
6371 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
6372 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
6373 the bitwise-OR of the two for use by the majority of applications
6374 wanting this behaviour, and update the docs. The documented
6375 behaviour and actual behaviour were inconsistent and had been
6376 changing anyway, so this is more a bug-fix than a behavioural
6378 [Geoff Thorpe, diagnosed by Nadav Har'El]
6380 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
6381 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
6384 *) Fix initialization code race conditions in
6385 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
6386 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
6387 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
6388 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
6389 ssl2_get_cipher_by_char(),
6390 ssl3_get_cipher_by_char().
6391 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
6393 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
6394 the cached sessions are flushed, as the remove_cb() might use ex_data
6395 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
6396 (see [openssl.org #212]).
6397 [Geoff Thorpe, Lutz Jaenicke]
6399 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
6400 length, instead of the encoding length to d2i_ASN1_OBJECT.
6403 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
6405 *) [In 0.9.6g-engine release:]
6406 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
6407 [Lynn Gazis <lgazis@rainbow.com>]
6409 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
6411 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
6412 and get fix the header length calculation.
6413 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
6414 Alon Kantor <alonk@checkpoint.com> (and others),
6417 *) Use proper error handling instead of 'assertions' in buffer
6418 overflow checks added in 0.9.6e. This prevents DoS (the
6419 assertions could call abort()).
6420 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
6422 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
6424 *) Add various sanity checks to asn1_get_length() to reject
6425 the ASN1 length bytes if they exceed sizeof(long), will appear
6426 negative or the content length exceeds the length of the
6428 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6430 *) Fix cipher selection routines: ciphers without encryption had no flags
6431 for the cipher strength set and where therefore not handled correctly
6432 by the selection routines (PR #130).
6435 *) Fix EVP_dsa_sha macro.
6439 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
6440 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
6441 that was added in OpenSSL 0.9.6d.
6443 As the countermeasure turned out to be incompatible with some
6444 broken SSL implementations, the new option is part of SSL_OP_ALL.
6445 SSL_OP_ALL is usually employed when compatibility with weird SSL
6446 implementations is desired (e.g. '-bugs' option to 's_client' and
6447 's_server'), so the new option is automatically set in many
6451 *) Changes in security patch:
6453 Changes marked "(CHATS)" were sponsored by the Defense Advanced
6454 Research Projects Agency (DARPA) and Air Force Research Laboratory,
6455 Air Force Materiel Command, USAF, under agreement number
6458 *) Add various sanity checks to asn1_get_length() to reject
6459 the ASN1 length bytes if they exceed sizeof(long), will appear
6460 negative or the content length exceeds the length of the
6461 supplied buffer. (CVE-2002-0659)
6462 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6464 *) Assertions for various potential buffer overflows, not known to
6466 [Ben Laurie (CHATS)]
6468 *) Various temporary buffers to hold ASCII versions of integers were
6469 too small for 64 bit platforms. (CVE-2002-0655)
6470 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6472 *) Remote buffer overflow in SSL3 protocol - an attacker could
6473 supply an oversized session ID to a client. (CVE-2002-0656)
6474 [Ben Laurie (CHATS)]
6476 *) Remote buffer overflow in SSL2 protocol - an attacker could
6477 supply an oversized client master key. (CVE-2002-0656)
6478 [Ben Laurie (CHATS)]
6480 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6482 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6483 encoded as NULL) with id-dsa-with-sha1.
6484 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6486 *) Check various X509_...() return values in apps/req.c.
6487 [Nils Larsch <nla@trustcenter.de>]
6489 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6490 an end-of-file condition would erronously be flagged, when the CRLF
6491 was just at the end of a processed block. The bug was discovered when
6492 processing data through a buffering memory BIO handing the data to a
6493 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6494 <ptsekov@syntrex.com> and Nedelcho Stanev.
6497 *) Implement a countermeasure against a vulnerability recently found
6498 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6499 before application data chunks to avoid the use of known IVs
6500 with data potentially chosen by the attacker.
6503 *) Fix length checks in ssl3_get_client_hello().
6506 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6507 to prevent ssl3_read_internal() from incorrectly assuming that
6508 ssl3_read_bytes() found application data while handshake
6509 processing was enabled when in fact s->s3->in_read_app_data was
6510 merely automatically cleared during the initial handshake.
6511 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6513 *) Fix object definitions for Private and Enterprise: they were not
6514 recognized in their shortname (=lowercase) representation. Extend
6515 obj_dat.pl to issue an error when using undefined keywords instead
6516 of silently ignoring the problem (Svenning Sorensen
6517 <sss@sss.dnsalias.net>).
6520 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6521 generators, i.e. generators other than 2 and 5. (Previously, the
6522 code did not properly initialise the 'add' and 'rem' values to
6523 BN_generate_prime().)
6525 In the new general case, we do not insist that 'generator' is
6526 actually a primitive root: This requirement is rather pointless;
6527 a generator of the order-q subgroup is just as good, if not
6531 *) Map new X509 verification errors to alerts. Discovered and submitted by
6532 Tom Wu <tom@arcot.com>.
6535 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6536 returning non-zero before the data has been completely received
6537 when using non-blocking I/O.
6538 [Bodo Moeller; problem pointed out by John Hughes]
6540 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6541 [Ben Laurie, Lutz Jaenicke]
6543 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6544 Yoram Zahavi <YoramZ@gilian.com>).
6547 *) Add information about CygWin 1.3 and on, and preserve proper
6548 configuration for the versions before that.
6549 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6551 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6552 check whether we deal with a copy of a session and do not delete from
6553 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6554 <izhar@checkpoint.com>.
6557 *) Do not store session data into the internal session cache, if it
6558 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6559 flag is set). Proposed by Aslam <aslam@funk.com>.
6562 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6566 *) [In 0.9.6d-engine release:]
6567 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6568 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6570 *) Add the configuration target linux-s390x.
6571 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6573 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6574 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6575 variable as an indication that a ClientHello message has been
6576 received. As the flag value will be lost between multiple
6577 invocations of ssl3_accept when using non-blocking I/O, the
6578 function may not be aware that a handshake has actually taken
6579 place, thus preventing a new session from being added to the
6582 To avoid this problem, we now set s->new_session to 2 instead of
6583 using a local variable.
6584 [Lutz Jaenicke, Bodo Moeller]
6586 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6587 if the SSL_R_LENGTH_MISMATCH error is detected.
6588 [Geoff Thorpe, Bodo Moeller]
6590 *) New 'shared_ldflag' column in Configure platform table.
6593 *) Fix EVP_CIPHER_mode macro.
6594 ["Dan S. Camper" <dan@bti.net>]
6596 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
6597 type, we must throw them away by setting rr->length to 0.
6598 [D P Chang <dpc@qualys.com>]
6600 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
6602 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
6603 <Dominikus.Scherkl@biodata.com>. (The previous implementation
6604 worked incorrectly for those cases where range = 10..._2 and
6605 3*range is two bits longer than range.)
6608 *) Only add signing time to PKCS7 structures if it is not already
6612 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
6613 OBJ_ld_ce should be OBJ_id_ce.
6614 Also some ip-pda OIDs in crypto/objects/objects.txt were
6615 incorrect (cf. RFC 3039).
6616 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
6618 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
6619 returns early because it has nothing to do.
6620 [Andy Schneider <andy.schneider@bjss.co.uk>]
6622 *) [In 0.9.6c-engine release:]
6623 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
6624 [Andy Schneider <andy.schneider@bjss.co.uk>]
6626 *) [In 0.9.6c-engine release:]
6627 Add support for Cryptographic Appliance's keyserver technology.
6628 (Use engine 'keyclient')
6629 [Cryptographic Appliances and Geoff Thorpe]
6631 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
6632 is called via tools/c89.sh because arguments have to be
6633 rearranged (all '-L' options must appear before the first object
6635 [Richard Shapiro <rshapiro@abinitio.com>]
6637 *) [In 0.9.6c-engine release:]
6638 Add support for Broadcom crypto accelerator cards, backported
6640 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
6642 *) [In 0.9.6c-engine release:]
6643 Add support for SureWare crypto accelerator cards from
6644 Baltimore Technologies. (Use engine 'sureware')
6645 [Baltimore Technologies and Mark Cox]
6647 *) [In 0.9.6c-engine release:]
6648 Add support for crypto accelerator cards from Accelerated
6649 Encryption Processing, www.aep.ie. (Use engine 'aep')
6650 [AEP Inc. and Mark Cox]
6652 *) Add a configuration entry for gcc on UnixWare.
6653 [Gary Benson <gbenson@redhat.com>]
6655 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
6656 messages are stored in a single piece (fixed-length part and
6657 variable-length part combined) and fix various bugs found on the way.
6660 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
6661 instead. BIO_gethostbyname() does not know what timeouts are
6662 appropriate, so entries would stay in cache even when they have
6664 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
6666 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
6667 faced with a pathologically small ClientHello fragment that does
6668 not contain client_version: Instead of aborting with an error,
6669 simply choose the highest available protocol version (i.e.,
6670 TLS 1.0 unless it is disabled). In practice, ClientHello
6671 messages are never sent like this, but this change gives us
6672 strictly correct behaviour at least for TLS.
6675 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
6676 never resets s->method to s->ctx->method when called from within
6677 one of the SSL handshake functions.
6678 [Bodo Moeller; problem pointed out by Niko Baric]
6680 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
6681 (sent using the client's version number) if client_version is
6682 smaller than the protocol version in use. Also change
6683 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
6684 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
6685 the client will at least see that alert.
6688 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
6692 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
6693 client receives HelloRequest while in a handshake.
6694 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
6696 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
6697 should end in 'break', not 'goto end' which circuments various
6698 cleanups done in state SSL_ST_OK. But session related stuff
6699 must be disabled for SSL_ST_OK in the case that we just sent a
6702 Also avoid some overhead by not calling ssl_init_wbio_buffer()
6703 before just sending a HelloRequest.
6704 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
6706 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
6707 reveal whether illegal block cipher padding was found or a MAC
6708 verification error occured. (Neither SSLerr() codes nor alerts
6709 are directly visible to potential attackers, but the information
6710 may leak via logfiles.)
6712 Similar changes are not required for the SSL 2.0 implementation
6713 because the number of padding bytes is sent in clear for SSL 2.0,
6714 and the extra bytes are just ignored. However ssl/s2_pkt.c
6715 failed to verify that the purported number of padding bytes is in
6719 *) Add OpenUNIX-8 support including shared libraries
6720 (Boyd Lynn Gerber <gerberb@zenez.com>).
6723 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
6724 'wristwatch attack' using huge encoding parameters (cf.
6725 James H. Manger's CRYPTO 2001 paper). Note that the
6726 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
6727 encoding parameters and hence was not vulnerable.
6730 *) BN_sqr() bug fix.
6731 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
6733 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
6734 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
6735 followed by modular reduction.
6736 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
6738 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
6739 equivalent based on BN_pseudo_rand() instead of BN_rand().
6742 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
6743 This function was broken, as the check for a new client hello message
6744 to handle SGC did not allow these large messages.
6745 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
6748 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
6751 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
6752 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
6755 *) Rework the configuration and shared library support for Tru64 Unix.
6756 The configuration part makes use of modern compiler features and
6757 still retains old compiler behavior for those that run older versions
6758 of the OS. The shared library support part includes a variant that
6759 uses the RPATH feature, and is available through the special
6760 configuration target "alpha-cc-rpath", which will never be selected
6762 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
6764 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
6765 with the same message size as in ssl3_get_certificate_request().
6766 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
6767 messages might inadvertently be reject as too long.
6768 [Petr Lampa <lampa@fee.vutbr.cz>]
6770 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6773 *) Modified SSL library such that the verify_callback that has been set
6774 specificly for an SSL object with SSL_set_verify() is actually being
6775 used. Before the change, a verify_callback set with this function was
6776 ignored and the verify_callback() set in the SSL_CTX at the time of
6777 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6778 to allow the necessary settings.
6781 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6782 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6783 done automatically (in contradiction to the requirements of the C
6784 standard). This made problems when used from OpenSSH.
6787 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6788 dh->length and always used
6790 BN_rand_range(priv_key, dh->p).
6792 BN_rand_range() is not necessary for Diffie-Hellman, and this
6793 specific range makes Diffie-Hellman unnecessarily inefficient if
6794 dh->length (recommended exponent length) is much smaller than the
6795 length of dh->p. We could use BN_rand_range() if the order of
6796 the subgroup was stored in the DH structure, but we only have
6801 BN_rand(priv_key, l, ...)
6803 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6809 RSA_eay_public_encrypt
6810 RSA_eay_private_decrypt
6811 RSA_eay_private_encrypt (signing)
6812 RSA_eay_public_decrypt (signature verification)
6814 (default implementations for RSA_public_encrypt,
6815 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6816 always reject numbers >= n.
6819 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6820 to synchronize access to 'locking_thread'. This is necessary on
6821 systems where access to 'locking_thread' (an 'unsigned long'
6822 variable) is not atomic.
6825 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6826 *before* setting the 'crypto_lock_rand' flag. The previous code had
6827 a race condition if 0 is a valid thread ID.
6828 [Travis Vitek <vitek@roguewave.com>]
6830 *) Add support for shared libraries under Irix.
6831 [Albert Chin-A-Young <china@thewrittenword.com>]
6833 *) Add configuration option to build on Linux on both big-endian and
6835 [Ralf Baechle <ralf@uni-koblenz.de>]
6837 *) Add the possibility to create shared libraries on HP-UX.
6840 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6842 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6843 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6844 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6845 PRNG state recovery was possible based on the output of
6846 one PRNG request appropriately sized to gain knowledge on
6847 'md' followed by enough consecutive 1-byte PRNG requests
6848 to traverse all of 'state'.
6850 1. When updating 'md_local' (the current thread's copy of 'md')
6851 during PRNG output generation, hash all of the previous
6852 'md_local' value, not just the half used for PRNG output.
6854 2. Make the number of bytes from 'state' included into the hash
6855 independent from the number of PRNG bytes requested.
6857 The first measure alone would be sufficient to avoid
6858 Markku-Juhani's attack. (Actually it had never occurred
6859 to me that the half of 'md_local' used for chaining was the
6860 half from which PRNG output bytes were taken -- I had always
6861 assumed that the secret half would be used.) The second
6862 measure makes sure that additional data from 'state' is never
6863 mixed into 'md_local' in small portions; this heuristically
6864 further strengthens the PRNG.
6867 *) Fix crypto/bn/asm/mips3.s.
6870 *) When only the key is given to "enc", the IV is undefined. Print out
6871 an error message in this case.
6874 *) Handle special case when X509_NAME is empty in X509 printing routines.
6877 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6878 positive and less than q.
6881 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6882 used: it isn't thread safe and the add_lock_callback should handle
6884 [Paul Rose <Paul.Rose@bridge.com>]
6886 *) Verify that incoming data obeys the block size in
6887 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6891 [Ulf Möller, Bodo Möller]
6893 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6894 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6895 when fixing the server behaviour for backwards-compatible 'client
6896 hello' messages. (Note that the attack is impractical against
6897 SSL 3.0 and TLS 1.0 anyway because length and version checking
6898 means that the probability of guessing a valid ciphertext is
6899 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6902 Before 0.9.5, the countermeasure (hide the error by generating a
6903 random 'decryption result') did not work properly because
6904 ERR_clear_error() was missing, meaning that SSL_get_error() would
6905 detect the supposedly ignored error.
6907 Both problems are now fixed.
6910 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6911 (previously it was 1024).
6914 *) Fix for compatibility mode trust settings: ignore trust settings
6915 unless some valid trust or reject settings are present.
6918 *) Fix for blowfish EVP: its a variable length cipher.
6921 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6922 parameters in DSA public key structures and return an error in the
6923 DSA routines if parameters are absent.
6926 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6927 in the current directory if neither $RANDFILE nor $HOME was set.
6928 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6929 caused some confusion to Windows users who haven't defined $HOME.
