5 Changes between 1.0.2 and 1.1.0 [xx XXX xxxx]
7 *) Experimental encrypt-then-mac support.
9 Experimental support for encrypt then mac from
10 draft-gutmann-tls-encrypt-then-mac-02.txt
12 To enable it set the appropriate extension number (0x42 for the test
13 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
15 For non-compliant peers (i.e. just about everything) this should have no
18 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
22 *) Add callbacks supporting generation and retrieval of supplemental
24 [Scott Deboy <sdeboy@apache.org>, Trevor Perrin and Ben Laurie]
26 *) Add EVP support for key wrapping algorithms, to avoid problems with
27 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
28 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
29 algorithms and include tests cases.
32 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
36 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
37 MGF1 digest and OAEP label.
40 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
41 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
42 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
45 *) Make openssl verify return errors.
46 [Chris Palmer <palmer@google.com> and Ben Laurie]
48 *) New function ASN1_TIME_diff to calculate the difference between two
49 ASN1_TIME structures or one structure and the current time.
52 *) Update fips_test_suite to support multiple command line options. New
53 test to induce all self test errors in sequence and check expected
57 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
58 sign or verify all in one operation.
61 *) Add fips_algvs: a multicall fips utility incorporaing all the algorithm
62 test programs and fips_test_suite. Includes functionality to parse
63 the minimal script output of fipsalgest.pl directly.
66 *) Add authorisation parameter to FIPS_module_mode_set().
69 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
72 *) Use separate DRBG fields for internal and external flags. New function
73 FIPS_drbg_health_check() to perform on demand health checking. Add
74 generation tests to fips_test_suite with reduced health check interval to
75 demonstrate periodic health checking. Add "nodh" option to
76 fips_test_suite to skip very slow DH test.
79 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
83 *) More extensive health check for DRBG checking many more failure modes.
84 New function FIPS_selftest_drbg_all() to handle every possible DRBG
85 combination: call this in fips_test_suite.
88 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
89 and POST to handle Dual EC cases.
92 *) Add support for canonical generation of DSA parameter 'g'. See
95 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
96 POST to handle HMAC cases.
99 *) Add functions FIPS_module_version() and FIPS_module_version_text()
100 to return numerical and string versions of the FIPS module number.
103 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
104 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
105 outside the validated module in the FIPS capable OpenSSL.
108 *) Minor change to DRBG entropy callback semantics. In some cases
109 there is no multiple of the block length between min_len and
110 max_len. Allow the callback to return more than max_len bytes
111 of entropy but discard any extra: it is the callback's responsibility
112 to ensure that the extra data discarded does not impact the
113 requested amount of entropy.
116 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
117 information in FIPS186-3, SP800-57 and SP800-131A.
120 *) CCM support via EVP. Interface is very similar to GCM case except we
121 must supply all data in one chunk (i.e. no update, final) and the
122 message length must be supplied if AAD is used. Add algorithm test
126 *) Initial version of POST overhaul. Add POST callback to allow the status
127 of POST to be monitored and/or failures induced. Modify fips_test_suite
128 to use callback. Always run all selftests even if one fails.
131 *) XTS support including algorithm test driver in the fips_gcmtest program.
132 Note: this does increase the maximum key length from 32 to 64 bytes but
133 there should be no binary compatibility issues as existing applications
134 will never use XTS mode.
137 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
138 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
139 performs algorithm blocking for unapproved PRNG types. Also do not
140 set PRNG type in FIPS_mode_set(): leave this to the application.
141 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
142 the standard OpenSSL PRNG: set additional data to a date time vector.
145 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
146 This shouldn't present any incompatibility problems because applications
147 shouldn't be using these directly and any that are will need to rethink
148 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
151 *) Extensive self tests and health checking required by SP800-90 DRBG.
152 Remove strength parameter from FIPS_drbg_instantiate and always
153 instantiate at maximum supported strength.
156 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
159 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
162 *) New function DH_compute_key_padded() to compute a DH key and pad with
163 leading zeroes if needed: this complies with SP800-56A et al.
166 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
167 anything, incomplete, subject to change and largely untested at present.
170 *) Modify fipscanisteronly build option to only build the necessary object
171 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
174 *) Add experimental option FIPSSYMS to give all symbols in
175 fipscanister.o and FIPS or fips prefix. This will avoid
176 conflicts with future versions of OpenSSL. Add perl script
177 util/fipsas.pl to preprocess assembly language source files
178 and rename any affected symbols.
181 *) Add selftest checks and algorithm block of non-fips algorithms in
182 FIPS mode. Remove DES2 from selftests.
185 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
186 return internal method without any ENGINE dependencies. Add new
187 tiny fips sign and verify functions.
190 *) New build option no-ec2m to disable characteristic 2 code.
193 *) New build option "fipscanisteronly". This only builds fipscanister.o
194 and (currently) associated fips utilities. Uses the file Makefile.fips
195 instead of Makefile.org as the prototype.
198 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
199 Update fips_gcmtest to use IV generator.
202 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
203 setting output buffer to NULL. The *Final function must be
204 called although it will not retrieve any additional data. The tag
205 can be set or retrieved with a ctrl. The IV length is by default 12
206 bytes (96 bits) but can be set to an alternative value. If the IV
207 length exceeds the maximum IV length (currently 16 bytes) it cannot be
211 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
212 underlying do_cipher function handles all cipher semantics itself
213 including padding and finalisation. This is useful if (for example)
214 an ENGINE cipher handles block padding itself. The behaviour of
215 do_cipher is subtly changed if this flag is set: the return value
216 is the number of characters written to the output buffer (zero is
217 no longer an error code) or a negative error code. Also if the
218 input buffer is NULL and length 0 finalisation should be performed.
221 *) If a candidate issuer certificate is already part of the constructed
222 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
225 *) Improve forward-security support: add functions
227 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
228 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
230 for use by SSL/TLS servers; the callback function will be called whenever a
231 new session is created, and gets to decide whether the session may be
232 cached to make it resumable (return 0) or not (return 1). (As by the
233 SSL/TLS protocol specifications, the session_id sent by the server will be
234 empty to indicate that the session is not resumable; also, the server will
235 not generate RFC 4507 (RFC 5077) session tickets.)
237 A simple reasonable callback implementation is to return is_forward_secure.
238 This parameter will be set to 1 or 0 depending on the ciphersuite selected
239 by the SSL/TLS server library, indicating whether it can provide forward
241 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
243 *) New -verify_name option in command line utilities to set verification
247 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
248 Add CMAC pkey methods.
251 *) Experimental regnegotiation in s_server -www mode. If the client
252 browses /reneg connection is renegotiated. If /renegcert it is
253 renegotiated requesting a certificate.
256 *) Add an "external" session cache for debugging purposes to s_server. This
257 should help trace issues which normally are only apparent in deployed
258 multi-process servers.
261 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
262 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
263 BIO_set_cipher() and some obscure PEM functions were changed so they
264 can now return an error. The RAND changes required a change to the
265 RAND_METHOD structure.
268 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
269 a gcc attribute to warn if the result of a function is ignored. This
270 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
271 whose return value is often ignored.
274 Changes between 1.0.1e and 1.0.2 [xx XXX xxxx]
276 *) TLS pad extension: draft-agl-tls-padding-02
278 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
279 TLS client Hello record length value would otherwise be > 255 and
280 less that 512 pad with a dummy extension containing zeroes so it
281 is at least 512 bytes long.
283 To enable it use an unused extension number (for example chrome uses
286 e.g. -DTLSEXT_TYPE_padding=35655
288 Since the extension is ignored the actual number doesn't matter as long
289 as it doesn't clash with any existing extension.
291 This will be updated when the extension gets an official number.
293 [Adam Langley, Steve Henson]
295 *) Add functions to allocate and set the fields of an ECDSA_METHOD
297 [Douglas E. Engert, Steve Henson]
299 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
300 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
301 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
302 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
303 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
304 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
305 [Rob Stradling, Adam Langley]
307 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
308 difference in days and seconds between two tm or ASN1_TIME structures.
311 *) Add -rev test option to s_server to just reverse order of characters
312 received by client and send back to server. Also prints an abbreviated
313 summary of the connection parameters.
316 *) New option -brief for s_client and s_server to print out a brief summary
317 of connection parameters.
320 *) Add callbacks for arbitrary TLS extensions.
321 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
323 *) New option -crl_download in several openssl utilities to download CRLs
324 from CRLDP extension in certificates.
327 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
330 *) New function X509_CRL_diff to generate a delta CRL from the difference
331 of two full CRLs. Add support to "crl" utility.
334 *) New functions to set lookup_crls function and to retrieve
335 X509_STORE from X509_STORE_CTX.
338 *) Print out deprecated issuer and subject unique ID fields in
342 *) Extend OCSP I/O functions so they can be used for simple general purpose
343 HTTP as well as OCSP. New wrapper function which can be used to download
344 CRLs using the OCSP API.
347 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
350 *) SSL_CONF* functions. These provide a common framework for application
351 configuration using configuration files or command lines.
354 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
355 message callback and prints the results. Needs compile time option
356 "enable-ssl-trace". New options to s_client and s_server to enable
360 *) New ctrl and macro to retrieve supported points extensions.
361 Print out extension in s_server and s_client.
364 *) New functions to retrieve certificate signature and signature
368 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
372 *) New Suite B modes for TLS code. These use and enforce the requirements
373 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
374 only use Suite B curves. The Suite B modes can be set by using the
375 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
378 *) New chain verification flags for Suite B levels of security. Check
379 algorithms are acceptable when flags are set in X509_verify_cert.
382 *) Make tls1_check_chain return a set of flags indicating checks passed
383 by a certificate chain. Add additional tests to handle client
384 certificates: checks for matching certificate type and issuer name
388 *) If an attempt is made to use a signature algorithm not in the peer
389 preference list abort the handshake. If client has no suitable
390 signature algorithms in response to a certificate request do not
394 *) If server EC tmp key is not in client preference list abort handshake.
397 *) Add support for certificate stores in CERT structure. This makes it
398 possible to have different stores per SSL structure or one store in
399 the parent SSL_CTX. Include distint stores for certificate chain
400 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
401 to build and store a certificate chain in CERT structure: returing
402 an error if the chain cannot be built: this will allow applications
403 to test if a chain is correctly configured.
405 Note: if the CERT based stores are not set then the parent SSL_CTX
406 store is used to retain compatibility with existing behaviour.
410 *) New function ssl_set_client_disabled to set a ciphersuite disabled
411 mask based on the current session, check mask when sending client
412 hello and checking the requested ciphersuite.
415 *) New ctrls to retrieve and set certificate types in a certificate
416 request message. Print out received values in s_client. If certificate
417 types is not set with custom values set sensible values based on
418 supported signature algorithms.
421 *) Support for distinct client and server supported signature algorithms.
424 *) Add certificate callback. If set this is called whenever a certificate
425 is required by client or server. An application can decide which
426 certificate chain to present based on arbitrary criteria: for example
427 supported signature algorithms. Add very simple example to s_server.
428 This fixes many of the problems and restrictions of the existing client
429 certificate callback: for example you can now clear an existing
430 certificate and specify the whole chain.
433 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
434 the certificate can be used for (if anything). Set valid_flags field
435 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
436 to have similar checks in it.
438 Add new "cert_flags" field to CERT structure and include a "strict mode".
439 This enforces some TLS certificate requirements (such as only permitting
440 certificate signature algorithms contained in the supported algorithms
441 extension) which some implementations ignore: this option should be used
442 with caution as it could cause interoperability issues.
445 *) Update and tidy signature algorithm extension processing. Work out
446 shared signature algorithms based on preferences and peer algorithms
447 and print them out in s_client and s_server. Abort handshake if no
448 shared signature algorithms.
451 *) Add new functions to allow customised supported signature algorithms
452 for SSL and SSL_CTX structures. Add options to s_client and s_server
456 *) New function SSL_certs_clear() to delete all references to certificates
457 from an SSL structure. Before this once a certificate had been added
458 it couldn't be removed.
461 *) Integrate hostname, email address and IP address checking with certificate
462 verification. New verify options supporting checking in opensl utility.
465 *) Fixes and wildcard matching support to hostname and email checking
466 functions. Add manual page.
467 [Florian Weimer (Red Hat Product Security Team)]
469 *) New functions to check a hostname email or IP address against a
470 certificate. Add options x509 utility to print results of checks against
474 *) Fix OCSP checking.
475 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
477 *) Initial experimental support for explicitly trusted non-root CAs.
478 OpenSSL still tries to build a complete chain to a root but if an
479 intermediate CA has a trust setting included that is used. The first
480 setting is used: whether to trust (e.g., -addtrust option to the x509
484 *) Add -trusted_first option which attempts to find certificates in the
485 trusted store even if an untrusted chain is also supplied.
488 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
489 platform support for Linux and Android.
492 *) Support for linux-x32, ILP32 environment in x86_64 framework.
495 *) RFC 5878 (TLS Authorization Extensions) support.
496 [Emilia Kasper, Adam Langley, Ben Laurie (Google)]
498 *) Experimental multi-implementation support for FIPS capable OpenSSL.
499 When in FIPS mode the approved implementations are used as normal,
500 when not in FIPS mode the internal unapproved versions are used instead.
501 This means that the FIPS capable OpenSSL isn't forced to use the
502 (often lower perfomance) FIPS implementations outside FIPS mode.
505 *) Transparently support X9.42 DH parameters when calling
506 PEM_read_bio_DHparameters. This means existing applications can handle
507 the new parameter format automatically.
510 *) Initial experimental support for X9.42 DH parameter format: mainly
511 to support use of 'q' parameter for RFC5114 parameters.
514 *) Add DH parameters from RFC5114 including test data to dhtest.
517 *) Support for automatic EC temporary key parameter selection. If enabled
518 the most preferred EC parameters are automatically used instead of
519 hardcoded fixed parameters. Now a server just has to call:
520 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
521 support ECDH and use the most appropriate parameters.
524 *) Enhance and tidy EC curve and point format TLS extension code. Use
525 static structures instead of allocation if default values are used.
526 New ctrls to set curves we wish to support and to retrieve shared curves.
527 Print out shared curves in s_server. New options to s_server and s_client
528 to set list of supported curves.
531 *) New ctrls to retrieve supported signature algorithms and
532 supported curve values as an array of NIDs. Extend openssl utility
533 to print out received values.
536 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
537 between NIDs and the more common NIST names such as "P-256". Enhance
538 ecparam utility and ECC method to recognise the NIST names for curves.
541 *) Enhance SSL/TLS certificate chain handling to support different
542 chains for each certificate instead of one chain in the parent SSL_CTX.
545 *) Support for fixed DH ciphersuite client authentication: where both
546 server and client use DH certificates with common parameters.
549 *) Support for fixed DH ciphersuites: those requiring DH server
553 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
555 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
556 supporting platforms or when small records were transferred.
557 [Andy Polyakov, Steve Henson]
559 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
561 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
563 This addresses the flaw in CBC record processing discovered by
564 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
565 at: http://www.isg.rhul.ac.uk/tls/
567 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
568 Security Group at Royal Holloway, University of London
569 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
570 Emilia Käsper for the initial patch.
572 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
574 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
575 ciphersuites which can be exploited in a denial of service attack.
576 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
577 and detecting this bug and to Wolfgang Ettlinger
578 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
582 *) Return an error when checking OCSP signatures when key is NULL.
583 This fixes a DoS attack. (CVE-2013-0166)
586 *) Make openssl verify return errors.
587 [Chris Palmer <palmer@google.com> and Ben Laurie]
589 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
590 the right response is stapled. Also change SSL_get_certificate()
591 so it returns the certificate actually sent.
592 See http://rt.openssl.org/Ticket/Display.html?id=2836.
593 [Rob Stradling <rob.stradling@comodo.com>]
595 *) Fix possible deadlock when decoding public keys.
598 *) Don't use TLS 1.0 record version number in initial client hello
602 Changes between 1.0.1b and 1.0.1c [10 May 2012]
604 *) Sanity check record length before skipping explicit IV in TLS
605 1.2, 1.1 and DTLS to fix DoS attack.
607 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
608 fuzzing as a service testing platform.
612 *) Initialise tkeylen properly when encrypting CMS messages.
613 Thanks to Solar Designer of Openwall for reporting this issue.
616 *) In FIPS mode don't try to use composite ciphers as they are not
620 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
622 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
623 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
624 mean any application compiled against OpenSSL 1.0.0 headers setting
625 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
626 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
627 0x10000000L Any application which was previously compiled against
628 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
629 will need to be recompiled as a result. Letting be results in
630 inability to disable specifically TLS 1.1 and in client context,
631 in unlike event, limit maximum offered version to TLS 1.0 [see below].
634 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
635 disable just protocol X, but all protocols above X *if* there are
636 protocols *below* X still enabled. In more practical terms it means
637 that if application wants to disable TLS1.0 in favor of TLS1.1 and
638 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
639 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
643 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
645 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
646 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
647 in CRYPTO_realloc_clean.
649 Thanks to Tavis Ormandy, Google Security Team, for discovering this
650 issue and to Adam Langley <agl@chromium.org> for fixing it.
652 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
654 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
657 *) Workarounds for some broken servers that "hang" if a client hello
658 record length exceeds 255 bytes.
660 1. Do not use record version number > TLS 1.0 in initial client
661 hello: some (but not all) hanging servers will now work.
662 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
663 the number of ciphers sent in the client hello. This should be
664 set to an even number, such as 50, for example by passing:
665 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
666 Most broken servers should now work.
667 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
668 TLS 1.2 client support entirely.
671 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
674 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
676 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
677 STRING form instead of a DigestInfo.
680 *) The format used for MDC2 RSA signatures is inconsistent between EVP
681 and the RSA_sign/RSA_verify functions. This was made more apparent when
682 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
683 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
684 the correct format in RSA_verify so both forms transparently work.
687 *) Some servers which support TLS 1.0 can choke if we initially indicate
688 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
689 encrypted premaster secret. As a workaround use the maximum pemitted
690 client version in client hello, this should keep such servers happy
691 and still work with previous versions of OpenSSL.
694 *) Add support for TLS/DTLS heartbeats.
695 [Robin Seggelmann <seggelmann@fh-muenster.de>]
697 *) Add support for SCTP.
698 [Robin Seggelmann <seggelmann@fh-muenster.de>]
700 *) Improved PRNG seeding for VOS.
701 [Paul Green <Paul.Green@stratus.com>]
703 *) Extensive assembler packs updates, most notably:
705 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
706 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
707 - x86_64: bit-sliced AES implementation;
708 - ARM: NEON support, contemporary platforms optimizations;
709 - s390x: z196 support;
710 - *: GHASH and GF(2^m) multiplication implementations;
714 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
715 (removal of unnecessary code)
716 [Peter Sylvester <peter.sylvester@edelweb.fr>]
718 *) Add TLS key material exporter from RFC 5705.
721 *) Add DTLS-SRTP negotiation from RFC 5764.
724 *) Add Next Protocol Negotiation,
725 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
726 disabled with a no-npn flag to config or Configure. Code donated
728 [Adam Langley <agl@google.com> and Ben Laurie]
730 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
731 NIST-P256, NIST-P521, with constant-time single point multiplication on
732 typical inputs. Compiler support for the nonstandard type __uint128_t is
733 required to use this (present in gcc 4.4 and later, for 64-bit builds).
734 Code made available under Apache License version 2.0.
736 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
737 line to include this in your build of OpenSSL, and run "make depend" (or
738 "make update"). This enables the following EC_METHODs:
740 EC_GFp_nistp224_method()
741 EC_GFp_nistp256_method()
742 EC_GFp_nistp521_method()
744 EC_GROUP_new_by_curve_name() will automatically use these (while
745 EC_GROUP_new_curve_GFp() currently prefers the more flexible
747 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
749 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
750 all platforms. Move ssize_t definition from e_os.h to the public
751 header file e_os2.h as it now appears in public header file cms.h
754 *) New -sigopt option to the ca, req and x509 utilities. Additional
755 signature parameters can be passed using this option and in
759 *) Add RSA PSS signing function. This will generate and set the
760 appropriate AlgorithmIdentifiers for PSS based on those in the
761 corresponding EVP_MD_CTX structure. No application support yet.
764 *) Support for companion algorithm specific ASN1 signing routines.
765 New function ASN1_item_sign_ctx() signs a pre-initialised
766 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
767 the appropriate parameters.
770 *) Add new algorithm specific ASN1 verification initialisation function
771 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
772 handling will be the same no matter what EVP_PKEY_METHOD is used.
773 Add a PSS handler to support verification of PSS signatures: checked
774 against a number of sample certificates.
777 *) Add signature printing for PSS. Add PSS OIDs.
778 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
780 *) Add algorithm specific signature printing. An individual ASN1 method
781 can now print out signatures instead of the standard hex dump.
783 More complex signatures (e.g. PSS) can print out more meaningful
784 information. Include DSA version that prints out the signature
788 *) Password based recipient info support for CMS library: implementing
792 *) Split password based encryption into PBES2 and PBKDF2 functions. This
793 neatly separates the code into cipher and PBE sections and is required
794 for some algorithms that split PBES2 into separate pieces (such as
798 *) Session-handling fixes:
799 - Fix handling of connections that are resuming with a session ID,
800 but also support Session Tickets.
801 - Fix a bug that suppressed issuing of a new ticket if the client
802 presented a ticket with an expired session.
803 - Try to set the ticket lifetime hint to something reasonable.
804 - Make tickets shorter by excluding irrelevant information.
805 - On the client side, don't ignore renewed tickets.
806 [Adam Langley, Bodo Moeller (Google)]
808 *) Fix PSK session representation.
811 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
813 This work was sponsored by Intel.
816 *) Add GCM support to TLS library. Some custom code is needed to split
817 the IV between the fixed (from PRF) and explicit (from TLS record)
818 portions. This adds all GCM ciphersuites supported by RFC5288 and
819 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
820 add a special AESGCM string for GCM only.
823 *) Expand range of ctrls for AES GCM. Permit setting invocation
824 field on decrypt and retrieval of invocation field only on encrypt.
827 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
828 As required by RFC5289 these ciphersuites cannot be used if for
829 versions of TLS earlier than 1.2.
832 *) For FIPS capable OpenSSL interpret a NULL default public key method
833 as unset and return the appopriate default but do *not* set the default.
834 This means we can return the appopriate method in applications that
835 swicth between FIPS and non-FIPS modes.
838 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
839 ENGINE is used then we cannot handle that in the FIPS module so we
840 keep original code iff non-FIPS operations are allowed.
843 *) Add -attime option to openssl utilities.
844 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
846 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
849 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
850 FIPS EC methods unconditionally for now.
853 *) New build option no-ec2m to disable characteristic 2 code.
856 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
857 all cases can be covered as some introduce binary incompatibilities.
860 *) Redirect RSA operations to FIPS module including keygen,
861 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
864 *) Add similar low level API blocking to ciphers.
867 *) Low level digest APIs are not approved in FIPS mode: any attempt
868 to use these will cause a fatal error. Applications that *really* want
869 to use them can use the private_* version instead.
872 *) Redirect cipher operations to FIPS module for FIPS builds.
875 *) Redirect digest operations to FIPS module for FIPS builds.
878 *) Update build system to add "fips" flag which will link in fipscanister.o
879 for static and shared library builds embedding a signature if needed.
882 *) Output TLS supported curves in preference order instead of numerical
883 order. This is currently hardcoded for the highest order curves first.
884 This should be configurable so applications can judge speed vs strength.
887 *) Add TLS v1.2 server support for client authentication.