6930 Thus RAND_file_name() is changed again: e_os.h can define a
6931 DEFAULT_HOME, which will be used if $HOME is not set.
6932 For Windows, we use "C:"; on other platforms, we still require
6933 environment variables.
6935 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6936 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6937 having multiple threads call RAND_poll() concurrently.
6940 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6941 combination of a flag and a thread ID variable.
6942 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6943 flag), *other* threads can enter ssleay_add_bytes without obeying
6944 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6945 that they do not hold after the first thread unsets add_do_not_lock).
6948 *) Change bctest again: '-x' expressions are not available in all
6952 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6954 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6955 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6957 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6958 the default extension for executables, if any. Also, make the perl
6959 scripts that use symlink() to test if it really exists and use "cp"
6960 if it doesn't. All this made OpenSSL compilable and installable in
6964 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6965 If SEQUENCE is length is indefinite just set c->slen to the total
6966 amount of data available.
6967 [Steve Henson, reported by shige@FreeBSD.org]
6968 [This change does not apply to 0.9.7.]
6970 *) Change bctest to avoid here-documents inside command substitution
6971 (workaround for FreeBSD /bin/sh bug).
6972 For compatibility with Ultrix, avoid shell functions (introduced
6973 in the bctest version that searches along $PATH).
6976 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6977 with des_encrypt() defined on some operating systems, like Solaris
6981 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6982 On the Importance of Eliminating Errors in Cryptographic
6983 Computations, J. Cryptology 14 (2001) 2, 101-119,
6984 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6987 *) MIPS assembler BIGNUM division bug fix.
6990 *) Disabled incorrect Alpha assembler code.
6993 *) Fix PKCS#7 decode routines so they correctly update the length
6994 after reading an EOC for the EXPLICIT tag.
6996 [This change does not apply to 0.9.7.]
6998 *) Fix bug in PKCS#12 key generation routines. This was triggered
6999 if a 3DES key was generated with a 0 initial byte. Include
7000 PKCS12_BROKEN_KEYGEN compilation option to retain the old
7001 (but broken) behaviour.
7004 *) Enhance bctest to search for a working bc along $PATH and print
7006 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7008 *) Fix memory leaks in err.c: free err_data string if necessary;
7009 don't write to the wrong index in ERR_set_error_data.
7012 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7016 *) Replace rdtsc with _emit statements for VC++ version 5.
7017 [Jeremy Cooper <jeremy@baymoo.org>]
7019 *) Make it possible to reuse SSLv2 sessions.
7022 *) In copy_email() check for >= 0 as a return value for
7023 X509_NAME_get_index_by_NID() since 0 is a valid index.
7024 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7026 *) Avoid coredump with unsupported or invalid public keys by checking if
7027 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7028 PKCS7_verify() fails with non detached data.
7031 *) Don't use getenv in library functions when run as setuid/setgid.
7032 New function OPENSSL_issetugid().
7035 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7036 due to incorrect handling of multi-threading:
7038 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7040 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7042 3. Count how many times MemCheck_off() has been called so that
7043 nested use can be treated correctly. This also avoids
7044 inband-signalling in the previous code (which relied on the
7045 assumption that thread ID 0 is impossible).
7048 *) Add "-rand" option also to s_client and s_server.
7051 *) Fix CPU detection on Irix 6.x.
7052 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7053 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7055 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7058 [This change does not apply to 0.9.7.]
7060 *) Use the cached encoding of an X509_NAME structure rather than
7061 copying it. This is apparently the reason for the libsafe "errors"
7062 but the code is actually correct.
7065 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7066 Bleichenbacher's DSA attack.
7067 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7068 to be set and top=0 forces the highest bit to be set; top=-1 is new
7069 and leaves the highest bit random.
7070 [Ulf Moeller, Bodo Moeller]
7072 *) In the NCONF_...-based implementations for CONF_... queries
7073 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7074 a temporary CONF structure with the data component set to NULL
7075 (which gives segmentation faults in lh_retrieve).
7076 Instead, use NULL for the CONF pointer in CONF_get_string and
7077 CONF_get_number (which may use environment variables) and directly
7078 return NULL from CONF_get_section.
7081 *) Fix potential buffer overrun for EBCDIC.
7084 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7085 keyUsage if basicConstraints absent for a CA.
7088 *) Make SMIME_write_PKCS7() write mail header values with a format that
7089 is more generally accepted (no spaces before the semicolon), since
7090 some programs can't parse those values properly otherwise. Also make
7091 sure BIO's that break lines after each write do not create invalid
7095 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7096 macros previously used would not encode an empty SEQUENCE OF
7097 and break the signature.
7099 [This change does not apply to 0.9.7.]
7101 *) Zero the premaster secret after deriving the master secret in
7105 *) Add some EVP_add_digest_alias registrations (as found in
7106 OpenSSL_add_all_digests()) to SSL_library_init()
7107 aka OpenSSL_add_ssl_algorithms(). This provides improved
7108 compatibility with peers using X.509 certificates
7109 with unconventional AlgorithmIdentifier OIDs.
7112 *) Fix for Irix with NO_ASM.
7113 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7115 *) ./config script fixes.
7116 [Ulf Moeller, Richard Levitte]
7118 *) Fix 'openssl passwd -1'.
7121 *) Change PKCS12_key_gen_asc() so it can cope with non null
7122 terminated strings whose length is passed in the passlen
7123 parameter, for example from PEM callbacks. This was done
7124 by adding an extra length parameter to asc2uni().
7125 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7127 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7128 call failed, free the DSA structure.
7131 *) Fix to uni2asc() to cope with zero length Unicode strings.
7132 These are present in some PKCS#12 files.
7135 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7136 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7137 when writing a 32767 byte record.
7138 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7140 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7141 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7143 (RSA objects have a reference count access to which is protected
7144 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7145 so they are meant to be shared between threads.)
7146 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7147 "Reddie, Steven" <Steven.Reddie@ca.com>]
7149 *) Fix a deadlock in CRYPTO_mem_leaks().
7152 *) Use better test patterns in bntest.
7155 *) rand_win.c fix for Borland C.
7158 *) BN_rshift bugfix for n == 0.
7161 *) Add a 'bctest' script that checks for some known 'bc' bugs
7162 so that 'make test' does not abort just because 'bc' is broken.
7165 *) Store verify_result within SSL_SESSION also for client side to
7166 avoid potential security hole. (Re-used sessions on the client side
7167 always resulted in verify_result==X509_V_OK, not using the original
7168 result of the server certificate verification.)
7171 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
7172 SSL3_RT_APPLICATION_DATA, return 0.
7173 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
7177 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
7178 releases, have been re-implemented by renaming the previous
7179 implementations of ssl2_read and ssl3_read to ssl2_read_internal
7180 and ssl3_read_internal, respectively, and adding 'peek' parameters
7181 to them. The new ssl[23]_{read,peek} functions are calls to
7182 ssl[23]_read_internal with the 'peek' flag set appropriately.
7183 A 'peek' parameter has also been added to ssl3_read_bytes, which
7184 does the actual work for ssl3_read_internal.
7187 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
7188 the method-specific "init()" handler. Also clean up ex_data after
7189 calling the method-specific "finish()" handler. Previously, this was
7190 happening the other way round.
7193 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
7194 The previous value, 12, was not always sufficient for BN_mod_exp().
7197 *) Make sure that shared libraries get the internal name engine with
7198 the full version number and not just 0. This should mark the
7199 shared libraries as not backward compatible. Of course, this should
7200 be changed again when we can guarantee backward binary compatibility.
7203 *) Fix typo in get_cert_by_subject() in by_dir.c
7204 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
7206 *) Rework the system to generate shared libraries:
7208 - Make note of the expected extension for the shared libraries and
7209 if there is a need for symbolic links from for example libcrypto.so.0
7210 to libcrypto.so.0.9.7. There is extended info in Configure for
7213 - Make as few rebuilds of the shared libraries as possible.
7215 - Still avoid linking the OpenSSL programs with the shared libraries.
7217 - When installing, install the shared libraries separately from the
7221 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
7223 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
7224 and not in SSL_clear because the latter is also used by the
7225 accept/connect functions; previously, the settings made by
7226 SSL_set_read_ahead would be lost during the handshake.
7227 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
7229 *) Correct util/mkdef.pl to be selective about disabled algorithms.
7230 Previously, it would create entries for disableed algorithms no
7234 *) Added several new manual pages for SSL_* function.
7237 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
7239 *) In ssl23_get_client_hello, generate an error message when faced
7240 with an initial SSL 3.0/TLS record that is too small to contain the
7241 first two bytes of the ClientHello message, i.e. client_version.
7242 (Note that this is a pathologic case that probably has never happened
7243 in real life.) The previous approach was to use the version number
7244 from the record header as a substitute; but our protocol choice
7245 should not depend on that one because it is not authenticated
7246 by the Finished messages.
7249 *) More robust randomness gathering functions for Windows.
7250 [Jeffrey Altman <jaltman@columbia.edu>]
7252 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
7253 not set then we don't setup the error code for issuer check errors
7254 to avoid possibly overwriting other errors which the callback does
7255 handle. If an application does set the flag then we assume it knows
7256 what it is doing and can handle the new informational codes
7260 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
7261 a general "ANY" type, as such it should be able to decode anything
7262 including tagged types. However it didn't check the class so it would
7263 wrongly interpret tagged types in the same way as their universal
7264 counterpart and unknown types were just rejected. Changed so that the
7265 tagged and unknown types are handled in the same way as a SEQUENCE:
7266 that is the encoding is stored intact. There is also a new type
7267 "V_ASN1_OTHER" which is used when the class is not universal, in this
7268 case we have no idea what the actual type is so we just lump them all
7272 *) On VMS, stdout may very well lead to a file that is written to
7273 in a record-oriented fashion. That means that every write() will
7274 write a separate record, which will be read separately by the
7275 programs trying to read from it. This can be very confusing.
7277 The solution is to put a BIO filter in the way that will buffer
7278 text until a linefeed is reached, and then write everything a
7279 line at a time, so every record written will be an actual line,
7280 not chunks of lines and not (usually doesn't happen, but I've
7281 seen it once) several lines in one record. BIO_f_linebuffer() is
7284 Currently, it's a VMS-only method, because that's where it has
7285 been tested well enough.
7288 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
7289 it can return incorrect results.
7290 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
7291 but it was in 0.9.6-beta[12].)
7294 *) Disable the check for content being present when verifying detached
7295 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
7296 include zero length content when signing messages.
7299 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
7300 BIO_ctrl (for BIO pairs).
7303 *) Add DSO method for VMS.
7306 *) Bug fix: Montgomery multiplication could produce results with the
7310 *) Add RPM specification openssl.spec and modify it to build three
7311 packages. The default package contains applications, application
7312 documentation and run-time libraries. The devel package contains
7313 include files, static libraries and function documentation. The
7314 doc package contains the contents of the doc directory. The original
7315 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
7318 *) Add a large number of documentation files for many SSL routines.
7319 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7321 *) Add a configuration entry for Sony News 4.
7322 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
7324 *) Don't set the two most significant bits to one when generating a
7325 random number < q in the DSA library.
7328 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
7329 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
7330 the underlying transport is blocking) if a handshake took place.
7331 (The default behaviour is needed by applications such as s_client
7332 and s_server that use select() to determine when to use SSL_read;
7333 but for applications that know in advance when to expect data, it
7334 just makes things more complicated.)
7337 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
7341 *) Add a few more EBCDIC conditionals that make `req' and `x509'
7342 work better on such systems.
7343 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7345 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
7346 Update PKCS12_parse() so it copies the friendlyName and the
7347 keyid to the certificates aux info.
7350 *) Fix bug in PKCS7_verify() which caused an infinite loop
7351 if there was more than one signature.
7352 [Sven Uszpelkat <su@celocom.de>]
7354 *) Major change in util/mkdef.pl to include extra information
7355 about each symbol, as well as presentig variables as well
7356 as functions. This change means that there's n more need
7357 to rebuild the .num files when some algorithms are excluded.
7360 *) Allow the verify time to be set by an application,
7361 rather than always using the current time.
7364 *) Phase 2 verify code reorganisation. The certificate
7365 verify code now looks up an issuer certificate by a
7366 number of criteria: subject name, authority key id
7367 and key usage. It also verifies self signed certificates
7368 by the same criteria. The main comparison function is
7369 X509_check_issued() which performs these checks.
7371 Lot of changes were necessary in order to support this
7372 without completely rewriting the lookup code.
7374 Authority and subject key identifier are now cached.
7376 The LHASH 'certs' is X509_STORE has now been replaced
7377 by a STACK_OF(X509_OBJECT). This is mainly because an
7378 LHASH can't store or retrieve multiple objects with
7379 the same hash value.
7381 As a result various functions (which were all internal
7382 use only) have changed to handle the new X509_STORE
7383 structure. This will break anything that messed round
7384 with X509_STORE internally.
7386 The functions X509_STORE_add_cert() now checks for an
7387 exact match, rather than just subject name.
7389 The X509_STORE API doesn't directly support the retrieval
7390 of multiple certificates matching a given criteria, however
7391 this can be worked round by performing a lookup first
7392 (which will fill the cache with candidate certificates)
7393 and then examining the cache for matches. This is probably
7394 the best we can do without throwing out X509_LOOKUP
7395 entirely (maybe later...).
7397 The X509_VERIFY_CTX structure has been enhanced considerably.
7399 All certificate lookup operations now go via a get_issuer()
7400 callback. Although this currently uses an X509_STORE it
7401 can be replaced by custom lookups. This is a simple way
7402 to bypass the X509_STORE hackery necessary to make this
7403 work and makes it possible to use more efficient techniques
7404 in future. A very simple version which uses a simple
7405 STACK for its trusted certificate store is also provided
7406 using X509_STORE_CTX_trusted_stack().
7408 The verify_cb() and verify() callbacks now have equivalents
7409 in the X509_STORE_CTX structure.
7411 X509_STORE_CTX also has a 'flags' field which can be used
7412 to customise the verify behaviour.
7415 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
7416 excludes S/MIME capabilities.
7419 *) When a certificate request is read in keep a copy of the
7420 original encoding of the signed data and use it when outputing
7421 again. Signatures then use the original encoding rather than
7422 a decoded, encoded version which may cause problems if the
7423 request is improperly encoded.
7426 *) For consistency with other BIO_puts implementations, call
7427 buffer_write(b, ...) directly in buffer_puts instead of calling
7430 In BIO_puts, increment b->num_write as in BIO_write.
7431 [Peter.Sylvester@EdelWeb.fr]
7433 *) Fix BN_mul_word for the case where the word is 0. (We have to use
7434 BN_zero, we may not return a BIGNUM with an array consisting of
7438 *) Avoid calling abort() from within the library when problems are
7439 detected, except if preprocessor symbols have been defined
7440 (such as REF_CHECK, BN_DEBUG etc.).
7443 *) New openssl application 'rsautl'. This utility can be
7444 used for low level RSA operations. DER public key
7445 BIO/fp routines also added.
7448 *) New Configure entry and patches for compiling on QNX 4.
7449 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
7451 *) A demo state-machine implementation was sponsored by
7452 Nuron (http://www.nuron.com/) and is now available in
7453 demos/state_machine.
7456 *) New options added to the 'dgst' utility for signature
7457 generation and verification.
7460 *) Unrecognized PKCS#7 content types are now handled via a
7461 catch all ASN1_TYPE structure. This allows unsupported
7462 types to be stored as a "blob" and an application can
7463 encode and decode it manually.
7466 *) Fix various signed/unsigned issues to make a_strex.c
7468 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7470 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7471 length if passed a buffer. ASN1_INTEGER_to_BN failed
7472 if passed a NULL BN and its argument was negative.
7473 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7475 *) Modification to PKCS#7 encoding routines to output definite
7476 length encoding. Since currently the whole structures are in
7477 memory there's not real point in using indefinite length
7478 constructed encoding. However if OpenSSL is compiled with
7479 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7482 *) Added BIO_vprintf() and BIO_vsnprintf().
7485 *) Added more prefixes to parse for in the the strings written
7486 through a logging bio, to cover all the levels that are available
7487 through syslog. The prefixes are now:
7489 PANIC, EMERG, EMR => LOG_EMERG
7490 ALERT, ALR => LOG_ALERT
7491 CRIT, CRI => LOG_CRIT
7492 ERROR, ERR => LOG_ERR
7493 WARNING, WARN, WAR => LOG_WARNING
7494 NOTICE, NOTE, NOT => LOG_NOTICE
7495 INFO, INF => LOG_INFO
7496 DEBUG, DBG => LOG_DEBUG
7498 and as before, if none of those prefixes are present at the
7499 beginning of the string, LOG_ERR is chosen.
7501 On Win32, the LOG_* levels are mapped according to this:
7503 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7504 LOG_WARNING => EVENTLOG_WARNING_TYPE
7505 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7509 *) Made it possible to reconfigure with just the configuration
7510 argument "reconf" or "reconfigure". The command line arguments
7511 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7512 and are retrieved from there when reconfiguring.
7516 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7518 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7521 *) The obj_dat.pl script was messing up the sorting of object
7522 names. The reason was that it compared the quoted version
7523 of strings as a result "OCSP" > "OCSP Signing" because
7524 " > SPACE. Changed script to store unquoted versions of
7525 names and add quotes on output. It was also omitting some
7526 names from the lookup table if they were given a default
7527 value (that is if SN is missing it is given the same
7528 value as LN and vice versa), these are now added on the
7529 grounds that if an object has a name we should be able to
7530 look it up. Finally added warning output when duplicate
7531 short or long names are found.
7534 *) Changes needed for Tandem NSK.
7535 [Scott Uroff <scott@xypro.com>]
7537 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7538 RSA_padding_check_SSLv23(), special padding was never detected
7539 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7540 version rollback attacks was not effective.
7542 In s23_clnt.c, don't use special rollback-attack detection padding
7543 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7544 client; similarly, in s23_srvr.c, don't do the rollback check if
7545 SSL 2.0 is the only protocol enabled in the server.
7548 *) Make it possible to get hexdumps of unprintable data with 'openssl
7549 asn1parse'. By implication, the functions ASN1_parse_dump() and
7550 BIO_dump_indent() are added.
7553 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7554 these print out strings and name structures based on various
7555 flags including RFC2253 support and proper handling of
7556 multibyte characters. Added options to the 'x509' utility
7557 to allow the various flags to be set.
7560 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7561 Also change the functions X509_cmp_current_time() and
7562 X509_gmtime_adj() work with an ASN1_TIME structure,
7563 this will enable certificates using GeneralizedTime in validity
7564 dates to be checked.
7567 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7568 negative public key encodings) on by default,
7569 NO_NEG_PUBKEY_BUG can be set to disable it.
7572 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7573 content octets. An i2c_ASN1_OBJECT is unnecessary because
7574 the encoding can be trivially obtained from the structure.
7577 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7578 not read locks (CRYPTO_r_[un]lock).
7581 *) A first attempt at creating official support for shared
7582 libraries through configuration. I've kept it so the
7583 default is static libraries only, and the OpenSSL programs
7584 are always statically linked for now, but there are
7585 preparations for dynamic linking in place.
7586 This has been tested on Linux and Tru64.
7589 *) Randomness polling function for Win9x, as described in:
7590 Peter Gutmann, Software Generation of Practically Strong
7594 *) Fix so PRNG is seeded in req if using an already existing
7598 *) New options to smime application. -inform and -outform
7599 allow alternative formats for the S/MIME message including
7600 PEM and DER. The -content option allows the content to be
7601 specified separately. This should allow things like Netscape
7602 form signing output easier to verify.
7605 *) Fix the ASN1 encoding of tags using the 'long form'.
7608 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
7609 STRING types. These convert content octets to and from the
7610 underlying type. The actual tag and length octets are
7611 already assumed to have been read in and checked. These
7612 are needed because all other string types have virtually
7613 identical handling apart from the tag. By having versions
7614 of the ASN1 functions that just operate on content octets
7615 IMPLICIT tagging can be handled properly. It also allows
7616 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
7617 and ASN1_INTEGER are identical apart from the tag.
7620 *) Change the handling of OID objects as follows:
7622 - New object identifiers are inserted in objects.txt, following
7623 the syntax given in objects.README.
7624 - objects.pl is used to process obj_mac.num and create a new
7626 - obj_dat.pl is used to create a new obj_dat.h, using the data in
7629 This is currently kind of a hack, and the perl code in objects.pl
7630 isn't very elegant, but it works as I intended. The simplest way
7631 to check that it worked correctly is to look in obj_dat.h and
7632 check the array nid_objs and make sure the objects haven't moved
7633 around (this is important!). Additions are OK, as well as
7634 consistent name changes.
7637 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
7640 *) Addition of the command line parameter '-rand file' to 'openssl req'.
7641 The given file adds to whatever has already been seeded into the
7642 random pool through the RANDFILE configuration file option or
7643 environment variable, or the default random state file.
7646 *) mkstack.pl now sorts each macro group into lexical order.
7647 Previously the output order depended on the order the files
7648 appeared in the directory, resulting in needless rewriting
7652 *) Patches to make OpenSSL compile under Win32 again. Mostly
7653 work arounds for the VC++ problem that it treats func() as
7654 func(void). Also stripped out the parts of mkdef.pl that
7655 added extra typesafe functions: these no longer exist.
7658 *) Reorganisation of the stack code. The macros are now all
7659 collected in safestack.h . Each macro is defined in terms of
7660 a "stack macro" of the form SKM_<name>(type, a, b). The
7661 DEBUG_SAFESTACK is now handled in terms of function casts,
7662 this has the advantage of retaining type safety without the
7663 use of additional functions. If DEBUG_SAFESTACK is not defined
7664 then the non typesafe macros are used instead. Also modified the
7665 mkstack.pl script to handle the new form. Needs testing to see
7666 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
7667 the default if no major problems. Similar behaviour for ASN1_SET_OF
7668 and PKCS12_STACK_OF.
7671 *) When some versions of IIS use the 'NET' form of private key the
7672 key derivation algorithm is different. Normally MD5(password) is
7673 used as a 128 bit RC4 key. In the modified case
7674 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
7675 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
7676 as the old Netscape_RSA functions except they have an additional
7677 'sgckey' parameter which uses the modified algorithm. Also added
7678 an -sgckey command line option to the rsa utility. Thanks to
7679 Adrian Peck <bertie@ncipher.com> for posting details of the modified
7680 algorithm to openssl-dev.
7683 *) The evp_local.h macros were using 'c.##kname' which resulted in
7684 invalid expansion on some systems (SCO 5.0.5 for example).
7685 Corrected to 'c.kname'.
7686 [Phillip Porch <root@theporch.com>]
7688 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
7689 a STACK of email addresses from a certificate or request, these look
7690 in the subject name and the subject alternative name extensions and
7691 omit any duplicate addresses.
7694 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
7695 This makes DSA verification about 2 % faster.
7698 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
7699 (meaning that now 2^5 values will be precomputed, which is only 4 KB
7700 plus overhead for 1024 bit moduli).
7701 This makes exponentiations about 0.5 % faster for 1024 bit
7702 exponents (as measured by "openssl speed rsa2048").
7705 *) Rename memory handling macros to avoid conflicts with other
7707 Malloc => OPENSSL_malloc
7708 Malloc_locked => OPENSSL_malloc_locked
7709 Realloc => OPENSSL_realloc
7710 Free => OPENSSL_free
7713 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
7714 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
7717 *) CygWin32 support.
7718 [John Jarvie <jjarvie@newsguy.com>]
7720 *) The type-safe stack code has been rejigged. It is now only compiled
7721 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
7722 by default all type-specific stack functions are "#define"d back to
7723 standard stack functions. This results in more streamlined output
7724 but retains the type-safety checking possibilities of the original
7728 *) The STACK code has been cleaned up, and certain type declarations
7729 that didn't make a lot of sense have been brought in line. This has
7730 also involved a cleanup of sorts in safestack.h to more correctly
7731 map type-safe stack functions onto their plain stack counterparts.
7732 This work has also resulted in a variety of "const"ifications of
7733 lots of the code, especially "_cmp" operations which should normally
7734 be prototyped with "const" parameters anyway.
7737 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
7738 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
7739 (The PRNG state consists of two parts, the large pool 'state' and 'md',
7740 where all of 'md' is used each time the PRNG is used, but 'state'
7741 is used only indexed by a cyclic counter. As entropy may not be
7742 well distributed from the beginning, 'md' is important as a
7743 chaining variable. However, the output function chains only half
7744 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
7745 all of 'md', and seeding with STATE_SIZE dummy bytes will result
7746 in all of 'state' being rewritten, with the new values depending
7747 on virtually all of 'md'. This overcomes the 80 bit limitation.)
7750 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
7751 the handshake is continued after ssl_verify_cert_chain();
7752 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
7753 can lead to 'unexplainable' connection aborts later.
7754 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
7756 *) Major EVP API cipher revision.
7757 Add hooks for extra EVP features. This allows various cipher
7758 parameters to be set in the EVP interface. Support added for variable
7759 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
7760 setting of RC2 and RC5 parameters.
7762 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
7765 Remove lots of duplicated code from the EVP library. For example *every*
7766 cipher init() function handles the 'iv' in the same way according to the
7767 cipher mode. They also all do nothing if the 'key' parameter is NULL and
7768 for CFB and OFB modes they zero ctx->num.
7770 New functionality allows removal of S/MIME code RC2 hack.
7772 Most of the routines have the same form and so can be declared in terms
7775 By shifting this to the top level EVP_CipherInit() it can be removed from
7776 all individual ciphers. If the cipher wants to handle IVs or keys
7777 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7780 Change lots of functions like EVP_EncryptUpdate() to now return a
7781 value: although software versions of the algorithms cannot fail
7782 any installed hardware versions can.
7785 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7786 this option is set, tolerate broken clients that send the negotiated
7787 protocol version number instead of the requested protocol version
7791 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7792 i.e. non-zero for export ciphersuites, zero otherwise.
7793 Previous versions had this flag inverted, inconsistent with
7794 rsa_tmp_cb (..._TMP_RSA_CB).
7795 [Bodo Moeller; problem reported by Amit Chopra]
7797 *) Add missing DSA library text string. Work around for some IIS
7798 key files with invalid SEQUENCE encoding.
7801 *) Add a document (doc/standards.txt) that list all kinds of standards
7802 and so on that are implemented in OpenSSL.
7805 *) Enhance c_rehash script. Old version would mishandle certificates
7806 with the same subject name hash and wouldn't handle CRLs at all.
7807 Added -fingerprint option to crl utility, to support new c_rehash
7811 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7814 *) Fix for SSL server purpose checking. Server checking was
7815 rejecting certificates which had extended key usage present
7816 but no ssl client purpose.
7817 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7819 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7820 is a little unclear about how a blank password is handled.
7821 Since the password in encoded as a BMPString with terminating
7822 double NULL a zero length password would end up as just the
7823 double NULL. However no password at all is different and is
7824 handled differently in the PKCS#12 key generation code. NS
7825 treats a blank password as zero length. MSIE treats it as no
7826 password on export: but it will try both on import. We now do
7827 the same: PKCS12_parse() tries zero length and no password if
7828 the password is set to "" or NULL (NULL is now a valid password:
7829 it wasn't before) as does the pkcs12 application.
7832 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7833 perror when PEM_read_bio_X509_REQ fails, the error message must
7834 be obtained from the error queue.
7837 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7838 it in ERR_remove_state if appropriate, and change ERR_get_state
7839 accordingly to avoid race conditions (this is necessary because
7840 thread_hash is no longer constant once set).
7843 *) Bugfix for linux-elf makefile.one.
7846 *) RSA_get_default_method() will now cause a default
7847 RSA_METHOD to be chosen if one doesn't exist already.
7848 Previously this was only set during a call to RSA_new()
7849 or RSA_new_method(NULL) meaning it was possible for
7850 RSA_get_default_method() to return NULL.
7853 *) Added native name translation to the existing DSO code
7854 that will convert (if the flag to do so is set) filenames
7855 that are sufficiently small and have no path information
7856 into a canonical native form. Eg. "blah" converted to
7857 "libblah.so" or "blah.dll" etc.
7860 *) New function ERR_error_string_n(e, buf, len) which is like
7861 ERR_error_string(e, buf), but writes at most 'len' bytes
7862 including the 0 terminator. For ERR_error_string_n, 'buf'
7864 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7866 *) CONF library reworked to become more general. A new CONF
7867 configuration file reader "class" is implemented as well as a
7868 new functions (NCONF_*, for "New CONF") to handle it. The now
7869 old CONF_* functions are still there, but are reimplemented to
7870 work in terms of the new functions. Also, a set of functions
7871 to handle the internal storage of the configuration data is
7872 provided to make it easier to write new configuration file
7873 reader "classes" (I can definitely see something reading a
7874 configuration file in XML format, for example), called _CONF_*,
7875 or "the configuration storage API"...
7877 The new configuration file reading functions are:
7879 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7880 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7882 NCONF_default, NCONF_WIN32
7884 NCONF_dump_fp, NCONF_dump_bio
7886 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7887 NCONF_new creates a new CONF object. This works in the same way
7888 as other interfaces in OpenSSL, like the BIO interface.
7889 NCONF_dump_* dump the internal storage of the configuration file,
7890 which is useful for debugging. All other functions take the same
7891 arguments as the old CONF_* functions wth the exception of the
7892 first that must be a `CONF *' instead of a `LHASH *'.
7894 To make it easer to use the new classes with the old CONF_* functions,
7895 the function CONF_set_default_method is provided.
7898 *) Add '-tls1' option to 'openssl ciphers', which was already
7899 mentioned in the documentation but had not been implemented.
7900 (This option is not yet really useful because even the additional
7901 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7904 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7905 OpenSSL-based applications) load shared libraries and bind to
7906 them in a portable way.
7907 [Geoff Thorpe, with contributions from Richard Levitte]
7909 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7911 *) Make sure _lrotl and _lrotr are only used with MSVC.
7913 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7914 (the default implementation of RAND_status).
7916 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7917 to '-clrext' (= clear extensions), as intended and documented.
7918 [Bodo Moeller; inconsistency pointed out by Michael Attili
7921 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7922 was larger than the MD block size.
7923 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7925 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7926 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7927 using the passed key: if the passed key was a private key the result
7928 of X509_print(), for example, would be to print out all the private key
7932 *) des_quad_cksum() byte order bug fix.
7933 [Ulf Möller, using the problem description in krb4-0.9.7, where
7934 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7936 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7938 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7940 *) For easily testing in shell scripts whether some command
7941 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7942 returns with exit code 0 iff no command of the given name is available.
7943 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7944 the output goes to stdout and nothing is printed to stderr.
7945 Additional arguments are always ignored.
7947 Since for each cipher there is a command of the same name,
7948 the 'no-cipher' compilation switches can be tested this way.
7950 ('openssl no-XXX' is not able to detect pseudo-commands such
7951 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7954 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7957 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7958 is set; it will be thrown away anyway because each handshake creates
7960 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7961 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7962 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7963 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7966 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7967 'Q' and 'R' lose their special meanings (quit/renegotiate).
7968 This is part of what -quiet does; unlike -quiet, -ign_eof
7969 does not suppress any output.
7972 *) Add compatibility options to the purpose and trust code. The
7973 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7974 accepts a certificate or CA, this was the previous behaviour,
7975 with all the associated security issues.
7977 X509_TRUST_COMPAT is the old trust behaviour: only and
7978 automatically trust self signed roots in certificate store. A
7979 new trust setting X509_TRUST_DEFAULT is used to specify that
7980 a purpose has no associated trust setting and it should instead
7981 use the value in the default purpose.