890 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
894 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
895 FIPS modules versions.
898 *) Add TLS v1.2 client side support for client authentication. Keep cache
899 of handshake records longer as we don't know the hash algorithm to use
900 until after the certificate request message is received.
903 *) Initial TLS v1.2 client support. Add a default signature algorithms
904 extension including all the algorithms we support. Parse new signature
905 format in client key exchange. Relax some ECC signing restrictions for
906 TLS v1.2 as indicated in RFC5246.
909 *) Add server support for TLS v1.2 signature algorithms extension. Switch
910 to new signature format when needed using client digest preference.
911 All server ciphersuites should now work correctly in TLS v1.2. No client
912 support yet and no support for client certificates.
915 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
916 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
917 ciphersuites. At present only RSA key exchange ciphersuites work with
918 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
919 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
920 and version checking.
923 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
924 with this defined it will not be affected by any changes to ssl internal
925 structures. Add several utility functions to allow openssl application
926 to work with OPENSSL_NO_SSL_INTERN defined.
930 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
932 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
935 *) Permit abbreviated handshakes when renegotiating using the function
936 SSL_renegotiate_abbreviated().
937 [Robin Seggelmann <seggelmann@fh-muenster.de>]
939 *) Add call to ENGINE_register_all_complete() to
940 ENGINE_load_builtin_engines(), so some implementations get used
941 automatically instead of needing explicit application support.
944 *) Add support for TLS key exporter as described in RFC5705.
945 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
947 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
948 a few changes are required:
950 Add SSL_OP_NO_TLSv1_1 flag.
952 Update version checking logic to handle version 1.1.
953 Add explicit IV handling (ported from DTLS code).
954 Add command line options to s_client/s_server.
957 Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
959 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
961 This addresses the flaw in CBC record processing discovered by
962 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
963 at: http://www.isg.rhul.ac.uk/tls/
965 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
966 Security Group at Royal Holloway, University of London
967 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
968 Emilia Käsper for the initial patch.
970 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
972 *) Return an error when checking OCSP signatures when key is NULL.
973 This fixes a DoS attack. (CVE-2013-0166)
976 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
977 the right response is stapled. Also change SSL_get_certificate()
978 so it returns the certificate actually sent.
979 See http://rt.openssl.org/Ticket/Display.html?id=2836.
981 [Rob Stradling <rob.stradling@comodo.com>]
983 *) Fix possible deadlock when decoding public keys.
986 Changes between 1.0.0i and 1.0.0j [10 May 2012]
988 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
991 *) Sanity check record length before skipping explicit IV in DTLS
994 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
995 fuzzing as a service testing platform.
999 *) Initialise tkeylen properly when encrypting CMS messages.
1000 Thanks to Solar Designer of Openwall for reporting this issue.
1003 Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
1005 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1006 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1007 in CRYPTO_realloc_clean.
1009 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1010 issue and to Adam Langley <agl@chromium.org> for fixing it.
1012 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1014 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
1016 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1017 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1018 content decryption and always return the same error. Note: this attack
1019 needs on average 2^20 messages so it only affects automated senders. The
1020 old behaviour can be reenabled in the CMS code by setting the
1021 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1022 an MMA defence is not necessary.
1023 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1024 this issue. (CVE-2012-0884)
1027 *) Fix CVE-2011-4619: make sure we really are receiving a
1028 client hello before rejecting multiple SGC restarts. Thanks to
1029 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1032 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
1034 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1035 Thanks to Antonio Martin, Enterprise Secure Access Research and
1036 Development, Cisco Systems, Inc. for discovering this bug and
1037 preparing a fix. (CVE-2012-0050)
1040 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
1042 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1043 of the Vaudenay padding oracle attack on CBC mode encryption
1044 which enables an efficient plaintext recovery attack against
1045 the OpenSSL implementation of DTLS. Their attack exploits timing
1046 differences arising during decryption processing. A research
1047 paper describing this attack can be found at:
1048 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1049 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1050 Security Group at Royal Holloway, University of London
1051 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1052 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1053 for preparing the fix. (CVE-2011-4108)
1054 [Robin Seggelmann, Michael Tuexen]
1056 *) Clear bytes used for block padding of SSL 3.0 records.
1058 [Adam Langley (Google)]
1060 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1061 Kadianakis <desnacked@gmail.com> for discovering this issue and
1062 Adam Langley for preparing the fix. (CVE-2011-4619)
1063 [Adam Langley (Google)]
1065 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
1066 [Andrey Kulikov <amdeich@gmail.com>]
1068 *) Prevent malformed RFC3779 data triggering an assertion failure.
1069 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1070 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1071 [Rob Austein <sra@hactrn.net>]
1073 *) Improved PRNG seeding for VOS.
1074 [Paul Green <Paul.Green@stratus.com>]
1076 *) Fix ssl_ciph.c set-up race.
1077 [Adam Langley (Google)]
1079 *) Fix spurious failures in ecdsatest.c.
1080 [Emilia Käsper (Google)]
1082 *) Fix the BIO_f_buffer() implementation (which was mixing different
1083 interpretations of the '..._len' fields).
1084 [Adam Langley (Google)]
1086 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1087 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1088 threads won't reuse the same blinding coefficients.
1090 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1091 lock to call BN_BLINDING_invert_ex, and avoids one use of
1092 BN_BLINDING_update for each BN_BLINDING structure (previously,
1093 the last update always remained unused).
1094 [Emilia Käsper (Google)]
1096 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
1097 [Bob Buckholz (Google)]
1099 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
1101 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
1102 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
1103 [Kaspar Brand <ossl@velox.ch>]
1105 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1106 for multi-threaded use of ECDH. (CVE-2011-3210)
1107 [Adam Langley (Google)]
1109 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1112 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
1113 signature public key algorithm by using OID xref utilities instead.
1114 Before this you could only use some ECC ciphersuites with SHA1 only.
1117 *) Add protection against ECDSA timing attacks as mentioned in the paper
1118 by Billy Bob Brumley and Nicola Tuveri, see:
1120 http://eprint.iacr.org/2011/232.pdf
1122 [Billy Bob Brumley and Nicola Tuveri]
1124 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
1126 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1127 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1129 *) Fix bug in string printing code: if *any* escaping is enabled we must
1130 escape the escape character (backslash) or the resulting string is
1134 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
1136 *) Disable code workaround for ancient and obsolete Netscape browsers
1137 and servers: an attacker can use it in a ciphersuite downgrade attack.
1138 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1141 *) Fixed J-PAKE implementation error, originally discovered by
1142 Sebastien Martini, further info and confirmation from Stefan
1143 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1146 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
1148 *) Fix extension code to avoid race conditions which can result in a buffer
1149 overrun vulnerability: resumed sessions must not be modified as they can
1150 be shared by multiple threads. CVE-2010-3864
1153 *) Fix WIN32 build system to correctly link an ENGINE directory into
1157 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
1159 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
1161 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
1163 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
1165 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
1166 context. The operation can be customised via the ctrl mechanism in
1167 case ENGINEs want to include additional functionality.
1170 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
1173 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
1174 output hashes compatible with older versions of OpenSSL.
1175 [Willy Weisz <weisz@vcpc.univie.ac.at>]
1177 *) Fix compression algorithm handling: if resuming a session use the
1178 compression algorithm of the resumed session instead of determining
1179 it from client hello again. Don't allow server to change algorithm.
1182 *) Add load_crls() function to apps tidying load_certs() too. Add option
1183 to verify utility to allow additional CRLs to be included.
1186 *) Update OCSP request code to permit adding custom headers to the request:
1187 some responders need this.
1190 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
1192 [Julia Lawall <julia@diku.dk>]
1194 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
1195 needlessly dereferenced structures, used obsolete functions and
1196 didn't handle all updated verify codes correctly.
1199 *) Disable MD2 in the default configuration.
1202 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
1203 indicate the initial BIO being pushed or popped. This makes it possible
1204 to determine whether the BIO is the one explicitly called or as a result
1205 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
1206 it handles reference counts correctly and doesn't zero out the I/O bio
1207 when it is not being explicitly popped. WARNING: applications which
1208 included workarounds for the old buggy behaviour will need to be modified
1209 or they could free up already freed BIOs.
1212 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
1213 renaming to all platforms (within the 0.9.8 branch, this was
1214 done conditionally on Netware platforms to avoid a name clash).
1215 [Guenter <lists@gknw.net>]
1217 *) Add ECDHE and PSK support to DTLS.
1218 [Michael Tuexen <tuexen@fh-muenster.de>]
1220 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
1224 *) Add "missing" function EVP_MD_flags() (without this the only way to
1225 retrieve a digest flags is by accessing the structure directly. Update
1226 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
1227 or cipher is registered as in the "from" argument. Print out all
1228 registered digests in the dgst usage message instead of manually
1229 attempting to work them out.
1232 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
1233 this allows the use of compression and extensions. Change default cipher
1234 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
1235 by default unless an application cipher string requests it.
1238 *) Alter match criteria in PKCS12_parse(). It used to try to use local
1239 key ids to find matching certificates and keys but some PKCS#12 files
1240 don't follow the (somewhat unwritten) rules and this strategy fails.
1241 Now just gather all certificates together and the first private key
1242 then look for the first certificate that matches the key.
1245 *) Support use of registered digest and cipher names for dgst and cipher
1246 commands instead of having to add each one as a special case. So now
1253 openssl dgst -sha256 foo
1255 and this works for ENGINE based algorithms too.
1259 *) Update Gost ENGINE to support parameter files.
1260 [Victor B. Wagner <vitus@cryptocom.ru>]
1262 *) Support GeneralizedTime in ca utility.
1263 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
1265 *) Enhance the hash format used for certificate directory links. The new
1266 form uses the canonical encoding (meaning equivalent names will work
1267 even if they aren't identical) and uses SHA1 instead of MD5. This form
1268 is incompatible with the older format and as a result c_rehash should
1269 be used to rebuild symbolic links.
1272 *) Make PKCS#8 the default write format for private keys, replacing the
1273 traditional format. This form is standardised, more secure and doesn't
1274 include an implicit MD5 dependency.
1277 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
1278 committed to OpenSSL should pass this lot as a minimum.
1281 *) Add session ticket override functionality for use by EAP-FAST.
1282 [Jouni Malinen <j@w1.fi>]
1284 *) Modify HMAC functions to return a value. Since these can be implemented
1285 in an ENGINE errors can occur.
1288 *) Type-checked OBJ_bsearch_ex.
1291 *) Type-checked OBJ_bsearch. Also some constification necessitated
1292 by type-checking. Still to come: TXT_DB, bsearch(?),
1293 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
1297 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
1298 seconds to a tm structure directly, instead of going through OS
1299 specific date routines. This avoids any issues with OS routines such
1300 as the year 2038 bug. New *_adj() functions for ASN1 time structures
1301 and X509_time_adj_ex() to cover the extended range. The existing
1302 X509_time_adj() is still usable and will no longer have any date issues.
1305 *) Delta CRL support. New use deltas option which will attempt to locate
1306 and search any appropriate delta CRLs available.
1308 This work was sponsored by Google.
1311 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
1312 code and add additional score elements. Validate alternate CRL paths
1313 as part of the CRL checking and indicate a new error "CRL path validation
1314 error" in this case. Applications wanting additional details can use
1315 the verify callback and check the new "parent" field. If this is not
1316 NULL CRL path validation is taking place. Existing applications wont
1317 see this because it requires extended CRL support which is off by
1320 This work was sponsored by Google.
1323 *) Support for freshest CRL extension.
1325 This work was sponsored by Google.
1328 *) Initial indirect CRL support. Currently only supported in the CRLs
1329 passed directly and not via lookup. Process certificate issuer
1330 CRL entry extension and lookup CRL entries by bother issuer name
1331 and serial number. Check and process CRL issuer entry in IDP extension.
1333 This work was sponsored by Google.
1336 *) Add support for distinct certificate and CRL paths. The CRL issuer
1337 certificate is validated separately in this case. Only enabled if
1338 an extended CRL support flag is set: this flag will enable additional
1339 CRL functionality in future.
1341 This work was sponsored by Google.
1344 *) Add support for policy mappings extension.
1346 This work was sponsored by Google.
1349 *) Fixes to pathlength constraint, self issued certificate handling,
1350 policy processing to align with RFC3280 and PKITS tests.
1352 This work was sponsored by Google.
1355 *) Support for name constraints certificate extension. DN, email, DNS
1356 and URI types are currently supported.
1358 This work was sponsored by Google.
1361 *) To cater for systems that provide a pointer-based thread ID rather
1362 than numeric, deprecate the current numeric thread ID mechanism and
1363 replace it with a structure and associated callback type. This
1364 mechanism allows a numeric "hash" to be extracted from a thread ID in
1365 either case, and on platforms where pointers are larger than 'long',
1366 mixing is done to help ensure the numeric 'hash' is usable even if it
1367 can't be guaranteed unique. The default mechanism is to use "&errno"
1368 as a pointer-based thread ID to distinguish between threads.
1370 Applications that want to provide their own thread IDs should now use
1371 CRYPTO_THREADID_set_callback() to register a callback that will call
1372 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
1374 Note that ERR_remove_state() is now deprecated, because it is tied
1375 to the assumption that thread IDs are numeric. ERR_remove_state(0)
1376 to free the current thread's error state should be replaced by
1377 ERR_remove_thread_state(NULL).
1379 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
1380 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
1381 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
1382 application was previously providing a numeric thread callback that
1383 was inappropriate for distinguishing threads, then uniqueness might
1384 have been obtained with &errno that happened immediately in the
1385 intermediate development versions of OpenSSL; this is no longer the
1386 case, the numeric thread callback will now override the automatic use
1388 [Geoff Thorpe, with help from Bodo Moeller]
1390 *) Initial support for different CRL issuing certificates. This covers a
1391 simple case where the self issued certificates in the chain exist and
1392 the real CRL issuer is higher in the existing chain.
1394 This work was sponsored by Google.
1397 *) Removed effectively defunct crypto/store from the build.
1400 *) Revamp of STACK to provide stronger type-checking. Still to come:
1401 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
1402 ASN1_STRING, CONF_VALUE.
1405 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
1406 RAM on SSL connections. This option can save about 34k per idle SSL.
1409 *) Revamp of LHASH to provide stronger type-checking. Still to come:
1410 STACK, TXT_DB, bsearch, qsort.
1413 *) Initial support for Cryptographic Message Syntax (aka CMS) based
1414 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
1415 support for data, signedData, compressedData, digestedData and
1416 encryptedData, envelopedData types included. Scripts to check against
1417 RFC4134 examples draft and interop and consistency checks of many
1418 content types and variants.
1421 *) Add options to enc utility to support use of zlib compression BIO.
1424 *) Extend mk1mf to support importing of options and assembly language
1425 files from Configure script, currently only included in VC-WIN32.
1426 The assembly language rules can now optionally generate the source
1427 files from the associated perl scripts.
1430 *) Implement remaining functionality needed to support GOST ciphersuites.
1431 Interop testing has been performed using CryptoPro implementations.
1432 [Victor B. Wagner <vitus@cryptocom.ru>]
1434 *) s390x assembler pack.
1437 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
1441 *) Implement Opaque PRF Input TLS extension as specified in
1442 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
1443 official specification yet and no extension type assignment by
1444 IANA exists, this extension (for now) will have to be explicitly
1445 enabled when building OpenSSL by providing the extension number
1446 to use. For example, specify an option
1448 -DTLSEXT_TYPE_opaque_prf_input=0x9527
1450 to the "config" or "Configure" script to enable the extension,
1451 assuming extension number 0x9527 (which is a completely arbitrary
1452 and unofficial assignment based on the MD5 hash of the Internet
1453 Draft). Note that by doing so, you potentially lose
1454 interoperability with other TLS implementations since these might
1455 be using the same extension number for other purposes.
1457 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
1458 opaque PRF input value to use in the handshake. This will create
1459 an interal copy of the length-'len' string at 'src', and will
1460 return non-zero for success.
1462 To get more control and flexibility, provide a callback function
1465 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
1466 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
1470 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
1473 Callback function 'cb' will be called in handshakes, and is
1474 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
1475 Argument 'arg' is for application purposes (the value as given to
1476 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
1477 be provided to the callback function). The callback function
1478 has to return non-zero to report success: usually 1 to use opaque
1479 PRF input just if possible, or 2 to enforce use of the opaque PRF
1480 input. In the latter case, the library will abort the handshake
1481 if opaque PRF input is not successfully negotiated.
1483 Arguments 'peerinput' and 'len' given to the callback function
1484 will always be NULL and 0 in the case of a client. A server will
1485 see the client's opaque PRF input through these variables if
1486 available (NULL and 0 otherwise). Note that if the server
1487 provides an opaque PRF input, the length must be the same as the
1488 length of the client's opaque PRF input.
1490 Note that the callback function will only be called when creating
1491 a new session (session resumption can resume whatever was
1492 previously negotiated), and will not be called in SSL 2.0
1493 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
1494 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
1495 for applications that need to enforce opaque PRF input.
1499 *) Update ssl code to support digests other than SHA1+MD5 for handshake
1502 [Victor B. Wagner <vitus@cryptocom.ru>]
1504 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1505 RFC4507bis. The encrypted ticket format is an encrypted encoded
1506 SSL_SESSION structure, that way new session features are automatically
1509 If a client application caches session in an SSL_SESSION structure
1510 support is transparent because tickets are now stored in the encoded
1513 The SSL_CTX structure automatically generates keys for ticket
1514 protection in servers so again support should be possible
1515 with no application modification.
1517 If a client or server wishes to disable RFC4507 support then the option
1518 SSL_OP_NO_TICKET can be set.
1520 Add a TLS extension debugging callback to allow the contents of any client
1521 or server extensions to be examined.
1523 This work was sponsored by Google.
1526 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1527 OpenSSL should now compile cleanly on gcc 4.2
1528 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1530 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1531 support including streaming MAC support: this is required for GOST
1532 ciphersuite support.
1533 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1535 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1536 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1537 to output in BER and PEM format.
1540 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1541 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1542 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1543 ENGINE support for HMAC keys which are unextractable. New -mac and
1544 -macopt options to dgst utility.
1547 *) New option -sigopt to dgst utility. Update dgst to use
1548 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1549 alternative signing paramaters such as X9.31 or PSS in the dgst
1553 *) Change ssl_cipher_apply_rule(), the internal function that does
1554 the work each time a ciphersuite string requests enabling
1555 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1556 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1557 the order of disabled ciphersuites such that those ciphersuites
1558 that most recently went from enabled to disabled not only stay
1559 in order with respect to each other, but also have higher priority
1560 than other disabled ciphersuites the next time ciphersuites are
1563 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1564 the same ciphersuites as with "HIGH" alone, but in a specific
1565 order where the PSK ciphersuites come first (since they are the
1566 most recently disabled ciphersuites when "HIGH" is parsed).
1568 Also, change ssl_create_cipher_list() (using this new
1569 funcionality) such that between otherwise identical
1570 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
1574 *) Change ssl_create_cipher_list() so that it automatically
1575 arranges the ciphersuites in reasonable order before starting
1576 to process the rule string. Thus, the definition for "DEFAULT"
1577 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
1578 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
1579 This makes it much easier to arrive at a reasonable default order
1580 in applications for which anonymous ciphers are OK (meaning
1581 that you can't actually use DEFAULT).
1582 [Bodo Moeller; suggested by Victor Duchovni]
1584 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
1585 processing) into multiple integers instead of setting
1586 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
1587 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
1588 (These masks as well as the individual bit definitions are hidden
1589 away into the non-exported interface ssl/ssl_locl.h, so this
1590 change to the definition of the SSL_CIPHER structure shouldn't
1591 affect applications.) This give us more bits for each of these
1592 categories, so there is no longer a need to coagulate AES128 and
1593 AES256 into a single algorithm bit, and to coagulate Camellia128
1594 and Camellia256 into a single algorithm bit, which has led to all
1597 Thus, among other things, the kludge introduced in 0.9.7m and
1598 0.9.8e for masking out AES256 independently of AES128 or masking
1599 out Camellia256 independently of AES256 is not needed here in 0.9.9.
1601 With the change, we also introduce new ciphersuite aliases that
1602 so far were missing: "AES128", "AES256", "CAMELLIA128", and
1606 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
1607 Use the leftmost N bytes of the signature input if the input is
1608 larger than the prime q (with N being the size in bytes of q).
1611 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1612 it yet and it is largely untested.
1615 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1618 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1619 some compilers (gcc 4.2 and later) reject their use. Safestack is
1620 reimplemented. Update ASN1 to avoid use of legacy functions.
1623 *) Win32/64 targets are linked with Winsock2.
1626 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1627 to external functions. This can be used to increase CRL handling
1628 efficiency especially when CRLs are very large by (for example) storing
1629 the CRL revoked certificates in a database.
1632 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1633 new CRLs added to a directory can be used. New command line option
1634 -verify_return_error to s_client and s_server. This causes real errors
1635 to be returned by the verify callback instead of carrying on no matter
1636 what. This reflects the way a "real world" verify callback would behave.
1639 *) GOST engine, supporting several GOST algorithms and public key formats.
1640 Kindly donated by Cryptocom.
1643 *) Partial support for Issuing Distribution Point CRL extension. CRLs
1644 partitioned by DP are handled but no indirect CRL or reason partitioning
1645 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
1646 selected via a scoring technique which handles IDP and AKID in CRLs.
1649 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1650 will ultimately be used for all verify operations: this will remove the
1651 X509_STORE dependency on certificate verification and allow alternative
1652 lookup methods. X509_STORE based implementations of these two callbacks.
1655 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1656 Modify get_crl() to find a valid (unexpired) CRL if possible.
1659 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1660 this would be called X509_CRL_cmp() but that name is already used by
1661 a function that just compares CRL issuer names. Cache several CRL
1662 extensions in X509_CRL structure and cache CRLDP in X509.
1665 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1666 this maps equivalent X509_NAME structures into a consistent structure.
1667 Name comparison can then be performed rapidly using memcmp().
1670 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1674 *) Allow digests to supply their own micalg string for S/MIME type using
1675 the ctrl EVP_MD_CTRL_MICALG.
1678 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1679 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1680 ctrl. It can then customise the structure before and/or after signing
1684 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1685 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1686 to free up any added signature OIDs.
1689 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1690 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1691 digest and cipher tables. New options added to openssl utility:
1692 list-message-digest-algorithms and list-cipher-algorithms.
1695 *) Change the array representation of binary polynomials: the list
1696 of degrees of non-zero coefficients is now terminated with -1.
1697 Previously it was terminated with 0, which was also part of the
1698 value; thus, the array representation was not applicable to
1699 polynomials where t^0 has coefficient zero. This change makes
1700 the array representation useful in a more general context.
1703 *) Various modifications and fixes to SSL/TLS cipher string
1704 handling. For ECC, the code now distinguishes between fixed ECDH
1705 with RSA certificates on the one hand and with ECDSA certificates
1706 on the other hand, since these are separate ciphersuites. The
1707 unused code for Fortezza ciphersuites has been removed.
1709 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1710 (not "ECDHE"). For consistency with the code for DH
1711 certificates, use of ECDH certificates is now considered ECDH
1712 authentication, not RSA or ECDSA authentication (the latter is
1713 merely the CA's signing algorithm and not actively used in the
1716 The temporary ciphersuite alias "ECCdraft" is no longer
1717 available, and ECC ciphersuites are no longer excluded from "ALL"
1718 and "DEFAULT". The following aliases now exist for RFC 4492
1719 ciphersuites, most of these by analogy with the DH case:
1721 kECDHr - ECDH cert, signed with RSA
1722 kECDHe - ECDH cert, signed with ECDSA
1723 kECDH - ECDH cert (signed with either RSA or ECDSA)
1724 kEECDH - ephemeral ECDH
1725 ECDH - ECDH cert or ephemeral ECDH
1731 AECDH - anonymous ECDH
1732 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1736 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1737 Use correct micalg parameters depending on digest(s) in signed message.