7984 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7985 and fix a memory leak.
7988 *) In util/mkerr.pl (which implements 'make errors'), preserve
7989 reason strings from the previous version of the .c file, as
7990 the default to have only downcase letters (and digits) in
7991 automatically generated reasons codes is not always appropriate.
7994 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7995 using strerror. Previously, ERR_reason_error_string() returned
7996 library names as reason strings for SYSerr; but SYSerr is a special
7997 case where small numbers are errno values, not library numbers.
8000 *) Add '-dsaparam' option to 'openssl dhparam' application. This
8001 converts DSA parameters into DH parameters. (When creating parameters,
8002 DSA_generate_parameters is used.)
8005 *) Include 'length' (recommended exponent length) in C code generated
8006 by 'openssl dhparam -C'.
8009 *) The second argument to set_label in perlasm was already being used
8010 so couldn't be used as a "file scope" flag. Moved to third argument
8014 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8015 instead of RAND_bytes for encryption IVs and salts.
8018 *) Include RAND_status() into RAND_METHOD instead of implementing
8019 it only for md_rand.c Otherwise replacing the PRNG by calling
8020 RAND_set_rand_method would be impossible.
8023 *) Don't let DSA_generate_key() enter an infinite loop if the random
8024 number generation fails.
8027 *) New 'rand' application for creating pseudo-random output.
8030 *) Added configuration support for Linux/IA64
8031 [Rolf Haberrecker <rolf@suse.de>]
8033 *) Assembler module support for Mingw32.
8036 *) Shared library support for HPUX (in shlib/).
8037 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8039 *) Shared library support for Solaris gcc.
8040 [Lutz Behnke <behnke@trustcenter.de>]
8042 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8044 *) PKCS7_encrypt() was adding text MIME headers twice because they
8045 were added manually and by SMIME_crlf_copy().
8048 *) In bntest.c don't call BN_rand with zero bits argument.
8049 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8051 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8052 case was implemented. This caused BN_div_recp() to fail occasionally.
8055 *) Add an optional second argument to the set_label() in the perl
8056 assembly language builder. If this argument exists and is set
8057 to 1 it signals that the assembler should use a symbol whose
8058 scope is the entire file, not just the current function. This
8059 is needed with MASM which uses the format label:: for this scope.
8060 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8062 *) Change the ASN1 types so they are typedefs by default. Before
8063 almost all types were #define'd to ASN1_STRING which was causing
8064 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8068 *) Change names of new functions to the new get1/get0 naming
8069 convention: After 'get1', the caller owns a reference count
8070 and has to call ..._free; 'get0' returns a pointer to some
8071 data structure without incrementing reference counters.
8072 (Some of the existing 'get' functions increment a reference
8073 counter, some don't.)
8074 Similarly, 'set1' and 'add1' functions increase reference
8075 counters or duplicate objects.
8078 *) Allow for the possibility of temp RSA key generation failure:
8079 the code used to assume it always worked and crashed on failure.
8082 *) Fix potential buffer overrun problem in BIO_printf().
8083 [Ulf Möller, using public domain code by Patrick Powell; problem
8084 pointed out by David Sacerdote <das33@cornell.edu>]
8086 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8087 RAND_egd() and RAND_status(). In the command line application,
8088 the EGD socket can be specified like a seed file using RANDFILE
8092 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8093 Some CAs (e.g. Verisign) distribute certificates in this form.
8096 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8097 list to exclude them. This means that no special compilation option
8098 is needed to use anonymous DH: it just needs to be included in the
8102 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8103 EVP_MD_type. The old functionality is available in a new macro called
8104 EVP_MD_md(). Change code that uses it and update docs.
8107 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8108 where the 'void *' argument is replaced by a function pointer argument.
8109 Previously 'void *' was abused to point to functions, which works on
8110 many platforms, but is not correct. As these functions are usually
8111 called by macros defined in OpenSSL header files, most source code
8112 should work without changes.
8115 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8116 sections with information on -D... compiler switches used for
8117 compiling the library so that applications can see them. To enable
8118 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8119 must be defined. E.g.,
8120 #define OPENSSL_ALGORITHM_DEFINES
8121 #include <openssl/opensslconf.h>
8122 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8123 [Richard Levitte, Ulf and Bodo Möller]
8125 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8129 *) Change the 'other' type in certificate aux info to a STACK_OF
8130 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8131 the required ASN1 format: arbitrary types determined by an OID.
8134 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8135 argument to 'req'. This is not because the function is newer or
8136 better than others it just uses the work 'NEW' in the certificate
8137 request header lines. Some software needs this.
8140 *) Reorganise password command line arguments: now passwords can be
8141 obtained from various sources. Delete the PEM_cb function and make
8142 it the default behaviour: i.e. if the callback is NULL and the
8143 usrdata argument is not NULL interpret it as a null terminated pass
8144 phrase. If usrdata and the callback are NULL then the pass phrase
8145 is prompted for as usual.
8148 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8149 the support is automatically enabled. The resulting binaries will
8150 autodetect the card and use it if present.
8151 [Ben Laurie and Compaq Inc.]
8153 *) Work around for Netscape hang bug. This sends certificate request
8154 and server done in one record. Since this is perfectly legal in the
8155 SSL/TLS protocol it isn't a "bug" option and is on by default. See
8156 the bugs/SSLv3 entry for more info.
8159 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
8162 *) Add -rand argument to smime and pkcs12 applications and read/write
8166 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
8169 *) Add command line password options to the remaining applications.
8172 *) Bug fix for BN_div_recp() for numerators with an even number of
8176 *) More tests in bntest.c, and changed test_bn output.
8179 *) ./config recognizes MacOS X now.
8182 *) Bug fix for BN_div() when the first words of num and divsor are
8183 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
8186 *) Add support for various broken PKCS#8 formats, and command line
8187 options to produce them.
8190 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
8191 get temporary BIGNUMs from a BN_CTX.
8194 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
8198 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
8199 include a #define from the old name to the new. The original intent
8200 was that statically linked binaries could for example just call
8201 SSLeay_add_all_ciphers() to just add ciphers to the table and not
8202 link with digests. This never worked becayse SSLeay_add_all_digests()
8203 and SSLeay_add_all_ciphers() were in the same source file so calling
8204 one would link with the other. They are now in separate source files.
8207 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
8210 *) Use a less unusual form of the Miller-Rabin primality test (it used
8211 a binary algorithm for exponentiation integrated into the Miller-Rabin
8212 loop, our standard modexp algorithms are faster).
8215 *) Support for the EBCDIC character set completed.
8216 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
8218 *) Source code cleanups: use const where appropriate, eliminate casts,
8219 use void * instead of char * in lhash.
8222 *) Bugfix: ssl3_send_server_key_exchange was not restartable
8223 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
8224 this the server could overwrite ephemeral keys that the client
8228 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
8229 using 50 iterations of the Rabin-Miller test.
8231 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
8232 iterations of the Rabin-Miller test as required by the appendix
8233 to FIPS PUB 186[-1]) instead of DSA_is_prime.
8234 As BN_is_prime_fasttest includes trial division, DSA parameter
8235 generation becomes much faster.
8237 This implies a change for the callback functions in DSA_is_prime
8238 and DSA_generate_parameters: The callback function is called once
8239 for each positive witness in the Rabin-Miller test, not just
8240 occasionally in the inner loop; and the parameters to the
8241 callback function now provide an iteration count for the outer
8242 loop rather than for the current invocation of the inner loop.
8243 DSA_generate_parameters additionally can call the callback
8244 function with an 'iteration count' of -1, meaning that a
8245 candidate has passed the trial division test (when q is generated
8246 from an application-provided seed, trial division is skipped).
8249 *) New function BN_is_prime_fasttest that optionally does trial
8250 division before starting the Rabin-Miller test and has
8251 an additional BN_CTX * argument (whereas BN_is_prime always
8252 has to allocate at least one BN_CTX).
8253 'callback(1, -1, cb_arg)' is called when a number has passed the
8254 trial division stage.
8257 *) Fix for bug in CRL encoding. The validity dates weren't being handled
8261 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
8264 *) New function BN_pseudo_rand().
8267 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
8268 bignum version of BN_from_montgomery() with the working code from
8269 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
8273 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
8274 made it impossible to use the same SSL_SESSION data structure in
8275 SSL2 clients in multiple threads.
8278 *) The return value of RAND_load_file() no longer counts bytes obtained
8279 by stat(). RAND_load_file(..., -1) is new and uses the complete file
8280 to seed the PRNG (previously an explicit byte count was required).
8281 [Ulf Möller, Bodo Möller]
8283 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
8284 used (char *) instead of (void *) and had casts all over the place.
8287 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
8290 *) Retain source code compatibility for BN_prime_checks macro:
8291 BN_is_prime(..., BN_prime_checks, ...) now uses
8292 BN_prime_checks_for_size to determine the appropriate number of
8293 Rabin-Miller iterations.
8296 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
8297 DH_CHECK_P_NOT_SAFE_PRIME.
8298 (Check if this is true? OpenPGP calls them "strong".)
8301 *) Merge the functionality of "dh" and "gendh" programs into a new program
8302 "dhparam". The old programs are retained for now but will handle DH keys
8303 (instead of parameters) in future.
8306 *) Make the ciphers, s_server and s_client programs check the return values
8307 when a new cipher list is set.
8310 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
8311 ciphers. Before when the 56bit ciphers were enabled the sorting was
8314 The syntax for the cipher sorting has been extended to support sorting by
8315 cipher-strength (using the strength_bits hard coded in the tables).
8316 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
8318 Fix a bug in the cipher-command parser: when supplying a cipher command
8319 string with an "undefined" symbol (neither command nor alphanumeric
8320 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
8321 an error is flagged.
8323 Due to the strength-sorting extension, the code of the
8324 ssl_create_cipher_list() function was completely rearranged. I hope that
8325 the readability was also increased :-)
8326 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8328 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
8329 for the first serial number and places 2 in the serial number file. This
8330 avoids problems when the root CA is created with serial number zero and
8331 the first user certificate has the same issuer name and serial number
8335 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
8336 the new code. Add documentation for this stuff.
8339 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
8340 X509_*() to X509at_*() on the grounds that they don't handle X509
8341 structures and behave in an analagous way to the X509v3 functions:
8342 they shouldn't be called directly but wrapper functions should be used
8345 So we also now have some wrapper functions that call the X509at functions
8346 when passed certificate requests. (TO DO: similar things can be done with
8347 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
8348 things. Some of these need some d2i or i2d and print functionality
8349 because they handle more complex structures.)
8352 *) Add missing #ifndefs that caused missing symbols when building libssl
8353 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
8354 NO_RSA in ssl/s2*.c.
8355 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
8357 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
8358 has a return value which indicates the quality of the random data
8359 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
8360 error queue. New function RAND_pseudo_bytes() generates output that is
8361 guaranteed to be unique but not unpredictable. RAND_add is like
8362 RAND_seed, but takes an extra argument for an entropy estimate
8363 (RAND_seed always assumes full entropy).
8366 *) Do more iterations of Rabin-Miller probable prime test (specifically,
8367 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
8368 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
8369 in crypto/bn/bn_prime.c for the complete table). This guarantees a
8370 false-positive rate of at most 2^-80 for random input.
8373 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
8376 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
8377 in the 0.9.5 release), this returns the chain
8378 from an X509_CTX structure with a dup of the stack and all
8379 the X509 reference counts upped: so the stack will exist
8380 after X509_CTX_cleanup() has been called. Modify pkcs12.c
8383 Also make SSL_SESSION_print() print out the verify return
8387 *) Add manpage for the pkcs12 command. Also change the default
8388 behaviour so MAC iteration counts are used unless the new
8389 -nomaciter option is used. This improves file security and
8390 only older versions of MSIE (4.0 for example) need it.
8393 *) Honor the no-xxx Configure options when creating .DEF files.
8396 *) Add PKCS#10 attributes to field table: challengePassword,
8397 unstructuredName and unstructuredAddress. These are taken from
8398 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
8399 international characters are used.
8401 More changes to X509_ATTRIBUTE code: allow the setting of types
8402 based on strings. Remove the 'loc' parameter when adding
8403 attributes because these will be a SET OF encoding which is sorted
8407 *) Initial changes to the 'req' utility to allow request generation
8408 automation. This will allow an application to just generate a template
8409 file containing all the field values and have req construct the
8412 Initial support for X509_ATTRIBUTE handling. Stacks of these are
8413 used all over the place including certificate requests and PKCS#7
8414 structures. They are currently handled manually where necessary with
8415 some primitive wrappers for PKCS#7. The new functions behave in a
8416 manner analogous to the X509 extension functions: they allow
8417 attributes to be looked up by NID and added.
8419 Later something similar to the X509V3 code would be desirable to
8420 automatically handle the encoding, decoding and printing of the
8421 more complex types. The string types like challengePassword can
8422 be handled by the string table functions.
8424 Also modified the multi byte string table handling. Now there is
8425 a 'global mask' which masks out certain types. The table itself
8426 can use the flag STABLE_NO_MASK to ignore the mask setting: this
8427 is useful when for example there is only one permissible type
8428 (as in countryName) and using the mask might result in no valid
8432 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
8433 SSL_get_peer_finished to allow applications to obtain the latest
8434 Finished messages sent to the peer or expected from the peer,
8435 respectively. (SSL_get_peer_finished is usually the Finished message
8436 actually received from the peer, otherwise the protocol will be aborted.)
8438 As the Finished message are message digests of the complete handshake
8439 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
8440 be used for external authentication procedures when the authentication
8441 provided by SSL/TLS is not desired or is not enough.
8444 *) Enhanced support for Alpha Linux is added. Now ./config checks if
8445 the host supports BWX extension and if Compaq C is present on the
8446 $PATH. Just exploiting of the BWX extension results in 20-30%
8447 performance kick for some algorithms, e.g. DES and RC4 to mention
8448 a couple. Compaq C in turn generates ~20% faster code for MD5 and
8452 *) Add support for MS "fast SGC". This is arguably a violation of the
8453 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
8454 weak crypto and after checking the certificate is SGC a second one
8455 with strong crypto. MS SGC stops the first handshake after receiving
8456 the server certificate message and sends a second client hello. Since
8457 a server will typically do all the time consuming operations before
8458 expecting any further messages from the client (server key exchange
8459 is the most expensive) there is little difference between the two.
8461 To get OpenSSL to support MS SGC we have to permit a second client
8462 hello message after we have sent server done. In addition we have to
8463 reset the MAC if we do get this second client hello.
8466 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
8467 if a DER encoded private key is RSA or DSA traditional format. Changed
8468 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
8469 format DER encoded private key. Newer code should use PKCS#8 format which
8470 has the key type encoded in the ASN1 structure. Added DER private key
8471 support to pkcs8 application.
8474 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8475 ciphersuites has been selected (as required by the SSL 3/TLS 1
8476 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8477 is set, we interpret this as a request to violate the specification
8478 (the worst that can happen is a handshake failure, and 'correct'
8479 behaviour would result in a handshake failure anyway).
8482 *) In SSL_CTX_add_session, take into account that there might be multiple
8483 SSL_SESSION structures with the same session ID (e.g. when two threads
8484 concurrently obtain them from an external cache).
8485 The internal cache can handle only one SSL_SESSION with a given ID,
8486 so if there's a conflict, we now throw out the old one to achieve
8490 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8491 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8492 some routines that use cipher OIDs: some ciphers do not have OIDs
8493 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8497 *) Simplify the trust setting structure and code. Now we just have
8498 two sequences of OIDs for trusted and rejected settings. These will
8499 typically have values the same as the extended key usage extension
8500 and any application specific purposes.
8502 The trust checking code now has a default behaviour: it will just
8503 check for an object with the same NID as the passed id. Functions can
8504 be provided to override either the default behaviour or the behaviour
8505 for a given id. SSL client, server and email already have functions
8506 in place for compatibility: they check the NID and also return "trusted"
8507 if the certificate is self signed.