1740 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1741 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1744 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1745 an engine to register a method. Add ENGINE lookups for methods and
1746 functional reference processing.
1749 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1750 EVP_{Sign,Verify}* which allow an application to customise the signature
1754 *) New -resign option to smime utility. This adds one or more signers
1755 to an existing PKCS#7 signedData structure. Also -md option to use an
1756 alternative message digest algorithm for signing.
1759 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1760 create PKCS7 structures containing multiple signers. Update smime
1761 application to support multiple signers.
1764 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1768 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1769 Reorganize PBE internals to lookup from a static table using NIDs,
1770 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1771 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1772 PRF which will be automatically used with PBES2.
1775 *) Replace the algorithm specific calls to generate keys in "req" with the
1779 *) Update PKCS#7 enveloped data routines to use new API. This is now
1780 supported by any public key method supporting the encrypt operation. A
1781 ctrl is added to allow the public key algorithm to examine or modify
1782 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1786 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1787 a default digest type to use. In most cases this will be SHA1 but some
1788 algorithms (such as GOST) need to specify an alternative digest. The
1789 return value indicates how strong the prefernce is 1 means optional and
1790 2 is mandatory (that is it is the only supported type). Modify
1791 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1792 use the default md. Update openssl utilities to use the default digest
1793 type for signing if it is not explicitly indicated.
1796 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1797 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1798 signing method from the key type. This effectively removes the link
1799 between digests and public key types.
1802 *) Add an OID cross reference table and utility functions. Its purpose is to
1803 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1804 rsaEncryption. This will allow some of the algorithm specific hackery
1805 needed to use the correct OID to be removed.
1808 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1809 structures for PKCS7_sign(). They are now set up by the relevant public
1813 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1816 *) Add support for key derivation (agreement) in the API, DH method and
1820 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1821 public and private key formats. As a side effect these add additional
1822 command line functionality not previously available: DSA signatures can be
1823 generated and verified using pkeyutl and DH key support and generation in
1828 [Oliver Tappe <zooey@hirschkaefer.de>]
1830 *) New make target "install_html_docs" installs HTML renditions of the
1832 [Oliver Tappe <zooey@hirschkaefer.de>]
1834 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1835 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1836 support key and parameter generation and add initial key generation
1837 functionality for RSA.
1840 *) Add functions for main EVP_PKEY_method operations. The undocumented
1841 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1842 EVP_PKEY_{encrypt,decrypt}_old.
1845 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1846 key API, doesn't do much yet.
1849 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1850 public key algorithms. New option to openssl utility:
1851 "list-public-key-algorithms" to print out info.
1854 *) Implement the Supported Elliptic Curves Extension for
1855 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1858 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1859 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1862 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1863 utilities such as rsa, dsa, dsaparam etc except they process any key
1867 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1868 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1869 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1873 *) Initial support for pluggable public key ASN1.
1874 De-spaghettify the public key ASN1 handling. Move public and private
1875 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1876 algorithm specific handling to a single module within the relevant
1877 algorithm directory. Add functions to allow (near) opaque processing
1878 of public and private key structures.
1881 *) Implement the Supported Point Formats Extension for
1882 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1885 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1886 for the psk identity [hint] and the psk callback functions to the
1887 SSL_SESSION, SSL and SSL_CTX structure.
1890 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1894 SSL_CTX_use_psk_identity_hint
1895 SSL_get_psk_identity_hint
1896 SSL_get_psk_identity
1897 SSL_use_psk_identity_hint
1899 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1901 *) Add RFC 3161 compliant time stamp request creation, response generation
1902 and response verification functionality.
1903 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1905 *) Add initial support for TLS extensions, specifically for the server_name
1906 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1907 have new members for a host name. The SSL data structure has an
1908 additional member SSL_CTX *initial_ctx so that new sessions can be
1909 stored in that context to allow for session resumption, even after the
1910 SSL has been switched to a new SSL_CTX in reaction to a client's
1911 server_name extension.
1913 New functions (subject to change):
1915 SSL_get_servername()
1916 SSL_get_servername_type()
1919 New CTRL codes and macros (subject to change):
1921 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1922 - SSL_CTX_set_tlsext_servername_callback()
1923 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1924 - SSL_CTX_set_tlsext_servername_arg()
1925 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1927 openssl s_client has a new '-servername ...' option.
1929 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1930 '-key2 ...', '-servername_fatal' (subject to change). This allows
1931 testing the HostName extension for a specific single host name ('-cert'
1932 and '-key' remain fallbacks for handshakes without HostName
1933 negotiation). If the unrecogninzed_name alert has to be sent, this by
1934 default is a warning; it becomes fatal with the '-servername_fatal'
1937 [Peter Sylvester, Remy Allais, Christophe Renou]
1939 *) Whirlpool hash implementation is added.
1942 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1943 bn(64,32). Because of instruction set limitations it doesn't have
1944 any negative impact on performance. This was done mostly in order
1945 to make it possible to share assembler modules, such as bn_mul_mont
1946 implementations, between 32- and 64-bit builds without hassle.
1949 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1950 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1954 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1955 dedicated Montgomery multiplication procedure, is introduced.
1956 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1957 "64-bit" performance on certain 32-bit targets.
1960 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1961 in SSL structures. New SSL ctrl to set maximum send fragment size.
1962 Save memory by seeting the I/O buffer sizes dynamically instead of
1963 using the maximum available value.
1966 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1967 in addition to the text details.
1970 *) Very, very preliminary EXPERIMENTAL support for printing of general
1971 ASN1 structures. This currently produces rather ugly output and doesn't
1972 handle several customised structures at all.
1975 *) Integrated support for PVK file format and some related formats such
1976 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1977 these in the 'rsa' and 'dsa' utilities.
1980 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1983 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1984 place for the (very old) "NETSCAPE" format certificates which are now
1985 handled using new ASN1 code equivalents.
1988 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1989 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1990 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1993 *) Modify CRL distribution points extension code to print out previously
1994 unsupported fields. Enhance extension setting code to allow setting of
1998 *) Add print and set support for Issuing Distribution Point CRL extension.
2001 *) Change 'Configure' script to enable Camellia by default.
2004 Changes between 0.9.8x and 0.9.8y [5 Feb 2013]
2006 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2008 This addresses the flaw in CBC record processing discovered by
2009 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2010 at: http://www.isg.rhul.ac.uk/tls/
2012 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2013 Security Group at Royal Holloway, University of London
2014 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2015 Emilia Käsper for the initial patch.
2017 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2019 *) Return an error when checking OCSP signatures when key is NULL.
2020 This fixes a DoS attack. (CVE-2013-0166)
2023 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2024 the right response is stapled. Also change SSL_get_certificate()
2025 so it returns the certificate actually sent.
2026 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2027 (This is a backport)
2028 [Rob Stradling <rob.stradling@comodo.com>]
2030 *) Fix possible deadlock when decoding public keys.
2033 Changes between 0.9.8w and 0.9.8x [10 May 2012]
2035 *) Sanity check record length before skipping explicit IV in DTLS
2038 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2039 fuzzing as a service testing platform.
2043 *) Initialise tkeylen properly when encrypting CMS messages.
2044 Thanks to Solar Designer of Openwall for reporting this issue.
2047 Changes between 0.9.8v and 0.9.8w [23 Apr 2012]
2049 *) The fix for CVE-2012-2110 did not take into account that the
2050 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an
2051 int in OpenSSL 0.9.8, making it still vulnerable. Fix by
2052 rejecting negative len parameter. (CVE-2012-2131)
2053 [Tomas Hoger <thoger@redhat.com>]
2055 Changes between 0.9.8u and 0.9.8v [19 Apr 2012]
2057 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2058 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2059 in CRYPTO_realloc_clean.
2061 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2062 issue and to Adam Langley <agl@chromium.org> for fixing it.
2064 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2066 Changes between 0.9.8t and 0.9.8u [12 Mar 2012]
2068 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2069 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2070 content decryption and always return the same error. Note: this attack
2071 needs on average 2^20 messages so it only affects automated senders. The
2072 old behaviour can be reenabled in the CMS code by setting the
2073 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2074 an MMA defence is not necessary.
2075 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2076 this issue. (CVE-2012-0884)
2079 *) Fix CVE-2011-4619: make sure we really are receiving a
2080 client hello before rejecting multiple SGC restarts. Thanks to
2081 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2084 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
2086 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2087 Thanks to Antonio Martin, Enterprise Secure Access Research and
2088 Development, Cisco Systems, Inc. for discovering this bug and
2089 preparing a fix. (CVE-2012-0050)
2092 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
2094 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2095 of the Vaudenay padding oracle attack on CBC mode encryption
2096 which enables an efficient plaintext recovery attack against
2097 the OpenSSL implementation of DTLS. Their attack exploits timing
2098 differences arising during decryption processing. A research
2099 paper describing this attack can be found at:
2100 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2101 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2102 Security Group at Royal Holloway, University of London
2103 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2104 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2105 for preparing the fix. (CVE-2011-4108)
2106 [Robin Seggelmann, Michael Tuexen]
2108 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
2109 [Ben Laurie, Kasper <ekasper@google.com>]
2111 *) Clear bytes used for block padding of SSL 3.0 records.
2113 [Adam Langley (Google)]
2115 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2116 Kadianakis <desnacked@gmail.com> for discovering this issue and
2117 Adam Langley for preparing the fix. (CVE-2011-4619)
2118 [Adam Langley (Google)]
2120 *) Prevent malformed RFC3779 data triggering an assertion failure.
2121 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2122 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2123 [Rob Austein <sra@hactrn.net>]
2125 *) Fix ssl_ciph.c set-up race.
2126 [Adam Langley (Google)]
2128 *) Fix spurious failures in ecdsatest.c.
2129 [Emilia Käsper (Google)]
2131 *) Fix the BIO_f_buffer() implementation (which was mixing different
2132 interpretations of the '..._len' fields).
2133 [Adam Langley (Google)]
2135 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2136 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2137 threads won't reuse the same blinding coefficients.
2139 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2140 lock to call BN_BLINDING_invert_ex, and avoids one use of
2141 BN_BLINDING_update for each BN_BLINDING structure (previously,
2142 the last update always remained unused).
2143 [Emilia Käsper (Google)]
2145 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2146 for multi-threaded use of ECDH.
2147 [Adam Langley (Google)]
2149 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2152 *) Add protection against ECDSA timing attacks as mentioned in the paper
2153 by Billy Bob Brumley and Nicola Tuveri, see:
2155 http://eprint.iacr.org/2011/232.pdf
2157 [Billy Bob Brumley and Nicola Tuveri]
2159 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
2161 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2162 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2164 *) Fix bug in string printing code: if *any* escaping is enabled we must
2165 escape the escape character (backslash) or the resulting string is
2169 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
2171 *) Disable code workaround for ancient and obsolete Netscape browsers
2172 and servers: an attacker can use it in a ciphersuite downgrade attack.
2173 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2176 *) Fixed J-PAKE implementation error, originally discovered by
2177 Sebastien Martini, further info and confirmation from Stefan
2178 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2181 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
2183 *) Fix extension code to avoid race conditions which can result in a buffer
2184 overrun vulnerability: resumed sessions must not be modified as they can
2185 be shared by multiple threads. CVE-2010-3864
2188 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
2191 *) Don't reencode certificate when calculating signature: cache and use
2192 the original encoding instead. This makes signature verification of
2193 some broken encodings work correctly.
2196 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
2197 is also one of the inputs.
2198 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
2200 *) Don't repeatedly append PBE algorithms to table if they already exist.