8510 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8511 traditional format into an EVP_PKEY structure.
8514 *) Add a password callback function PEM_cb() which either prompts for
8515 a password if usr_data is NULL or otherwise assumes it is a null
8516 terminated password. Allow passwords to be passed on command line
8517 environment or config files in a few more utilities.
8520 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8521 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8522 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8523 Update documentation.
8526 *) Support for ASN1 "NULL" type. This could be handled before by using
8527 ASN1_TYPE but there wasn't any function that would try to read a NULL
8528 and produce an error if it couldn't. For compatibility we also have
8529 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8530 don't allocate anything because they don't need to.
8533 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8535 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8537 *) Rebuild of the memory allocation routines used by OpenSSL code and
8538 possibly others as well. The purpose is to make an interface that
8539 provide hooks so anyone can build a separate set of allocation and
8540 deallocation routines to be used by OpenSSL, for example memory
8541 pool implementations, or something else, which was previously hard
8542 since Malloc(), Realloc() and Free() were defined as macros having
8543 the values malloc, realloc and free, respectively (except for Win32
8544 compilations). The same is provided for memory debugging code.
8545 OpenSSL already comes with functionality to find memory leaks, but
8546 this gives people a chance to debug other memory problems.
8548 With these changes, a new set of functions and macros have appeared:
8550 CRYPTO_set_mem_debug_functions() [F]
8551 CRYPTO_get_mem_debug_functions() [F]
8552 CRYPTO_dbg_set_options() [F]
8553 CRYPTO_dbg_get_options() [F]
8554 CRYPTO_malloc_debug_init() [M]
8556 The memory debug functions are NULL by default, unless the library
8557 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8558 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8559 gives the standard debugging functions that come with OpenSSL) or
8560 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8561 provided by the library user) must be used. When the standard
8562 debugging functions are used, CRYPTO_dbg_set_options can be used to
8563 request additional information:
8564 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8565 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8567 Also, things like CRYPTO_set_mem_functions will always give the
8568 expected result (the new set of functions is used for allocation
8569 and deallocation) at all times, regardless of platform and compiler
8572 To finish it up, some functions that were never use in any other
8573 way than through macros have a new API and new semantic:
8576 CRYPTO_dbg_realloc()
8579 All macros of value have retained their old syntax.
8580 [Richard Levitte and Bodo Moeller]
8582 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8583 ordering of SMIMECapabilities wasn't in "strength order" and there
8584 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8588 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8589 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8590 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
8592 *) Merge in my S/MIME library for OpenSSL. This provides a simple
8593 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
8594 functionality to handle multipart/signed properly) and a utility
8595 called 'smime' to call all this stuff. This is based on code I
8596 originally wrote for Celo who have kindly allowed it to be
8597 included in OpenSSL.
8600 *) Add variants des_set_key_checked and des_set_key_unchecked of
8601 des_set_key (aka des_key_sched). Global variable des_check_key
8602 decides which of these is called by des_set_key; this way
8603 des_check_key behaves as it always did, but applications and
8604 the library itself, which was buggy for des_check_key == 1,
8605 have a cleaner way to pick the version they need.
8608 *) New function PKCS12_newpass() which changes the password of a
8612 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
8613 dynamic mix. In both cases the ids can be used as an index into the
8614 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
8615 functions so they accept a list of the field values and the
8616 application doesn't need to directly manipulate the X509_TRUST
8620 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
8624 *) Modify the way the V3 extension code looks up extensions. This now
8625 works in a similar way to the object code: we have some "standard"
8626 extensions in a static table which is searched with OBJ_bsearch()
8627 and the application can add dynamic ones if needed. The file
8628 crypto/x509v3/ext_dat.h now has the info: this file needs to be
8629 updated whenever a new extension is added to the core code and kept
8630 in ext_nid order. There is a simple program 'tabtest.c' which checks
8631 this. New extensions are not added too often so this file can readily
8632 be maintained manually.
8634 There are two big advantages in doing things this way. The extensions
8635 can be looked up immediately and no longer need to be "added" using
8636 X509V3_add_standard_extensions(): this function now does nothing.
8637 [Side note: I get *lots* of email saying the extension code doesn't
8638 work because people forget to call this function]
8639 Also no dynamic allocation is done unless new extensions are added:
8640 so if we don't add custom extensions there is no need to call
8641 X509V3_EXT_cleanup().
8644 *) Modify enc utility's salting as follows: make salting the default. Add a
8645 magic header, so unsalted files fail gracefully instead of just decrypting
8646 to garbage. This is because not salting is a big security hole, so people
8647 should be discouraged from doing it.
8650 *) Fixes and enhancements to the 'x509' utility. It allowed a message
8651 digest to be passed on the command line but it only used this
8652 parameter when signing a certificate. Modified so all relevant
8653 operations are affected by the digest parameter including the
8654 -fingerprint and -x509toreq options. Also -x509toreq choked if a
8655 DSA key was used because it didn't fix the digest.
8658 *) Initial certificate chain verify code. Currently tests the untrusted
8659 certificates for consistency with the verify purpose (which is set
8660 when the X509_STORE_CTX structure is set up) and checks the pathlength.
8662 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
8663 this is because it will reject chains with invalid extensions whereas
8664 every previous version of OpenSSL and SSLeay made no checks at all.
8666 Trust code: checks the root CA for the relevant trust settings. Trust
8667 settings have an initial value consistent with the verify purpose: e.g.
8668 if the verify purpose is for SSL client use it expects the CA to be
8669 trusted for SSL client use. However the default value can be changed to
8670 permit custom trust settings: one example of this would be to only trust
8671 certificates from a specific "secure" set of CAs.
8673 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
8674 which should be used for version portability: especially since the
8675 verify structure is likely to change more often now.
8677 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
8678 to set them. If not set then assume SSL clients will verify SSL servers
8681 Two new options to the verify program: -untrusted allows a set of
8682 untrusted certificates to be passed in and -purpose which sets the
8683 intended purpose of the certificate. If a purpose is set then the
8684 new chain verify code is used to check extension consistency.
8687 *) Support for the authority information access extension.
8690 *) Modify RSA and DSA PEM read routines to transparently handle
8691 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
8692 public keys in a format compatible with certificate
8693 SubjectPublicKeyInfo structures. Unfortunately there were already
8694 functions called *_PublicKey_* which used various odd formats so
8695 these are retained for compatibility: however the DSA variants were
8696 never in a public release so they have been deleted. Changed dsa/rsa
8697 utilities to handle the new format: note no releases ever handled public
8698 keys so we should be OK.
8700 The primary motivation for this change is to avoid the same fiasco
8701 that dogs private keys: there are several incompatible private key
8702 formats some of which are standard and some OpenSSL specific and
8703 require various evil hacks to allow partial transparent handling and
8704 even then it doesn't work with DER formats. Given the option anything
8705 other than PKCS#8 should be dumped: but the other formats have to
8706 stay in the name of compatibility.
8708 With public keys and the benefit of hindsight one standard format
8709 is used which works with EVP_PKEY, RSA or DSA structures: though
8710 it clearly returns an error if you try to read the wrong kind of key.
8712 Added a -pubkey option to the 'x509' utility to output the public key.
8713 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
8714 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
8715 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
8716 that do the same as the EVP_PKEY_assign_*() except they up the
8717 reference count of the added key (they don't "swallow" the
8721 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
8722 CRLs would fail if the file contained no certificates or no CRLs:
8723 added a new function to read in both types and return the number
8724 read: this means that if none are read it will be an error. The
8725 DER versions of the certificate and CRL reader would always fail
8726 because it isn't possible to mix certificates and CRLs in DER format
8727 without choking one or the other routine. Changed this to just read
8728 a certificate: this is the best we can do. Also modified the code
8729 in apps/verify.c to take notice of return codes: it was previously
8730 attempting to read in certificates from NULL pointers and ignoring
8731 any errors: this is one reason why the cert and CRL reader seemed
8732 to work. It doesn't check return codes from the default certificate
8733 routines: these may well fail if the certificates aren't installed.
8736 *) Code to support otherName option in GeneralName.
8739 *) First update to verify code. Change the verify utility
8740 so it warns if it is passed a self signed certificate:
8741 for consistency with the normal behaviour. X509_verify
8742 has been modified to it will now verify a self signed
8743 certificate if *exactly* the same certificate appears
8744 in the store: it was previously impossible to trust a
8745 single self signed certificate. This means that:
8746 openssl verify ss.pem
8747 now gives a warning about a self signed certificate but
8748 openssl verify -CAfile ss.pem ss.pem
8752 *) For servers, store verify_result in SSL_SESSION data structure
8753 (and add it to external session representation).
8754 This is needed when client certificate verifications fails,
8755 but an application-provided verification callback (set by
8756 SSL_CTX_set_cert_verify_callback) allows accepting the session
8757 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
8758 but returns 1): When the session is reused, we have to set
8759 ssl->verify_result to the appropriate error code to avoid
8761 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
8763 *) Fix a bug in the new PKCS#7 code: it didn't consider the
8764 case in PKCS7_dataInit() where the signed PKCS7 structure
8765 didn't contain any existing data because it was being created.
8766 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
8768 *) Add a salt to the key derivation routines in enc.c. This
8769 forms the first 8 bytes of the encrypted file. Also add a
8770 -S option to allow a salt to be input on the command line.
8773 *) New function X509_cmp(). Oddly enough there wasn't a function
8774 to compare two certificates. We do this by working out the SHA1
8775 hash and comparing that. X509_cmp() will be needed by the trust
8779 *) SSL_get1_session() is like SSL_get_session(), but increments
8780 the reference count in the SSL_SESSION returned.
8781 [Geoff Thorpe <geoff@eu.c2.net>]
8783 *) Fix for 'req': it was adding a null to request attributes.
8784 Also change the X509_LOOKUP and X509_INFO code to handle
8785 certificate auxiliary information.
8788 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8792 *) Add the possibility to add extra information to the memory leak
8793 detecting output, to form tracebacks, showing from where each
8794 allocation was originated: CRYPTO_push_info("constant string") adds
8795 the string plus current file name and line number to a per-thread
8796 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8797 is like calling CYRPTO_pop_info() until the stack is empty.
8798 Also updated memory leak detection code to be multi-thread-safe.
8801 *) Add options -text and -noout to pkcs7 utility and delete the
8802 encryption options which never did anything. Update docs.
8805 *) Add options to some of the utilities to allow the pass phrase
8806 to be included on either the command line (not recommended on
8807 OSes like Unix) or read from the environment. Update the
8808 manpages and fix a few bugs.
8811 *) Add a few manpages for some of the openssl commands.
8814 *) Fix the -revoke option in ca. It was freeing up memory twice,
8815 leaking and not finding already revoked certificates.
8818 *) Extensive changes to support certificate auxiliary information.
8819 This involves the use of X509_CERT_AUX structure and X509_AUX
8820 functions. An X509_AUX function such as PEM_read_X509_AUX()
8821 can still read in a certificate file in the usual way but it
8822 will also read in any additional "auxiliary information". By
8823 doing things this way a fair degree of compatibility can be
8824 retained: existing certificates can have this information added
8825 using the new 'x509' options.
8827 Current auxiliary information includes an "alias" and some trust
8828 settings. The trust settings will ultimately be used in enhanced
8829 certificate chain verification routines: currently a certificate
8830 can only be trusted if it is self signed and then it is trusted
8834 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8835 The problem was that one of the replacement routines had not been working
8836 since SSLeay releases. For now the offending routine has been replaced
8837 with non-optimised assembler. Even so, this now gives around 95%
8838 performance improvement for 1024 bit RSA signs.
8841 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8842 handling. Most clients have the effective key size in bits equal to
8843 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8844 A few however don't do this and instead use the size of the decrypted key
8845 to determine the RC2 key length and the AlgorithmIdentifier to determine
8846 the effective key length. In this case the effective key length can still
8847 be 40 bits but the key length can be 168 bits for example. This is fixed
8848 by manually forcing an RC2 key into the EVP_PKEY structure because the
8849 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8850 the key length and effective key length are equal.
8853 *) Add a bunch of functions that should simplify the creation of
8854 X509_NAME structures. Now you should be able to do:
8855 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8856 and have it automatically work out the correct field type and fill in
8857 the structures. The more adventurous can try:
8858 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8859 and it will (hopefully) work out the correct multibyte encoding.
8862 *) Change the 'req' utility to use the new field handling and multibyte
8863 copy routines. Before the DN field creation was handled in an ad hoc
8864 way in req, ca, and x509 which was rather broken and didn't support
8865 BMPStrings or UTF8Strings. Since some software doesn't implement
8866 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8867 using the dirstring_type option. See the new comment in the default
8868 openssl.cnf for more info.
8871 *) Make crypto/rand/md_rand.c more robust:
8872 - Assure unique random numbers after fork().
8873 - Make sure that concurrent threads access the global counter and
8874 md serializably so that we never lose entropy in them
8875 or use exactly the same state in multiple threads.
8876 Access to the large state is not always serializable because
8877 the additional locking could be a performance killer, and
8878 md should be large enough anyway.
8881 *) New file apps/app_rand.c with commonly needed functionality
8882 for handling the random seed file.
8884 Use the random seed file in some applications that previously did not:
8886 dsaparam -genkey (which also ignored its '-rand' option),
8889 x509 (when signing).
8890 Except on systems with /dev/urandom, it is crucial to have a random
8891 seed file at least for key creation, DSA signing, and for DH exchanges;
8892 for RSA signatures we could do without one.
8894 gendh and gendsa (unlike genrsa) used to read only the first byte
8895 of each file listed in the '-rand' option. The function as previously
8896 found in genrsa is now in app_rand.c and is used by all programs
8897 that support '-rand'.
8900 *) In RAND_write_file, use mode 0600 for creating files;
8901 don't just chmod when it may be too late.
8904 *) Report an error from X509_STORE_load_locations
8905 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8908 *) New function ASN1_mbstring_copy() this copies a string in either
8909 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8910 into an ASN1_STRING type. A mask of permissible types is passed
8911 and it chooses the "minimal" type to use or an error if not type
8915 *) Add function equivalents to the various macros in asn1.h. The old
8916 macros are retained with an M_ prefix. Code inside the library can
8917 use the M_ macros. External code (including the openssl utility)
8918 should *NOT* in order to be "shared library friendly".
8921 *) Add various functions that can check a certificate's extensions
8922 to see if it usable for various purposes such as SSL client,
8923 server or S/MIME and CAs of these types. This is currently
8924 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8925 verification. Also added a -purpose flag to x509 utility to
8926 print out all the purposes.
8929 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8933 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8934 for, obtain and decode and extension and obtain its critical flag.
8935 This allows all the necessary extension code to be handled in a
8936 single function call.
8939 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8940 platforms. See crypto/rc4/rc4_enc.c for further details.
8943 *) New -noout option to asn1parse. This causes no output to be produced
8944 its main use is when combined with -strparse and -out to extract data
8945 from a file (which may not be in ASN.1 format).
8948 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8949 when producing the local key id.
8950 [Richard Levitte <levitte@stacken.kth.se>]
8952 *) New option -dhparam in s_server. This allows a DH parameter file to be
8953 stated explicitly. If it is not stated then it tries the first server
8954 certificate file. The previous behaviour hard coded the filename
8958 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8959 a public key to be input or output. For example:
8960 openssl rsa -in key.pem -pubout -out pubkey.pem
8961 Also added necessary DSA public key functions to handle this.
8964 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8965 in the message. This was handled by allowing
8966 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8967 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8969 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8970 to the end of the strings whereas this didn't. This would cause problems
8971 if strings read with d2i_ASN1_bytes() were later modified.