2201 Sort table on each new add. This effectively makes the table read only
2202 after all algorithms are added and subsequent calls to PKCS12_pbe_add
2206 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
2208 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
2211 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
2212 access or freeing data twice (CVE-2010-0742)
2213 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
2215 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
2216 common in certificates and some applications which only call
2217 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
2221 Reduce copying into .apps and .test in makevms.com
2222 Don't try to use blank CA certificate in CA.com
2223 Allow use of C files from original directories in maketests.com
2224 [Steven M. Schweda" <sms@antinode.info>]
2226 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2228 *) When rejecting SSL/TLS records due to an incorrect version number, never
2229 update s->server with a new major version number. As of
2230 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2231 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2232 the previous behavior could result in a read attempt at NULL when
2233 receiving specific incorrect SSL/TLS records once record payload
2234 protection is active. (CVE-2010-0740)
2235 [Bodo Moeller, Adam Langley <agl@chromium.org>]
2237 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
2238 could be crashed if the relevant tables were not present (e.g. chrooted).
2239 [Tomas Hoger <thoger@redhat.com>]
2241 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
2243 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
2244 [Martin Olsson, Neel Mehta]
2246 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
2247 accommodate for stack sorting, always a write lock!).
2250 *) On some versions of WIN32 Heap32Next is very slow. This can cause
2251 excessive delays in the RAND_poll(): over a minute. As a workaround
2252 include a time check in the inner Heap32Next loop too.
2255 *) The code that handled flushing of data in SSL/TLS originally used the
2256 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
2257 the problem outlined in PR#1949. The fix suggested there however can
2258 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
2259 of Apache). So instead simplify the code to flush unconditionally.
2260 This should be fine since flushing with no data to flush is a no op.
2263 *) Handle TLS versions 2.0 and later properly and correctly use the
2264 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
2265 off ancient servers have a habit of sticking around for a while...
2268 *) Modify compression code so it frees up structures without using the
2269 ex_data callbacks. This works around a problem where some applications
2270 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
2271 restarting) then use compression (e.g. SSL with compression) later.
2272 This results in significant per-connection memory leaks and
2273 has caused some security issues including CVE-2008-1678 and
2277 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
2278 change when encrypting or decrypting.
2281 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
2282 connect and renegotiate with servers which do not support RI.
2283 Until RI is more widely deployed this option is enabled by default.
2286 *) Add "missing" ssl ctrls to clear options and mode.
2289 *) If client attempts to renegotiate and doesn't support RI respond with
2290 a no_renegotiation alert as required by RFC5746. Some renegotiating
2291 TLS clients will continue a connection gracefully when they receive
2292 the alert. Unfortunately OpenSSL mishandled this alert and would hang
2293 waiting for a server hello which it will never receive. Now we treat a
2294 received no_renegotiation alert as a fatal error. This is because
2295 applications requesting a renegotiation might well expect it to succeed
2296 and would have no code in place to handle the server denying it so the
2297 only safe thing to do is to terminate the connection.
2300 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
2301 peer supports secure renegotiation and 0 otherwise. Print out peer
2302 renegotiation support in s_client/s_server.
2305 *) Replace the highly broken and deprecated SPKAC certification method with
2306 the updated NID creation version. This should correctly handle UTF8.
2309 *) Implement RFC5746. Re-enable renegotiation but require the extension
2310 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2311 turns out to be a bad idea. It has been replaced by
2312 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2313 SSL_CTX_set_options(). This is really not recommended unless you
2314 know what you are doing.
2315 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2317 *) Fixes to stateless session resumption handling. Use initial_ctx when
2318 issuing and attempting to decrypt tickets in case it has changed during
2319 servername handling. Use a non-zero length session ID when attempting
2320 stateless session resumption: this makes it possible to determine if
2321 a resumption has occurred immediately after receiving server hello
2322 (several places in OpenSSL subtly assume this) instead of later in
2326 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2327 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
2328 fixes for a few places where the return code is not checked
2330 [Julia Lawall <julia@diku.dk>]
2332 *) Add --strict-warnings option to Configure script to include devteam
2333 warnings in other configurations.
2336 *) Add support for --libdir option and LIBDIR variable in makefiles. This
2337 makes it possible to install openssl libraries in locations which
2338 have names other than "lib", for example "/usr/lib64" which some
2340 [Steve Henson, based on patch from Jeremy Utley]
2342 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
2343 X690 8.9.12 and can produce some misleading textual output of OIDs.
2344 [Steve Henson, reported by Dan Kaminsky]
2346 *) Delete MD2 from algorithm tables. This follows the recommendation in
2347 several standards that it is not used in new applications due to
2348 several cryptographic weaknesses. For binary compatibility reasons
2349 the MD2 API is still compiled in by default.
2352 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
2356 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
2357 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
2359 [Guenter <lists@gknw.net>]
2361 *) Fix the server certificate chain building code to use X509_verify_cert(),
2362 it used to have an ad-hoc builder which was unable to cope with anything
2363 other than a simple chain.
2364 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
2366 *) Don't check self signed certificate signatures in X509_verify_cert()
2367 by default (a flag can override this): it just wastes time without
2368 adding any security. As a useful side effect self signed root CAs
2369 with non-FIPS digests are now usable in FIPS mode.
2372 *) In dtls1_process_out_of_seq_message() the check if the current message
2373 is already buffered was missing. For every new message was memory
2374 allocated, allowing an attacker to perform an denial of service attack
2375 with sending out of seq handshake messages until there is no memory
2376 left. Additionally every future messege was buffered, even if the
2377 sequence number made no sense and would be part of another handshake.
2378 So only messages with sequence numbers less than 10 in advance will be
2379 buffered. (CVE-2009-1378)
2380 [Robin Seggelmann, discovered by Daniel Mentz]
2382 *) Records are buffered if they arrive with a future epoch to be
2383 processed after finishing the corresponding handshake. There is
2384 currently no limitation to this buffer allowing an attacker to perform
2385 a DOS attack with sending records with future epochs until there is no
2386 memory left. This patch adds the pqueue_size() function to detemine
2387 the size of a buffer and limits the record buffer to 100 entries.
2389 [Robin Seggelmann, discovered by Daniel Mentz]
2391 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2392 parent structure is freed. (CVE-2009-1379)
2395 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2396 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2399 [Ilya O. <vrghost@gmail.com>]
2401 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2403 *) Disable renegotiation completely - this fixes a severe security
2404 problem (CVE-2009-3555) at the cost of breaking all
2405 renegotiation. Renegotiation can be re-enabled by setting
2406 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2407 run-time. This is really not recommended unless you know what
2411 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2413 *) Don't set val to NULL when freeing up structures, it is freed up by
2414 underlying code. If sizeof(void *) > sizeof(long) this can result in
2415 zeroing past the valid field. (CVE-2009-0789)
2416 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2418 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2419 checked correctly. This would allow some invalid signed attributes to
2420 appear to verify correctly. (CVE-2009-0591)
2421 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2423 *) Reject UniversalString and BMPString types with invalid lengths. This
2424 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2425 a legal length. (CVE-2009-0590)
2428 *) Set S/MIME signing as the default purpose rather than setting it
2429 unconditionally. This allows applications to override it at the store
2433 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2434 to handle some structures.
2437 *) Improve efficiency of mem_gets: don't search whole buffer each time
2439 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2441 *) New -hex option for openssl rand.
2444 *) Print out UTF8String and NumericString when parsing ASN1.
2447 *) Support NumericString type for name components.
2450 *) Allow CC in the environment to override the automatically chosen
2451 compiler. Note that nothing is done to ensure flags work with the
2455 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2457 *) Properly check EVP_VerifyFinal() and similar return values
2459 [Ben Laurie, Bodo Moeller, Google Security Team]
2461 *) Enable TLS extensions by default.
2464 *) Allow the CHIL engine to be loaded, whether the application is
2465 multithreaded or not. (This does not release the developer from the
2466 obligation to set up the dynamic locking callbacks.)
2467 [Sander Temme <sander@temme.net>]
2469 *) Use correct exit code if there is an error in dgst command.
2470 [Steve Henson; problem pointed out by Roland Dirlewanger]
2472 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2473 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2476 *) Add experimental JPAKE support, including demo authentication in
2477 s_client and s_server.
2480 *) Set the comparison function in v3_addr_canonize().
2481 [Rob Austein <sra@hactrn.net>]
2483 *) Add support for XMPP STARTTLS in s_client.
2484 [Philip Paeps <philip@freebsd.org>]
2486 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2487 to ensure that even with this option, only ciphersuites in the
2488 server's preference list will be accepted. (Note that the option
2489 applies only when resuming a session, so the earlier behavior was
2490 just about the algorithm choice for symmetric cryptography.)
2493 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2495 *) Fix NULL pointer dereference if a DTLS server received
2496 ChangeCipherSpec as first record (CVE-2009-1386).
2499 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
2500 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2503 *) The fix in 0.9.8c that supposedly got rid of unsafe
2504 double-checked locking was incomplete for RSA blinding,
2505 addressing just one layer of what turns out to have been
2506 doubly unsafe triple-checked locking.
2508 So now fix this for real by retiring the MONT_HELPER macro
2509 in crypto/rsa/rsa_eay.c.
2511 [Bodo Moeller; problem pointed out by Marius Schilder]
2513 *) Various precautionary measures:
2515 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2517 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2518 (NB: This would require knowledge of the secret session ticket key
2519 to exploit, in which case you'd be SOL either way.)
2521 - Change bn_nist.c so that it will properly handle input BIGNUMs
2522 outside the expected range.
2524 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2527 [Neel Mehta, Bodo Moeller]
2529 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2530 the load fails. Useful for distros.
2531 [Ben Laurie and the FreeBSD team]
2533 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2536 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2539 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2541 This work was sponsored by Logica.
2544 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2545 keystores. Support for SSL/TLS client authentication too.
2546 Not compiled unless enable-capieng specified to Configure.
2548 This work was sponsored by Logica.
2551 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2552 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2553 attribute creation routines such as certifcate requests and PKCS#12
2557 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2559 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2560 handshake which could lead to a cilent crash as found using the
2561 Codenomicon TLS test suite (CVE-2008-1672)
2562 [Steve Henson, Mark Cox]
2564 *) Fix double free in TLS server name extensions which could lead to
2565 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
2568 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
2570 Clear the error queue to ensure that error entries left from
2571 older function calls do not interfere with the correct operation.
2572 [Lutz Jaenicke, Erik de Castro Lopo]
2574 *) Remove root CA certificates of commercial CAs:
2576 The OpenSSL project does not recommend any specific CA and does not
2577 have any policy with respect to including or excluding any CA.
2578 Therefore it does not make any sense to ship an arbitrary selection
2579 of root CA certificates with the OpenSSL software.
2582 *) RSA OAEP patches to fix two separate invalid memory reads.
2583 The first one involves inputs when 'lzero' is greater than
2584 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
2585 before the beginning of from). The second one involves inputs where
2586 the 'db' section contains nothing but zeroes (there is a one-byte
2587 invalid read after the end of 'db').
2588 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2590 *) Partial backport from 0.9.9-dev:
2592 Introduce bn_mul_mont (dedicated Montgomery multiplication
2593 procedure) as a candidate for BIGNUM assembler implementation.
2594 While 0.9.9-dev uses assembler for various architectures, only
2595 x86_64 is available by default here in the 0.9.8 branch, and
2596 32-bit x86 is available through a compile-time setting.
2598 To try the 32-bit x86 assembler implementation, use Configure
2599 option "enable-montasm" (which exists only for this backport).
2601 As "enable-montasm" for 32-bit x86 disclaims code stability
2602 anyway, in this constellation we activate additional code
2603 backported from 0.9.9-dev for further performance improvements,
2604 namely BN_from_montgomery_word. (To enable this otherwise,
2605 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
2607 [Andy Polyakov (backport partially by Bodo Moeller)]
2609 *) Add TLS session ticket callback. This allows an application to set
2610 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
2611 values. This is useful for key rollover for example where several key
2612 sets may exist with different names.