8972 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8974 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8975 data and it contains EOF it will end up returning an error. This is
8976 caused by input 46 bytes long. The cause is due to the way base64
8977 BIOs find the start of base64 encoded data. They do this by trying a
8978 trial decode on each line until they find one that works. When they
8979 do a flag is set and it starts again knowing it can pass all the
8980 data directly through the decoder. Unfortunately it doesn't reset
8981 the context it uses. This means that if EOF is reached an attempt
8982 is made to pass two EOFs through the context and this causes the
8983 resulting error. This can also cause other problems as well. As is
8984 usual with these problems it takes *ages* to find and the fix is
8985 trivial: move one line.
8986 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8988 *) Ugly workaround to get s_client and s_server working under Windows. The
8989 old code wouldn't work because it needed to select() on sockets and the
8990 tty (for keypresses and to see if data could be written). Win32 only
8991 supports select() on sockets so we select() with a 1s timeout on the
8992 sockets and then see if any characters are waiting to be read, if none
8993 are present then we retry, we also assume we can always write data to
8994 the tty. This isn't nice because the code then blocks until we've
8995 received a complete line of data and it is effectively polling the
8996 keyboard at 1s intervals: however it's quite a bit better than not
8997 working at all :-) A dedicated Windows application might handle this
8998 with an event loop for example.
9001 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9002 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9003 will be called when RSA_sign() and RSA_verify() are used. This is useful
9004 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9005 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9006 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9007 This necessitated the support of an extra signature type NID_md5_sha1
9008 for SSL signatures and modifications to the SSL library to use it instead
9009 of calling RSA_public_decrypt() and RSA_private_encrypt().
9012 *) Add new -verify -CAfile and -CApath options to the crl program, these
9013 will lookup a CRL issuers certificate and verify the signature in a
9014 similar way to the verify program. Tidy up the crl program so it
9015 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9016 less strict. It will now permit CRL extensions even if it is not
9017 a V2 CRL: this will allow it to tolerate some broken CRLs.
9020 *) Initialize all non-automatic variables each time one of the openssl
9021 sub-programs is started (this is necessary as they may be started
9022 multiple times from the "OpenSSL>" prompt).
9023 [Lennart Bang, Bodo Moeller]
9025 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9026 removing all other RSA functionality (this is what NO_RSA does). This
9027 is so (for example) those in the US can disable those operations covered
9028 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9032 *) Non-copying interface to BIO pairs.
9033 (still largely untested)
9036 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9037 ASCII string. This was handled independently in various places before.
9040 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9041 UTF8 strings a character at a time.
9044 *) Use client_version from client hello to select the protocol
9045 (s23_srvr.c) and for RSA client key exchange verification
9046 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9049 *) Add various utility functions to handle SPKACs, these were previously
9050 handled by poking round in the structure internals. Added new function
9051 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9052 print, verify and generate SPKACs. Based on an original idea from
9053 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9056 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9059 *) Allow the config file extension section to be overwritten on the
9060 command line. Based on an original idea from Massimiliano Pala
9061 <madwolf@comune.modena.it>. The new option is called -extensions
9062 and can be applied to ca, req and x509. Also -reqexts to override
9063 the request extensions in req and -crlexts to override the crl extensions
9067 *) Add new feature to the SPKAC handling in ca. Now you can include
9068 the same field multiple times by preceding it by "XXXX." for example:
9071 this is the same syntax as used in the req config file.
9074 *) Allow certificate extensions to be added to certificate requests. These
9075 are specified in a 'req_extensions' option of the req section of the
9076 config file. They can be printed out with the -text option to req but
9077 are otherwise ignored at present.
9080 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9081 data read consists of only the final block it would not decrypted because
9082 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9083 A misplaced 'break' also meant the decrypted final block might not be
9084 copied until the next read.
9087 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9088 a few extra parameters to the DH structure: these will be useful if
9089 for example we want the value of 'q' or implement X9.42 DH.
9092 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9093 provides hooks that allow the default DSA functions or functions on a
9094 "per key" basis to be replaced. This allows hardware acceleration and
9095 hardware key storage to be handled without major modification to the
9096 library. Also added low level modexp hooks and CRYPTO_EX structure and
9097 associated functions.
9100 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9101 as "read only": it can't be written to and the buffer it points to will
9102 not be freed. Reading from a read only BIO is much more efficient than
9103 a normal memory BIO. This was added because there are several times when
9104 an area of memory needs to be read from a BIO. The previous method was
9105 to create a memory BIO and write the data to it, this results in two
9106 copies of the data and an O(n^2) reading algorithm. There is a new
9107 function BIO_new_mem_buf() which creates a read only memory BIO from
9108 an area of memory. Also modified the PKCS#7 routines to use read only
9112 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9113 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9114 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9115 but a retry condition occured while trying to read the rest.
9118 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9119 NID_pkcs7_encrypted by default: this was wrong since this should almost
9120 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9121 the encrypted data type: this is a more sensible place to put it and it
9122 allows the PKCS#12 code to be tidied up that duplicated this
9126 *) Changed obj_dat.pl script so it takes its input and output files on
9127 the command line. This should avoid shell escape redirection problems
9131 *) Initial support for certificate extension requests, these are included
9132 in things like Xenroll certificate requests. Included functions to allow
9133 extensions to be obtained and added.
9136 *) -crlf option to s_client and s_server for sending newlines as
9137 CRLF (as required by many protocols).
9140 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9142 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9143 [Ralf S. Engelschall]
9145 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9146 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9148 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
9152 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
9153 DH parameters/keys (q is lost during that conversion, but the resulting
9154 DH parameters contain its length).
9156 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
9157 much faster than DH_generate_parameters (which creates parameters
9158 where p = 2*q + 1), and also the smaller q makes DH computations
9159 much more efficient (160-bit exponentiation instead of 1024-bit
9160 exponentiation); so this provides a convenient way to support DHE
9161 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
9162 utter importance to use
9163 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9165 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9166 when such DH parameters are used, because otherwise small subgroup
9167 attacks may become possible!
9170 *) Avoid memory leak in i2d_DHparams.
9173 *) Allow the -k option to be used more than once in the enc program:
9174 this allows the same encrypted message to be read by multiple recipients.
9177 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
9178 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
9179 it will always use the numerical form of the OID, even if it has a short
9183 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
9184 method only got called if p,q,dmp1,dmq1,iqmp components were present,
9185 otherwise bn_mod_exp was called. In the case of hardware keys for example
9186 no private key components need be present and it might store extra data
9187 in the RSA structure, which cannot be accessed from bn_mod_exp.
9188 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
9189 private key operations.
9192 *) Added support for SPARC Linux.
9195 *) pem_password_cb function type incompatibly changed from
9196 typedef int pem_password_cb(char *buf, int size, int rwflag);
9198 ....(char *buf, int size, int rwflag, void *userdata);
9199 so that applications can pass data to their callbacks:
9200 The PEM[_ASN1]_{read,write}... functions and macros now take an
9201 additional void * argument, which is just handed through whenever
9202 the password callback is called.
9203 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
9205 New function SSL_CTX_set_default_passwd_cb_userdata.
9207 Compatibility note: As many C implementations push function arguments
9208 onto the stack in reverse order, the new library version is likely to
9209 interoperate with programs that have been compiled with the old
9210 pem_password_cb definition (PEM_whatever takes some data that
9211 happens to be on the stack as its last argument, and the callback
9212 just ignores this garbage); but there is no guarantee whatsoever that
9215 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
9216 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
9217 problems not only on Windows, but also on some Unix platforms.
9218 To avoid problematic command lines, these definitions are now in an
9219 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
9220 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
9223 *) MIPS III/IV assembler module is reimplemented.
9226 *) More DES library cleanups: remove references to srand/rand and
9227 delete an unused file.
9230 *) Add support for the the free Netwide assembler (NASM) under Win32,
9231 since not many people have MASM (ml) and it can be hard to obtain.
9232 This is currently experimental but it seems to work OK and pass all
9233 the tests. Check out INSTALL.W32 for info.
9236 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
9237 without temporary keys kept an extra copy of the server key,
9238 and connections with temporary keys did not free everything in case
9242 *) New function RSA_check_key and new openssl rsa option -check
9243 for verifying the consistency of RSA keys.
9244 [Ulf Moeller, Bodo Moeller]
9246 *) Various changes to make Win32 compile work:
9247 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
9248 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
9249 comparison" warnings.
9250 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
9253 *) Add a debugging option to PKCS#5 v2 key generation function: when
9254 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
9255 derived keys are printed to stderr.
9258 *) Copy the flags in ASN1_STRING_dup().
9259 [Roman E. Pavlov <pre@mo.msk.ru>]
9261 *) The x509 application mishandled signing requests containing DSA
9262 keys when the signing key was also DSA and the parameters didn't match.
9264 It was supposed to omit the parameters when they matched the signing key:
9265 the verifying software was then supposed to automatically use the CA's
9266 parameters if they were absent from the end user certificate.
9268 Omitting parameters is no longer recommended. The test was also
9269 the wrong way round! This was probably due to unusual behaviour in
9270 EVP_cmp_parameters() which returns 1 if the parameters match.
9271 This meant that parameters were omitted when they *didn't* match and
9272 the certificate was useless. Certificates signed with 'ca' didn't have
9274 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
9276 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
9277 The interface is as follows:
9278 Applications can use
9279 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
9280 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
9281 "off" is now the default.
9282 The library internally uses
9283 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
9284 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
9285 to disable memory-checking temporarily.
9287 Some inconsistent states that previously were possible (and were
9288 even the default) are now avoided.
9290 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
9291 with each memory chunk allocated; this is occasionally more helpful
9292 than just having a counter.
9294 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
9296 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
9300 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
9301 which largely parallels "options", but is for changing API behaviour,
9302 whereas "options" are about protocol behaviour.
9303 Initial "mode" flags are:
9305 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
9306 a single record has been written.
9307 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
9308 retries use the same buffer location.
9309 (But all of the contents must be
9313 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
9316 *) Fix problems with no-hmac etc.
9317 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
9319 *) New functions RSA_get_default_method(), RSA_set_method() and
9320 RSA_get_method(). These allows replacement of RSA_METHODs without having
9321 to mess around with the internals of an RSA structure.
9324 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
9325 Also really enable memory leak checks in openssl.c and in some
9327 [Chad C. Mulligan, Bodo Moeller]
9329 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
9330 up the length of negative integers. This has now been simplified to just
9331 store the length when it is first determined and use it later, rather
9332 than trying to keep track of where data is copied and updating it to
9334 [Steve Henson, reported by Brien Wheeler
9335 <bwheeler@authentica-security.com>]
9337 *) Add a new function PKCS7_signatureVerify. This allows the verification
9338 of a PKCS#7 signature but with the signing certificate passed to the
9339 function itself. This contrasts with PKCS7_dataVerify which assumes the
9340 certificate is present in the PKCS#7 structure. This isn't always the
9341 case: certificates can be omitted from a PKCS#7 structure and be
9342 distributed by "out of band" means (such as a certificate database).
9345 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
9346 function prototypes in pem.h, also change util/mkdef.pl to add the
9347 necessary function names.
9350 *) mk1mf.pl (used by Windows builds) did not properly read the
9351 options set by Configure in the top level Makefile, and Configure
9352 was not even able to write more than one option correctly.
9353 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
9356 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
9357 file to be loaded from a BIO or FILE pointer. The BIO version will
9358 for example allow memory BIOs to contain config info.
9361 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
9362 Whoever hopes to achieve shared-library compatibility across versions
9363 must use this, not the compile-time macro.
9364 (Exercise 0.9.4: Which is the minimum library version required by
9366 Note: All this applies only to multi-threaded programs, others don't
9370 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
9371 through a BIO pair triggered the default case, i.e.
9372 SSLerr(...,SSL_R_UNKNOWN_STATE).
9375 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
9376 can use the SSL library even if none of the specific BIOs is
9380 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
9381 for the encoded length.
9382 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
9384 *) Add initial documentation of the X509V3 functions.
9387 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
9388 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
9389 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
9390 secure PKCS#8 private key format with a high iteration count.
9393 *) Fix determination of Perl interpreter: A perl or perl5
9394 _directory_ in $PATH was also accepted as the interpreter.
9395 [Ralf S. Engelschall]
9397 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
9398 wrong with it but it was very old and did things like calling
9399 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
9403 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
9404 to use the new extension code.
9407 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
9408 with macros. This should make it easier to change their form, add extra
9409 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
9413 *) Add to configuration table a new entry that can specify an alternative
9414 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
9415 according to Mark Crispin <MRC@Panda.COM>.
9419 *) DES CBC did not update the IV. Weird.
9422 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
9423 Changing the behaviour of the former might break existing programs --
9424 where IV updating is needed, des_ncbc_encrypt can be used.
9427 *) When bntest is run from "make test" it drives bc to check its
9428 calculations, as well as internally checking them. If an internal check
9429 fails, it needs to cause bc to give a non-zero result or make test carries
9430 on without noticing the failure. Fixed.
9433 *) DES library cleanups.
9436 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
9437 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
9438 ciphers. NOTE: although the key derivation function has been verified
9439 against some published test vectors it has not been extensively tested
9440 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
9444 *) Instead of "mkdir -p", which is not fully portable, use new
9445 Perl script "util/mkdir-p.pl".
9448 *) Rewrite the way password based encryption (PBE) is handled. It used to
9449 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
9450 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
9451 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
9452 the 'parameter' field of the AlgorithmIdentifier is passed to the
9453 underlying key generation function so it must do its own ASN1 parsing.
9454 This has also changed the EVP_PBE_CipherInit() function which now has a
9455 'parameter' argument instead of literal salt and iteration count values
9456 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
9459 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
9460 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
9461 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
9462 KEY" because this clashed with PKCS#8 unencrypted string. Since this
9463 value was just used as a "magic string" and not used directly its
9464 value doesn't matter.
9467 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9471 *) "linux-sparc64" configuration (ultrapenguin).
9472 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9473 "linux-sparc" configuration.
9474 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9476 *) config now generates no-xxx options for missing ciphers.
9479 *) Support the EBCDIC character set (work in progress).
9480 File ebcdic.c not yet included because it has a different license.
9481 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9483 *) Support BS2000/OSD-POSIX.
9484 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9486 *) Make callbacks for key generation use void * instead of char *.
9489 *) Make S/MIME samples compile (not yet tested).
9492 *) Additional typesafe stacks.
9495 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9499 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9501 *) New configuration variant "sco5-gcc".
9503 *) Updated some demos.
9504 [Sean O Riordain, Wade Scholine]
9506 *) Add missing BIO_free at exit of pkcs12 application.
9509 *) Fix memory leak in conf.c.
9512 *) Updates for Win32 to assembler version of MD5.
9515 *) Set #! path to perl in apps/der_chop to where we found it
9516 instead of using a fixed path.
9519 *) SHA library changes for irix64-mips4-cc.
9522 *) Improvements for VMS support.
9526 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9528 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9529 This also avoids the problems with SC4.2 and unpatched SC5.
9530 [Andy Polyakov <appro@fy.chalmers.se>]
9532 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9533 These are required because of the typesafe stack would otherwise break
9534 existing code. If old code used a structure member which used to be STACK
9535 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9536 sk_num or sk_value it would produce an error because the num, data members
9537 are not present in STACK_OF. Now it just produces a warning. sk_set
9538 replaces the old method of assigning a value to sk_value
9539 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9540 that does this will no longer work (and should use sk_set instead) but
9541 this could be regarded as a "questionable" behaviour anyway.
9544 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9545 correctly handle encrypted S/MIME data.
9548 *) Change type of various DES function arguments from des_cblock
9549 (which means, in function argument declarations, pointer to char)
9550 to des_cblock * (meaning pointer to array with 8 char elements),
9551 which allows the compiler to do more typechecking; it was like
9552 that back in SSLeay, but with lots of ugly casts.
9554 Introduce new type const_des_cblock.
9557 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9558 problems: find RecipientInfo structure that matches recipient certificate
9559 and initialise the ASN1 structures properly based on passed cipher.
9562 *) Belatedly make the BN tests actually check the results.