2615 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
2616 This was broken until now in 0.9.8 releases, such that the only way
2617 a registered ENGINE could be used (assuming it initialises
2618 successfully on the host) was to explicitly set it as the default
2619 for the relevant algorithms. This is in contradiction with 0.9.7
2620 behaviour and the documentation. With this fix, when an ENGINE is
2621 registered into a given algorithm's table of implementations, the
2622 'uptodate' flag is reset so that auto-discovery will be used next
2623 time a new context for that algorithm attempts to select an
2625 [Ian Lister (tweaked by Geoff Thorpe)]
2627 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
2628 implemention in the following ways:
2630 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
2633 Lack of BER streaming support means one pass streaming processing is
2634 only supported if data is detached: setting the streaming flag is
2635 ignored for embedded content.
2637 CMS support is disabled by default and must be explicitly enabled
2638 with the enable-cms configuration option.
2641 *) Update the GMP engine glue to do direct copies between BIGNUM and
2642 mpz_t when openssl and GMP use the same limb size. Otherwise the
2643 existing "conversion via a text string export" trick is still used.
2644 [Paul Sheer <paulsheer@gmail.com>]
2646 *) Zlib compression BIO. This is a filter BIO which compressed and
2647 uncompresses any data passed through it.
2650 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
2651 RFC3394 compatible AES key wrapping.
2654 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
2655 sets string data without copying. X509_ALGOR_set0() and
2656 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
2657 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
2658 from an X509_ATTRIBUTE structure optionally checking it occurs only
2659 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
2663 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
2664 to get the expected BN_FLG_CONSTTIME behavior.
2665 [Bodo Moeller (Google)]
2669 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
2670 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
2671 - added some more tests to do_tests.pl
2672 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
2673 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
2674 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
2675 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
2676 - various changes to netware.pl to enable gcc-cross builds on Win32
2678 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
2679 - various changes to fix missing prototype warnings
2680 - fixed x86nasm.pl to create correct asm files for NASM COFF output
2681 - added AES, WHIRLPOOL and CPUID assembler code to build files
2682 - added missing AES assembler make rules to mk1mf.pl
2683 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
2684 [Guenter Knauf <eflash@gmx.net>]
2686 *) Implement certificate status request TLS extension defined in RFC3546.
2687 A client can set the appropriate parameters and receive the encoded
2688 OCSP response via a callback. A server can query the supplied parameters
2689 and set the encoded OCSP response in the callback. Add simplified examples
2690 to s_client and s_server.
2693 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
2695 *) Fix various bugs:
2696 + Binary incompatibility of ssl_ctx_st structure
2697 + DTLS interoperation with non-compliant servers
2698 + Don't call get_session_cb() without proposed session
2699 + Fix ia64 assembler code
2700 [Andy Polyakov, Steve Henson]
2702 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
2704 *) DTLS Handshake overhaul. There were longstanding issues with
2705 OpenSSL DTLS implementation, which were making it impossible for
2706 RFC 4347 compliant client to communicate with OpenSSL server.
2707 Unfortunately just fixing these incompatibilities would "cut off"
2708 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
2709 server keeps tolerating non RFC compliant syntax. The opposite is
2710 not true, 0.9.8f client can not communicate with earlier server.
2711 This update even addresses CVE-2007-4995.
2714 *) Changes to avoid need for function casts in OpenSSL: some compilers
2715 (gcc 4.2 and later) reject their use.
2716 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
2719 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2720 RFC4507bis. The encrypted ticket format is an encrypted encoded
2721 SSL_SESSION structure, that way new session features are automatically
2724 If a client application caches session in an SSL_SESSION structure
2725 support is transparent because tickets are now stored in the encoded
2728 The SSL_CTX structure automatically generates keys for ticket
2729 protection in servers so again support should be possible
2730 with no application modification.
2732 If a client or server wishes to disable RFC4507 support then the option
2733 SSL_OP_NO_TICKET can be set.
2735 Add a TLS extension debugging callback to allow the contents of any client
2736 or server extensions to be examined.
2738 This work was sponsored by Google.
2741 *) Add initial support for TLS extensions, specifically for the server_name
2742 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2743 have new members for a host name. The SSL data structure has an
2744 additional member SSL_CTX *initial_ctx so that new sessions can be
2745 stored in that context to allow for session resumption, even after the
2746 SSL has been switched to a new SSL_CTX in reaction to a client's
2747 server_name extension.
2749 New functions (subject to change):
2751 SSL_get_servername()
2752 SSL_get_servername_type()
2755 New CTRL codes and macros (subject to change):
2757 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2758 - SSL_CTX_set_tlsext_servername_callback()
2759 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2760 - SSL_CTX_set_tlsext_servername_arg()
2761 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2763 openssl s_client has a new '-servername ...' option.
2765 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2766 '-key2 ...', '-servername_fatal' (subject to change). This allows
2767 testing the HostName extension for a specific single host name ('-cert'
2768 and '-key' remain fallbacks for handshakes without HostName
2769 negotiation). If the unrecogninzed_name alert has to be sent, this by
2770 default is a warning; it becomes fatal with the '-servername_fatal'
2773 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2775 *) Add AES and SSE2 assembly language support to VC++ build.
2778 *) Mitigate attack on final subtraction in Montgomery reduction.
2781 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2782 (which previously caused an internal error).
2785 *) Squeeze another 10% out of IGE mode when in != out.
2788 *) AES IGE mode speedup.
2789 [Dean Gaudet (Google)]
2791 *) Add the Korean symmetric 128-bit cipher SEED (see
2792 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2793 add SEED ciphersuites from RFC 4162:
2795 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2796 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2797 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2798 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2800 To minimize changes between patchlevels in the OpenSSL 0.9.8
2801 series, SEED remains excluded from compilation unless OpenSSL
2802 is configured with 'enable-seed'.
2803 [KISA, Bodo Moeller]
2805 *) Mitigate branch prediction attacks, which can be practical if a
2806 single processor is shared, allowing a spy process to extract
2807 information. For detailed background information, see
2808 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2809 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2810 and Necessary Software Countermeasures"). The core of the change
2811 are new versions BN_div_no_branch() and
2812 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2813 respectively, which are slower, but avoid the security-relevant
2814 conditional branches. These are automatically called by BN_div()
2815 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2816 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2817 remove a conditional branch.
2819 BN_FLG_CONSTTIME is the new name for the previous
2820 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2821 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2822 in the exponent causes BN_mod_exp_mont() to use the alternative
2823 implementation in BN_mod_exp_mont_consttime().) The old name
2824 remains as a deprecated alias.
2826 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2827 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2828 constant-time implementations for more than just exponentiation.
2829 Here too the old name is kept as a deprecated alias.
2831 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2832 the BN_BLINDING structure gets an independent copy of the
2833 modulus. This means that the previous "BIGNUM *m" argument to
2834 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2835 essentially becomes "const BIGNUM *m", although we can't actually
2836 change this in the header file before 0.9.9. It allows
2837 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2838 enable BN_FLG_CONSTTIME.
2840 [Matthew D Wood (Intel Corp)]
2842 *) In the SSL/TLS server implementation, be strict about session ID
2843 context matching (which matters if an application uses a single
2844 external cache for different purposes). Previously,
2845 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2846 set. This did ensure strict client verification, but meant that,
2847 with applications using a single external cache for quite
2848 different requirements, clients could circumvent ciphersuite
2849 restrictions for a given session ID context by starting a session
2850 in a different context.
2853 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2854 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2855 authentication-only ciphersuites.
2858 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2859 not complete and could lead to a possible single byte overflow
2860 (CVE-2007-5135) [Ben Laurie]
2862 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2864 *) Since AES128 and AES256 (and similarly Camellia128 and
2865 Camellia256) share a single mask bit in the logic of
2866 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2867 kludge to work properly if AES128 is available and AES256 isn't
2868 (or if Camellia128 is available and Camellia256 isn't).
2871 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2872 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2873 When a point or a seed is encoded in a BIT STRING, we need to
2874 prevent the removal of trailing zero bits to get the proper DER
2875 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2876 of a NamedBitList, for which trailing 0 bits need to be removed.)
2879 *) Have SSL/TLS server implementation tolerate "mismatched" record
2880 protocol version while receiving ClientHello even if the
2881 ClientHello is fragmented. (The server can't insist on the
2882 particular protocol version it has chosen before the ServerHello
2883 message has informed the client about his choice.)
2886 *) Add RFC 3779 support.
2887 [Rob Austein for ARIN, Ben Laurie]
2889 *) Load error codes if they are not already present instead of using a
2890 static variable. This allows them to be cleanly unloaded and reloaded.
2891 Improve header file function name parsing.
2894 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2895 or CAPABILITY handshake as required by RFCs.
2898 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2900 *) Introduce limits to prevent malicious keys being able to
2901 cause a denial of service. (CVE-2006-2940)
2902 [Steve Henson, Bodo Moeller]
2904 *) Fix ASN.1 parsing of certain invalid structures that can result
2905 in a denial of service. (CVE-2006-2937) [Steve Henson]
2907 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2908 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2910 *) Fix SSL client code which could crash if connecting to a
2911 malicious SSLv2 server. (CVE-2006-4343)
2912 [Tavis Ormandy and Will Drewry, Google Security Team]
2914 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2915 match only those. Before that, "AES256-SHA" would be interpreted
2916 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2917 the same strength classification in 0.9.7h) as we currently only
2918 have a single AES bit in the ciphersuite description bitmap.
2919 That change, however, also applied to ciphersuite strings such as
2920 "RC4-MD5" that intentionally matched multiple ciphersuites --
2921 namely, SSL 2.0 ciphersuites in addition to the more common ones
2922 from SSL 3.0/TLS 1.0.
2924 So we change the selection algorithm again: Naming an explicit
2925 ciphersuite selects this one ciphersuite, and any other similar
2926 ciphersuite (same bitmap) from *other* protocol versions.
2927 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2928 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2930 Since SSL 2.0 does not have any ciphersuites for which the
2931 128/256 bit distinction would be relevant, this works for now.
2932 The proper fix will be to use different bits for AES128 and
2933 AES256, which would have avoided the problems from the beginning;
2934 however, bits are scarce, so we can only do this in a new release
2935 (not just a patchlevel) when we can change the SSL_CIPHER
2936 definition to split the single 'unsigned long mask' bitmap into
2937 multiple values to extend the available space.
2941 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2943 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2944 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2946 *) Add AES IGE and biIGE modes.
2949 *) Change the Unix randomness entropy gathering to use poll() when
2950 possible instead of select(), since the latter has some
2951 undesirable limitations.
2952 [Darryl Miles via Richard Levitte and Bodo Moeller]
2954 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2955 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2956 cannot be implicitly activated as part of, e.g., the "AES" alias.
2957 However, please upgrade to OpenSSL 0.9.9[-dev] for
2958 non-experimental use of the ECC ciphersuites to get TLS extension
2959 support, which is required for curve and point format negotiation
2960 to avoid potential handshake problems.
2963 *) Disable rogue ciphersuites:
2965 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2966 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2967 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2969 The latter two were purportedly from
2970 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2973 Also deactivate the remaining ciphersuites from
2974 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2975 unofficial, and the ID has long expired.
2978 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2979 dual-core machines) and other potential thread-safety issues.
2982 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2983 versions), which is now available for royalty-free use
2984 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2985 Also, add Camellia TLS ciphersuites from RFC 4132.
2987 To minimize changes between patchlevels in the OpenSSL 0.9.8
2988 series, Camellia remains excluded from compilation unless OpenSSL
2989 is configured with 'enable-camellia'.
2992 *) Disable the padding bug check when compression is in use. The padding
2993 bug check assumes the first packet is of even length, this is not
2994 necessarily true if compresssion is enabled and can result in false
2995 positives causing handshake failure. The actual bug test is ancient
2996 code so it is hoped that implementations will either have fixed it by
2997 now or any which still have the bug do not support compression.