9565 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9566 to and from BNs: it was completely broken. New compilation option
9567 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9568 key elements as negative integers.
9571 *) Reorganize and speed up MD5.
9572 [Andy Polyakov <appro@fy.chalmers.se>]
9575 [Richard Levitte <richard@levitte.org>]
9577 *) New option -out to asn1parse to allow the parsed structure to be
9578 output to a file. This is most useful when combined with the -strparse
9579 option to examine the output of things like OCTET STRINGS.
9582 *) Make SSL library a little more fool-proof by not requiring any longer
9583 that SSL_set_{accept,connect}_state be called before
9584 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9585 in many applications because usually everything *appeared* to work as
9586 intended anyway -- now it really works as intended).
9589 *) Move openssl.cnf out of lib/.
9592 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
9593 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
9594 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
9595 [Ralf S. Engelschall]
9597 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
9598 handle PKCS#7 enveloped data properly.
9599 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
9601 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
9602 copying pointers. The cert_st handling is changed by this in
9603 various ways (and thus what used to be known as ctx->default_cert
9604 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
9605 any longer when s->cert does not give us what we need).
9606 ssl_cert_instantiate becomes obsolete by this change.
9607 As soon as we've got the new code right (possibly it already is?),
9608 we have solved a couple of bugs of the earlier code where s->cert
9609 was used as if it could not have been shared with other SSL structures.
9611 Note that using the SSL API in certain dirty ways now will result
9612 in different behaviour than observed with earlier library versions:
9613 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
9614 does not influence s as it used to.
9616 In order to clean up things more thoroughly, inside SSL_SESSION
9617 we don't use CERT any longer, but a new structure SESS_CERT
9618 that holds per-session data (if available); currently, this is
9619 the peer's certificate chain and, for clients, the server's certificate
9620 and temporary key. CERT holds only those values that can have
9621 meaningful defaults in an SSL_CTX.
9624 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
9625 from the internal representation. Various PKCS#7 fixes: remove some
9626 evil casts and set the enc_dig_alg field properly based on the signing
9630 *) Allow PKCS#12 password to be set from the command line or the
9631 environment. Let 'ca' get its config file name from the environment
9632 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
9636 *) Allow certificate policies extension to use an IA5STRING for the
9637 organization field. This is contrary to the PKIX definition but
9638 VeriSign uses it and IE5 only recognises this form. Document 'x509'
9642 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
9643 without disallowing inline assembler and the like for non-pedantic builds.
9646 *) Support Borland C++ builder.
9647 [Janez Jere <jj@void.si>, modified by Ulf Möller]
9652 *) SHA-1 cleanups and performance enhancements.
9653 [Andy Polyakov <appro@fy.chalmers.se>]
9655 *) Sparc v8plus assembler for the bignum library.
9656 [Andy Polyakov <appro@fy.chalmers.se>]
9658 *) Accept any -xxx and +xxx compiler options in Configure.
9661 *) Update HPUX configuration.
9664 *) Add missing sk_<type>_unshift() function to safestack.h
9665 [Ralf S. Engelschall]
9667 *) New function SSL_CTX_use_certificate_chain_file that sets the
9668 "extra_cert"s in addition to the certificate. (This makes sense
9669 only for "PEM" format files, as chains as a whole are not
9673 *) Support verify_depth from the SSL API.
9674 x509_vfy.c had what can be considered an off-by-one-error:
9675 Its depth (which was not part of the external interface)
9676 was actually counting the number of certificates in a chain;
9677 now it really counts the depth.
9680 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
9681 instead of X509err, which often resulted in confusing error
9682 messages since the error codes are not globally unique
9683 (e.g. an alleged error in ssl3_accept when a certificate
9684 didn't match the private key).
9686 *) New function SSL_CTX_set_session_id_context that allows to set a default
9687 value (so that you don't need SSL_set_session_id_context for each
9688 connection using the SSL_CTX).
9691 *) OAEP decoding bug fix.
9694 *) Support INSTALL_PREFIX for package builders, as proposed by
9698 *) New Configure options "threads" and "no-threads". For systems
9699 where the proper compiler options are known (currently Solaris
9700 and Linux), "threads" is the default.
9703 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
9706 *) Install various scripts to $(OPENSSLDIR)/misc, not to
9707 $(INSTALLTOP)/bin -- they shouldn't clutter directories
9708 such as /usr/local/bin.
9711 *) "make linux-shared" to build shared libraries.
9712 [Niels Poppe <niels@netbox.org>]
9714 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
9717 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
9718 extension adding in x509 utility.
9721 *) Remove NOPROTO sections and error code comments.
9724 *) Partial rewrite of the DEF file generator to now parse the ANSI
9728 *) New Configure options --prefix=DIR and --openssldir=DIR.
9731 *) Complete rewrite of the error code script(s). It is all now handled
9732 by one script at the top level which handles error code gathering,
9733 header rewriting and C source file generation. It should be much better
9734 than the old method: it now uses a modified version of Ulf's parser to
9735 read the ANSI prototypes in all header files (thus the old K&R definitions
9736 aren't needed for error creation any more) and do a better job of
9737 translating function codes into names. The old 'ASN1 error code imbedded
9738 in a comment' is no longer necessary and it doesn't use .err files which
9739 have now been deleted. Also the error code call doesn't have to appear all
9740 on one line (which resulted in some large lines...).
9743 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
9746 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
9747 0 (which usually indicates a closed connection), but continue reading.
9750 *) Fix some race conditions.
9753 *) Add support for CRL distribution points extension. Add Certificate
9754 Policies and CRL distribution points documentation.
9757 *) Move the autogenerated header file parts to crypto/opensslconf.h.
9760 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
9761 8 of keying material. Merlin has also confirmed interop with this fix
9762 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
9763 [Merlin Hughes <merlin@baltimore.ie>]
9765 *) Fix lots of warnings.
9766 [Richard Levitte <levitte@stacken.kth.se>]
9768 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
9769 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9770 [Richard Levitte <levitte@stacken.kth.se>]
9772 *) Fix problems with sizeof(long) == 8.
9773 [Andy Polyakov <appro@fy.chalmers.se>]
9775 *) Change functions to ANSI C.
9778 *) Fix typos in error codes.
9779 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9781 *) Remove defunct assembler files from Configure.
9784 *) SPARC v8 assembler BIGNUM implementation.
9785 [Andy Polyakov <appro@fy.chalmers.se>]
9787 *) Support for Certificate Policies extension: both print and set.
9788 Various additions to support the r2i method this uses.
9791 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9792 return a const string when you are expecting an allocated buffer.
9795 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9796 types DirectoryString and DisplayText.
9799 *) Add code to allow r2i extensions to access the configuration database,
9800 add an LHASH database driver and add several ctx helper functions.
9803 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9804 fail when they extended the size of a BIGNUM.
9807 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9808 support typesafe stack.
9811 *) Fix typo in SSL_[gs]et_options().
9812 [Nils Frostberg <nils@medcom.se>]
9814 *) Delete various functions and files that belonged to the (now obsolete)
9815 old X509V3 handling code.
9818 *) New Configure option "rsaref".
9821 *) Don't auto-generate pem.h.
9824 *) Introduce type-safe ASN.1 SETs.
9827 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9828 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9830 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9831 that links with OpenSSL (well at least cause lots of warnings), but fear
9832 not: the conversion is trivial, and it eliminates loads of evil casts. A
9833 few STACKed things have been converted already. Feel free to convert more.
9834 In the fullness of time, I'll do away with the STACK type altogether.
9837 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9838 specified in <certfile> by updating the entry in the index.txt file.
9839 This way one no longer has to edit the index.txt file manually for
9840 revoking a certificate. The -revoke option does the gory details now.
9841 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9843 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9844 `-text' option at all and this way the `-noout -text' combination was
9845 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9846 [Ralf S. Engelschall]
9848 *) Make sure a corresponding plain text error message exists for the
9849 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9850 verify callback function determined that a certificate was revoked.
9851 [Ralf S. Engelschall]
9853 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9854 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9855 all available cipers including rc5, which was forgotten until now.
9856 In order to let the testing shell script know which algorithms
9857 are available, a new (up to now undocumented) command
9858 "openssl list-cipher-commands" is used.
9861 *) Bugfix: s_client occasionally would sleep in select() when
9862 it should have checked SSL_pending() first.
9865 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9866 the raw DSA values prior to ASN.1 encoding.
9869 *) Tweaks to Configure
9870 [Niels Poppe <niels@netbox.org>]
9872 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9876 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9879 *) New config option to avoid instructions that are illegal on the 80386.
9880 The default code is faster, but requires at least a 486.
9883 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9884 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9885 same as SSL2_VERSION anyway.
9888 *) New "-showcerts" option for s_client.
9891 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9892 application. Various cleanups and fixes.
9895 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9896 modify error routines to work internally. Add error codes and PBE init
9897 to library startup routines.
9900 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9901 packing functions to asn1 and evp. Changed function names and error
9902 codes along the way.
9905 *) PKCS12 integration: and so it begins... First of several patches to
9906 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9907 objects to objects.h
9910 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9911 and display support for Thawte strong extranet extension.
9914 *) Add LinuxPPC support.
9915 [Jeff Dubrule <igor@pobox.org>]
9917 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9918 bn_div_words in alpha.s.
9919 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9921 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9922 OAEP isn't supported when OpenSSL is built with RSAref.
9923 [Ulf Moeller <ulf@fitug.de>]
9925 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9926 so they no longer are missing under -DNOPROTO.
9927 [Soren S. Jorvang <soren@t.dk>]
9930 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9932 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9933 doesn't work when the session is reused. Coming soon!
9936 *) Fix a security hole, that allows sessions to be reused in the wrong
9937 context thus bypassing client cert protection! All software that uses
9938 client certs and session caches in multiple contexts NEEDS PATCHING to
9939 allow session reuse! A fuller solution is in the works.
9940 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9942 *) Some more source tree cleanups (removed obsolete files
9943 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9944 permission on "config" script to be executable) and a fix for the INSTALL
9946 [Ulf Moeller <ulf@fitug.de>]
9948 *) Remove some legacy and erroneous uses of malloc, free instead of
9950 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9952 *) Make rsa_oaep_test return non-zero on error.
9953 [Ulf Moeller <ulf@fitug.de>]
9955 *) Add support for native Solaris shared libraries. Configure
9956 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9957 if someone would make that last step automatic.
9958 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9960 *) ctx_size was not built with the right compiler during "make links". Fixed.
9963 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9964 except NULL ciphers". This means the default cipher list will no longer
9965 enable NULL ciphers. They need to be specifically enabled e.g. with
9966 the string "DEFAULT:eNULL".
9969 *) Fix to RSA private encryption routines: if p < q then it would
9970 occasionally produce an invalid result. This will only happen with
9971 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9974 *) Be less restrictive and allow also `perl util/perlpath.pl
9975 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9976 because this way one can also use an interpreter named `perl5' (which is
9977 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9978 installed as `perl').
9979 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9981 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9982 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9984 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9985 advapi32.lib to Win32 build and change the pem test comparision
9986 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9987 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9988 and crypto/des/ede_cbcm_enc.c.
9991 *) DES quad checksum was broken on big-endian architectures. Fixed.
9994 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9995 Win32 test batch file so it (might) work again. The Win32 test batch file
9996 is horrible: I feel ill....
9999 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
10000 in e_os.h. Audit of header files to check ANSI and non ANSI
10001 sections: 10 functions were absent from non ANSI section and not exported
10002 from Windows DLLs. Fixed up libeay.num for new functions.
10005 *) Make `openssl version' output lines consistent.
10006 [Ralf S. Engelschall]
10008 *) Fix Win32 symbol export lists for BIO functions: Added
10009 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10010 to ms/libeay{16,32}.def.
10011 [Ralf S. Engelschall]
10013 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10014 fine under Unix and passes some trivial tests I've now added. But the
10015 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10016 added to make sure no one expects that this stuff really works in the
10017 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10018 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10020 [Ralf S. Engelschall]
10022 *) Fix the generation of two part addresses in perl.
10023 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10025 *) Add config entry for Linux on MIPS.
10026 [John Tobey <jtobey@channel1.com>]
10028 *) Make links whenever Configure is run, unless we are on Windoze.
10031 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10032 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10036 *) Add a useful kludge to allow package maintainers to specify compiler and
10037 other platforms details on the command line without having to patch the
10038 Configure script everytime: One now can use ``perl Configure
10039 <id>:<details>'', i.e. platform ids are allowed to have details appended
10040 to them (seperated by colons). This is treated as there would be a static
10041 pre-configured entry in Configure's %table under key <id> with value
10042 <details> and ``perl Configure <id>'' is called. So, when you want to
10043 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10044 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10045 now, which overrides the FreeBSD-elf entry on-the-fly.
10046 [Ralf S. Engelschall]
10048 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10051 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10052 on the `perl Configure ...' command line. This way one can compile
10053 OpenSSL libraries with Position Independent Code (PIC) which is needed
10054 for linking it into DSOs.
10055 [Ralf S. Engelschall]
10057 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10061 *) Cleaned up the LICENSE document: The official contact for any license
10062 questions now is the OpenSSL core team under openssl-core@openssl.org.
10063 And add a paragraph about the dual-license situation to make sure people
10064 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10065 to the OpenSSL toolkit.
10066 [Ralf S. Engelschall]
10068 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10069 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10070 Additonally cleaned up the `make links' target: Remove unnecessary
10071 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10072 to speed processing and no longer clutter the display with confusing
10073 stuff. Instead only the actually done links are displayed.
10074 [Ralf S. Engelschall]
10076 *) Permit null encryption ciphersuites, used for authentication only. It used
10077 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10078 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10082 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10083 signed attributes when verifying signatures (this would break them),
10084 the detached data encoding was wrong and public keys obtained using
10085 X509_get_pubkey() weren't freed.
10088 *) Add text documentation for the BUFFER functions. Also added a work around
10089 to a Win95 console bug. This was triggered by the password read stuff: the
10090 last character typed gets carried over to the next fread(). If you were
10091 generating a new cert request using 'req' for example then the last
10092 character of the passphrase would be CR which would then enter the first
10096 *) Added the new `Includes OpenSSL Cryptography Software' button as
10097 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10098 button and can be used by applications based on OpenSSL to show the
10099 relationship to the OpenSSL project.
10100 [Ralf S. Engelschall]
10102 *) Remove confusing variables in function signatures in files
10103 ssl/ssl_lib.c and ssl/ssl.h.
10104 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10106 *) Don't install bss_file.c under PREFIX/include/
10107 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10109 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10110 functions that return function pointers and has support for NT specific
10111 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10112 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10113 unsigned to signed types: this was killing the Win32 compile.
10116 *) Add new certificate file to stack functions,
10117 SSL_add_dir_cert_subjects_to_stack() and
10118 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10119 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10120 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10121 This means that Apache-SSL and similar packages don't have to mess around
10122 to add as many CAs as they want to the preferred list.
10125 *) Experiment with doxygen documentation. Currently only partially applied to
10127 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10128 openssl.doxy as the configuration file.
10131 *) Get rid of remaining C++-style comments which strict C compilers hate.
10132 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10134 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10135 compiled in by default: it has problems with large keys.
10138 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10139 DH private keys and/or callback functions which directly correspond to
10140 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10141 is needed for applications which have to configure certificates on a
10142 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10144 For the RSA certificate situation is makes no difference, but
10145 for the DSA certificate situation this fixes the "no shared cipher"
10146 problem where the OpenSSL cipher selection procedure failed because the
10147 temporary keys were not overtaken from the context and the API provided
10148 no way to reconfigure them.
10149 The new functions now let applications reconfigure the stuff and they
10150 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
10151 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
10152 non-public-API function ssl_cert_instantiate() is used as a helper
10153 function and also to reduce code redundancy inside ssl_rsa.c.
10154 [Ralf S. Engelschall]
10156 *) Move s_server -dcert and -dkey options out of the undocumented feature
10157 area because they are useful for the DSA situation and should be
10158 recognized by the users.
10159 [Ralf S. Engelschall]
10161 *) Fix the cipher decision scheme for export ciphers: the export bits are
10162 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
10163 SSL_EXP_MASK. So, the original variable has to be used instead of the
10164 already masked variable.
10165 [Richard Levitte <levitte@stacken.kth.se>]
10167 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
10168 [Richard Levitte <levitte@stacken.kth.se>]
10170 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
10171 from `int' to `unsigned int' because it's a length and initialized by
10172 EVP_DigestFinal() which expects an `unsigned int *'.
10173 [Richard Levitte <levitte@stacken.kth.se>]
10175 *) Don't hard-code path to Perl interpreter on shebang line of Configure
10176 script. Instead use the usual Shell->Perl transition trick.
10177 [Ralf S. Engelschall]
10179 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
10180 (in addition to RSA certificates) to match the behaviour of `openssl dsa
10181 -noout -modulus' as it's already the case for `openssl rsa -noout
10182 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
10183 currently the public key is printed (a decision which was already done by
10184 `openssl dsa -modulus' in the past) which serves a similar purpose.
10185 Additionally the NO_RSA no longer completely removes the whole -modulus
10186 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
10188 [Ralf S. Engelschall]
10190 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
10191 BIO. See the source (crypto/evp/bio_ok.c) for more info.
10192 [Arne Ansper <arne@ats.cyber.ee>]
10194 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
10195 to be added. Now both 'req' and 'ca' can use new objects defined in the
10199 *) Add cool BIO that does syslog (or event log on NT).
10200 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
10202 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
10203 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
10204 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
10205 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
10208 *) Add preliminary config info for new extension code.
10211 *) Make RSA_NO_PADDING really use no padding.
10212 [Ulf Moeller <ulf@fitug.de>]
10214 *) Generate errors when private/public key check is done.
10217 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
10218 for some CRL extensions and new objects added.
10221 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
10222 key usage extension and fuller support for authority key id.
10225 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
10226 padding method for RSA, which is recommended for new applications in PKCS
10227 #1 v2.0 (RFC 2437, October 1998).
10228 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
10229 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
10230 against Bleichbacher's attack on RSA.
10231 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
10234 *) Updates to the new SSL compression code
10235 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10237 *) Fix so that the version number in the master secret, when passed
10238 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
10239 (because the server will not accept higher), that the version number
10240 is 0x03,0x01, not 0x03,0x00
10241 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10243 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
10244 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
10245 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
10248 *) Support for RAW extensions where an arbitrary extension can be
10249 created by including its DER encoding. See apps/openssl.cnf for
10253 *) Make sure latest Perl versions don't interpret some generated C array
10254 code as Perl array code in the crypto/err/err_genc.pl script.
10255 [Lars Weber <3weber@informatik.uni-hamburg.de>]
10257 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
10258 not many people have the assembler. Various Win32 compilation fixes and
10259 update to the INSTALL.W32 file with (hopefully) more accurate Win32
10260 build instructions.
10263 *) Modify configure script 'Configure' to automatically create crypto/date.h
10264 file under Win32 and also build pem.h from pem.org. New script
10265 util/mkfiles.pl to create the MINFO file on environments that can't do a
10266 'make files': perl util/mkfiles.pl >MINFO should work.
10269 *) Major rework of DES function declarations, in the pursuit of correctness
10270 and purity. As a result, many evil casts evaporated, and some weirdness,
10271 too. You may find this causes warnings in your code. Zapping your evil
10272 casts will probably fix them. Mostly.
10275 *) Fix for a typo in asn1.h. Bug fix to object creation script
10276 obj_dat.pl. It considered a zero in an object definition to mean
10277 "end of object": none of the objects in objects.h have any zeros
10278 so it wasn't spotted.
10279 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
10281 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
10282 Masking (CBCM). In the absence of test vectors, the best I have been able
10283 to do is check that the decrypt undoes the encrypt, so far. Send me test
10284 vectors if you have them.
10287 *) Correct calculation of key length for export ciphers (too much space was
10288 allocated for null ciphers). This has not been tested!
10291 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
10292 message is now correct (it understands "crypto" and "ssl" on its
10293 command line). There is also now an "update" option. This will update
10294 the util/ssleay.num and util/libeay.num files with any new functions.
10296 perl util/mkdef.pl crypto ssl update
10297 it will update them.
10300 *) Overhauled the Perl interface (perl/*):
10301 - ported BN stuff to OpenSSL's different BN library
10302 - made the perl/ source tree CVS-aware
10303 - renamed the package from SSLeay to OpenSSL (the files still contain
10304 their history because I've copied them in the repository)
10305 - removed obsolete files (the test scripts will be replaced
10306 by better Test::Harness variants in the future)
10307 [Ralf S. Engelschall]
10309 *) First cut for a very conservative source tree cleanup:
10310 1. merge various obsolete readme texts into doc/ssleay.txt
10311 where we collect the old documents and readme texts.
10312 2. remove the first part of files where I'm already sure that we no
10313 longer need them because of three reasons: either they are just temporary
10314 files which were left by Eric or they are preserved original files where
10315 I've verified that the diff is also available in the CVS via "cvs diff
10316 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
10317 the crypto/md/ stuff).
10318 [Ralf S. Engelschall]
10320 *) More extension code. Incomplete support for subject and issuer alt
10321 name, issuer and authority key id. Change the i2v function parameters
10322 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
10323 what that's for :-) Fix to ASN1 macro which messed up
10324 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
10327 *) Preliminary support for ENUMERATED type. This is largely copied from the
10331 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
10332 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10334 *) Make sure `make rehash' target really finds the `openssl' program.
10335 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10337 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
10338 like to hear about it if this slows down other processors.
10341 *) Add CygWin32 platform information to Configure script.
10342 [Alan Batie <batie@aahz.jf.intel.com>]
10344 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
10345 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
10347 *) New program nseq to manipulate netscape certificate sequences
10350 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
10354 *) Fixes to BN code. Previously the default was to define BN_RECURSION
10355 but the BN code had some problems that would cause failures when
10356 doing certificate verification and some other functions.
10357 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10359 *) Add ASN1 and PEM code to support netscape certificate sequences.
10362 *) Add ASN1 and PEM code to support netscape certificate sequences.
10365 *) Add several PKIX and private extended key usage OIDs.
10368 *) Modify the 'ca' program to handle the new extension code. Modify
10369 openssl.cnf for new extension format, add comments.
10372 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
10373 and add a sample to openssl.cnf so req -x509 now adds appropriate
10377 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
10378 error code, add initial support to X509_print() and x509 application.
10381 *) Takes a deep breath and start addding X509 V3 extension support code. Add
10382 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
10383 stuff is currently isolated and isn't even compiled yet.
10386 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
10387 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
10388 Removed the versions check from X509 routines when loading extensions:
10389 this allows certain broken certificates that don't set the version
10390 properly to be processed.
10393 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
10394 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
10395 can still be regenerated with "make depend".
10398 *) Spelling mistake in C version of CAST-128.
10399 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
10401 *) Changes to the error generation code. The perl script err-code.pl
10402 now reads in the old error codes and retains the old numbers, only
10403 adding new ones if necessary. It also only changes the .err files if new
10404 codes are added. The makefiles have been modified to only insert errors
10405 when needed (to avoid needlessly modifying header files). This is done
10406 by only inserting errors if the .err file is newer than the auto generated
10407 C file. To rebuild all the error codes from scratch (the old behaviour)
10408 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
10409 or delete all the .err files.
10412 *) CAST-128 was incorrectly implemented for short keys. The C version has
10413 been fixed, but is untested. The assembler versions are also fixed, but
10414 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
10415 to regenerate it if needed.
10416 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
10417 Hagino <itojun@kame.net>]
10419 *) File was opened incorrectly in randfile.c.
10420 [Ulf Möller <ulf@fitug.de>]
10422 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
10423 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
10424 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
10425 al: it's just almost always a UTCTime. Note this patch adds new error
10426 codes so do a "make errors" if there are problems.
10429 *) Correct Linux 1 recognition in config.
10430 [Ulf Möller <ulf@fitug.de>]
10432 *) Remove pointless MD5 hash when using DSA keys in ca.
10433 [Anonymous <nobody@replay.com>]
10435 *) Generate an error if given an empty string as a cert directory. Also
10436 generate an error if handed NULL (previously returned 0 to indicate an
10437 error, but didn't set one).
10438 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
10440 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
10443 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
10444 parameters. This was causing a warning which killed off the Win32 compile.
10447 *) Remove C++ style comments from crypto/bn/bn_local.h.
10448 [Neil Costigan <neil.costigan@celocom.com>]
10450 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
10451 based on a text string, looking up short and long names and finally
10452 "dot" format. The "dot" format stuff didn't work. Added new function
10453 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
10454 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
10455 OID is not part of the table.
10458 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
10459 X509_LOOKUP_by_alias().
10462 *) Sort openssl functions by name.
10465 *) Get the gendsa program working (hopefully) and add it to app list. Remove
10466 encryption from sample DSA keys (in case anyone is interested the password
10470 *) Make _all_ *_free functions accept a NULL pointer.
10471 [Frans Heymans <fheymans@isaserver.be>]
10473 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10475 [Anonymous <nobody@replay.com>]
10477 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10478 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10480 *) Don't blow it for numeric -newkey arguments to apps/req.
10481 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10483 *) Temp key "for export" tests were wrong in s3_srvr.c.
10484 [Anonymous <nobody@replay.com>]
10486 *) Add prototype for temp key callback functions
10487 SSL_CTX_set_tmp_{rsa,dh}_callback().
10490 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10491 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10494 *) X509_name_add_entry() freed the wrong thing after an error.
10495 [Arne Ansper <arne@ats.cyber.ee>]
10497 *) rsa_eay.c would attempt to free a NULL context.
10498 [Arne Ansper <arne@ats.cyber.ee>]
10500 *) BIO_s_socket() had a broken should_retry() on Windoze.
10501 [Arne Ansper <arne@ats.cyber.ee>]
10503 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10504 [Arne Ansper <arne@ats.cyber.ee>]
10506 *) Make sure the already existing X509_STORE->depth variable is initialized
10507 in X509_STORE_new(), but document the fact that this variable is still
10508 unused in the certificate verification process.
10509 [Ralf S. Engelschall]
10511 *) Fix the various library and apps files to free up pkeys obtained from
10512 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10515 *) Fix reference counting in X509_PUBKEY_get(). This makes
10516 demos/maurice/example2.c work, amongst others, probably.
10517 [Steve Henson and Ben Laurie]
10519 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10520 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10521 are no longer created. This way we have a single and consistent command
10522 line interface `openssl <command>', similar to `cvs <command>'.
10523 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10525 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10526 BIT STRING wrapper always have zero unused bits.
10529 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10532 *) Make the top-level INSTALL documentation easier to understand.
10535 *) Makefiles updated to exit if an error occurs in a sub-directory
10536 make (including if user presses ^C) [Paul Sutton]
10538 *) Make Montgomery context stuff explicit in RSA data structure.
10541 *) Fix build order of pem and err to allow for generated pem.h.
10544 *) Fix renumbering bug in X509_NAME_delete_entry().
10547 *) Enhanced the err-ins.pl script so it makes the error library number
10548 global and can add a library name. This is needed for external ASN1 and
10549 other error libraries.
10552 *) Fixed sk_insert which never worked properly.
10555 *) Fix ASN1 macros so they can handle indefinite length construted
10556 EXPLICIT tags. Some non standard certificates use these: they can now
10560 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10561 into a single doc/ssleay.txt bundle. This way the information is still
10562 preserved but no longer messes up this directory. Now it's new room for
10563 the new set of documenation files.
10564 [Ralf S. Engelschall]
10566 *) SETs were incorrectly DER encoded. This was a major pain, because they
10567 shared code with SEQUENCEs, which aren't coded the same. This means that
10568 almost everything to do with SETs or SEQUENCEs has either changed name or
10569 number of arguments.
10570 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10572 *) Fix test data to work with the above.
10575 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10576 was already fixed by Eric for 0.9.1 it seems.
10577 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10579 *) Autodetect FreeBSD3.
10582 *) Fix various bugs in Configure. This affects the following platforms:
10586 unixware-2.0-pentium
10590 *) Eliminate generated files from CVS. Reorder tests to regenerate files
10591 before they are needed.
10594 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
10598 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
10600 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
10601 changed SSLeay to OpenSSL in version strings.
10602 [Ralf S. Engelschall]
10604 *) Some fixups to the top-level documents.
10607 *) Fixed the nasty bug where rsaref.h was not found under compile-time
10608 because the symlink to include/ was missing.
10609 [Ralf S. Engelschall]
10611 *) Incorporated the popular no-RSA/DSA-only patches
10612 which allow to compile a RSA-free SSLeay.
10613 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
10615 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
10616 when "ssleay" is still not found.
10617 [Ralf S. Engelschall]
10619 *) Added more platforms to Configure: Cray T3E, HPUX 11,
10620 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
10622 *) Updated the README file.
10623 [Ralf S. Engelschall]
10625 *) Added various .cvsignore files in the CVS repository subdirs
10626 to make a "cvs update" really silent.
10627 [Ralf S. Engelschall]
10629 *) Recompiled the error-definition header files and added
10630 missing symbols to the Win32 linker tables.
10631 [Ralf S. Engelschall]
10633 *) Cleaned up the top-level documents;
10634 o new files: CHANGES and LICENSE
10635 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
10636 o merged COPYRIGHT into LICENSE
10637 o removed obsolete TODO file
10638 o renamed MICROSOFT to INSTALL.W32
10639 [Ralf S. Engelschall]
10641 *) Removed dummy files from the 0.9.1b source tree:
10642 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
10643 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
10644 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
10645 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
10646 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
10647 [Ralf S. Engelschall]
10649 *) Added various platform portability fixes.
10652 *) The Genesis of the OpenSSL rpject:
10653 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
10654 Young and Tim J. Hudson created while they were working for C2Net until
10656 [The OpenSSL Project]
10659 Changes between 0.9.0b and 0.9.1b [not released]
10661 *) Updated a few CA certificates under certs/
10664 *) Changed some BIGNUM api stuff.
10667 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
10668 DGUX x86, Linux Alpha, etc.
10671 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
10672 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
10676 *) Add -strparse option to asn1pars program which parses nested
10678 [Dr Stephen Henson <shenson@bigfoot.com>]
10680 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
10683 *) DSA fix for "ca" program.
10686 *) Added "-genkey" option to "dsaparam" program.
10689 *) Added RIPE MD160 (rmd160) message digest.
10692 *) Added -a (all) option to "ssleay version" command.
10695 *) Added PLATFORM define which is the id given to Configure.
10698 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
10701 *) Extended the ASN.1 parser routines.
10704 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
10707 *) Added a BN_CTX to the BN library.
10710 *) Fixed the weak key values in DES library
10713 *) Changed API in EVP library for cipher aliases.
10716 *) Added support for RC2/64bit cipher.
10719 *) Converted the lhash library to the crypto/mem.c functions.
10722 *) Added more recognized ASN.1 object ids.
10725 *) Added more RSA padding checks for SSL/TLS.
10728 *) Added BIO proxy/filter functionality.
10731 *) Added extra_certs to SSL_CTX which can be used
10732 send extra CA certificates to the client in the CA cert chain sending
10733 process. It can be configured with SSL_CTX_add_extra_chain_cert().
10736 *) Now Fortezza is denied in the authentication phase because
10737 this is key exchange mechanism is not supported by SSLeay at all.
10740 *) Additional PKCS1 checks.
10743 *) Support the string "TLSv1" for all TLS v1 ciphers.
10746 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
10747 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
10750 *) Fixed a few memory leaks.
10753 *) Fixed various code and comment typos.
10756 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
10757 bytes sent in the client random.
10758 [Edward Bishop <ebishop@spyglass.com>]