3000 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3002 *) When applying a cipher rule check to see if string match is an explicit
3003 cipher suite and only match that one cipher suite if it is.
3006 *) Link in manifests for VC++ if needed.
3007 [Austin Ziegler <halostatue@gmail.com>]
3009 *) Update support for ECC-based TLS ciphersuites according to
3010 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3011 TLS extensions, which are supported starting with the 0.9.9
3012 branch, not in the OpenSSL 0.9.8 branch).
3015 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3016 opaque EVP_CIPHER_CTX handling.
3019 *) Fixes and enhancements to zlib compression code. We now only use
3020 "zlib1.dll" and use the default __cdecl calling convention on Win32
3021 to conform with the standards mentioned here:
3022 http://www.zlib.net/DLL_FAQ.txt
3023 Static zlib linking now works on Windows and the new --with-zlib-include
3024 --with-zlib-lib options to Configure can be used to supply the location
3025 of the headers and library. Gracefully handle case where zlib library
3029 *) Several fixes and enhancements to the OID generation code. The old code
3030 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3031 handle numbers larger than ULONG_MAX, truncated printing and had a
3032 non standard OBJ_obj2txt() behaviour.
3035 *) Add support for building of engines under engine/ as shared libraries
3036 under VC++ build system.
3039 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3040 Hopefully, we will not see any false combination of paths any more.
3043 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3045 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3046 (part of SSL_OP_ALL). This option used to disable the
3047 countermeasure against man-in-the-middle protocol-version
3048 rollback in the SSL 2.0 server implementation, which is a bad
3049 idea. (CVE-2005-2969)
3051 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3052 for Information Security, National Institute of Advanced Industrial
3053 Science and Technology [AIST], Japan)]
3055 *) Add two function to clear and return the verify parameter flags.
3058 *) Keep cipherlists sorted in the source instead of sorting them at
3059 runtime, thus removing the need for a lock.
3062 *) Avoid some small subgroup attacks in Diffie-Hellman.
3063 [Nick Mathewson and Ben Laurie]
3065 *) Add functions for well-known primes.
3068 *) Extended Windows CE support.
3069 [Satoshi Nakamura and Andy Polyakov]
3071 *) Initialize SSL_METHOD structures at compile time instead of during
3072 runtime, thus removing the need for a lock.
3075 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3076 attempting to decrypt each encrypted key in turn. Add support to
3080 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3082 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3085 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3088 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3089 key into the same file any more.
3092 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3095 *) Add -utf8 command line and config file option to 'ca'.
3096 [Stefan <stf@udoma.org]
3098 *) Removed the macro des_crypt(), as it seems to conflict with some
3099 libraries. Use DES_crypt().
3102 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3103 involves renaming the source and generated shared-libs for
3104 both. The engines will accept the corrected or legacy ids
3105 ('ncipher' and '4758_cca' respectively) when binding. NB,
3106 this only applies when building 'shared'.
3107 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3109 *) Add attribute functions to EVP_PKEY structure. Modify
3110 PKCS12_create() to recognize a CSP name attribute and
3111 use it. Make -CSP option work again in pkcs12 utility.
3114 *) Add new functionality to the bn blinding code:
3115 - automatic re-creation of the BN_BLINDING parameters after
3116 a fixed number of uses (currently 32)
3117 - add new function for parameter creation
3118 - introduce flags to control the update behaviour of the
3119 BN_BLINDING parameters
3120 - hide BN_BLINDING structure
3121 Add a second BN_BLINDING slot to the RSA structure to improve
3122 performance when a single RSA object is shared among several
3126 *) Add support for DTLS.
3127 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3129 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3130 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3133 *) Remove buggy and incompletet DH cert support from
3134 ssl/ssl_rsa.c and ssl/s3_both.c
3137 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3138 the apps/openssl applications.
3141 *) Compile clean with "-Wall -Wmissing-prototypes
3142 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
3143 DEBUG_SAFESTACK must also be set.
3146 *) Change ./Configure so that certain algorithms can be disabled by default.
3147 The new counterpiece to "no-xxx" is "enable-xxx".
3149 The patented RC5 and MDC2 algorithms will now be disabled unless
3150 "enable-rc5" and "enable-mdc2", respectively, are specified.
3152 (IDEA remains enabled despite being patented. This is because IDEA
3153 is frequently required for interoperability, and there is no license
3154 fee for non-commercial use. As before, "no-idea" can be used to
3155 avoid this algorithm.)
3159 *) Add processing of proxy certificates (see RFC 3820). This work was
3160 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
3161 EGEE (Enabling Grids for E-science in Europe).
3164 *) RC4 performance overhaul on modern architectures/implementations, such
3165 as Intel P4, IA-64 and AMD64.
3168 *) New utility extract-section.pl. This can be used specify an alternative
3169 section number in a pod file instead of having to treat each file as
3170 a separate case in Makefile. This can be done by adding two lines to the
3173 =for comment openssl_section:XXX
3175 The blank line is mandatory.
3179 *) New arguments -certform, -keyform and -pass for s_client and s_server
3180 to allow alternative format key and certificate files and passphrase
3184 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
3185 update associated structures and add various utility functions.
3187 Add new policy related verify parameters, include policy checking in
3188 standard verify code. Enhance 'smime' application with extra parameters
3189 to support policy checking and print out.
3192 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
3193 Nehemiah processors. These extensions support AES encryption in hardware
3194 as well as RNG (though RNG support is currently disabled).
3195 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
3197 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
3200 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
3201 [Andy Polyakov and a number of other people]
3203 *) Improved PowerPC platform support. Most notably BIGNUM assembler
3204 implementation contributed by IBM.
3205 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
3207 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
3208 exponent rather than 'unsigned long'. There is a corresponding change to
3209 the new 'rsa_keygen' element of the RSA_METHOD structure.
3210 [Jelte Jansen, Geoff Thorpe]
3212 *) Functionality for creating the initial serial number file is now
3213 moved from CA.pl to the 'ca' utility with a new option -create_serial.
3215 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
3216 number file to 1, which is bound to cause problems. To avoid
3217 the problems while respecting compatibility between different 0.9.7
3218 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
3219 CA.pl for serial number initialization. With the new release 0.9.8,
3220 we can fix the problem directly in the 'ca' utility.)
3223 *) Reduced header interdepencies by declaring more opaque objects in
3224 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
3225 give fewer recursive includes, which could break lazy source code - so
3226 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
3227 developers should define this symbol when building and using openssl to
3228 ensure they track the recommended behaviour, interfaces, [etc], but
3229 backwards-compatible behaviour prevails when this isn't defined.
3232 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
3235 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
3236 This will generate a random key of the appropriate length based on the
3237 cipher context. The EVP_CIPHER can provide its own random key generation
3238 routine to support keys of a specific form. This is used in the des and
3239 3des routines to generate a key of the correct parity. Update S/MIME
3240 code to use new functions and hence generate correct parity DES keys.
3241 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
3242 valid (weak or incorrect parity).
3245 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
3246 as looking them up. This is useful when the verified structure may contain
3247 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
3248 present unless the new PKCS7_NO_CRL flag is asserted.
3251 *) Extend ASN1 oid configuration module. It now additionally accepts the
3254 shortName = some long name, 1.2.3.4
3257 *) Reimplemented the BN_CTX implementation. There is now no more static
3258 limitation on the number of variables it can handle nor the depth of the
3259 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
3260 information can now expand as required, and rather than having a single
3261 static array of bignums, BN_CTX now uses a linked-list of such arrays
3262 allowing it to expand on demand whilst maintaining the usefulness of
3263 BN_CTX's "bundling".
3266 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
3267 to allow all RSA operations to function using a single BN_CTX.
3270 *) Preliminary support for certificate policy evaluation and checking. This
3271 is initially intended to pass the tests outlined in "Conformance Testing
3272 of Relying Party Client Certificate Path Processing Logic" v1.07.
3275 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
3276 remained unused and not that useful. A variety of other little bignum
3277 tweaks and fixes have also been made continuing on from the audit (see
3281 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
3282 associated ASN1, EVP and SSL functions and old ASN1 macros.
3285 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
3286 and this should never fail. So the return value from the use of
3287 BN_set_word() (which can fail due to needless expansion) is now deprecated;
3288 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
3291 *) BN_CTX_get() should return zero-valued bignums, providing the same
3292 initialised value as BN_new().
3293 [Geoff Thorpe, suggested by Ulf Möller]
3295 *) Support for inhibitAnyPolicy certificate extension.
3298 *) An audit of the BIGNUM code is underway, for which debugging code is
3299 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
3300 is considered valid when processing BIGNUMs, and causes execution to
3301 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
3302 further steps are taken to deliberately pollute unused data in BIGNUM
3303 structures to try and expose faulty code further on. For now, openssl will
3304 (in its default mode of operation) continue to tolerate the inconsistent
3305 forms that it has tolerated in the past, but authors and packagers should
3306 consider trying openssl and their own applications when compiled with
3307 these debugging symbols defined. It will help highlight potential bugs in
3308 their own code, and will improve the test coverage for OpenSSL itself. At
3309 some point, these tighter rules will become openssl's default to improve
3310 maintainability, though the assert()s and other overheads will remain only
3311 in debugging configurations. See bn.h for more details.
3312 [Geoff Thorpe, Nils Larsch, Ulf Möller]
3314 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
3315 that can only be obtained through BN_CTX_new() (which implicitly
3316 initialises it). The presence of this function only made it possible
3317 to overwrite an existing structure (and cause memory leaks).
3320 *) Because of the callback-based approach for implementing LHASH as a
3321 template type, lh_insert() adds opaque objects to hash-tables and
3322 lh_doall() or lh_doall_arg() are typically used with a destructor callback
3323 to clean up those corresponding objects before destroying the hash table
3324 (and losing the object pointers). So some over-zealous constifications in
3325 LHASH have been relaxed so that lh_insert() does not take (nor store) the
3326 objects as "const" and the lh_doall[_arg] callback wrappers are not
3327 prototyped to have "const" restrictions on the object pointers they are
3328 given (and so aren't required to cast them away any more).
3331 *) The tmdiff.h API was so ugly and minimal that our own timing utility
3332 (speed) prefers to use its own implementation. The two implementations
3333 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
3334 its object type properly exposed (MS_TM) instead of casting to/from "char
3335 *". This may still change yet if someone realises MS_TM and "ms_time_***"
3336 aren't necessarily the greatest nomenclatures - but this is what was used
3337 internally to the implementation so I've used that for now.
3340 *) Ensure that deprecated functions do not get compiled when
3341 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
3342 the self-tests were still using deprecated key-generation functions so
3343 these have been updated also.
3346 *) Reorganise PKCS#7 code to separate the digest location functionality
3347 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
3348 New function PKCS7_set_digest() to set the digest type for PKCS#7
3349 digestedData type. Add additional code to correctly generate the
3350 digestedData type and add support for this type in PKCS7 initialization
3354 *) New function PKCS7_set0_type_other() this initializes a PKCS7
3355 structure of type "other".
3358 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
3359 sure the loop does correctly stop and breaking ("division by zero")
3360 modulus operations are not performed. The (pre-generated) prime
3361 table crypto/bn/bn_prime.h was already correct, but it could not be
3362 re-generated on some platforms because of the "division by zero"
3363 situation in the script.
3364 [Ralf S. Engelschall]
3366 *) Update support for ECC-based TLS ciphersuites according to
3367 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
3368 SHA-1 now is only used for "small" curves (where the
3369 representation of a field element takes up to 24 bytes); for
3370 larger curves, the field element resulting from ECDH is directly
3371 used as premaster secret.
3372 [Douglas Stebila (Sun Microsystems Laboratories)]
3374 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
3375 curve secp160r1 to the tests.
3376 [Douglas Stebila (Sun Microsystems Laboratories)]
3378 *) Add the possibility to load symbols globally with DSO.
3379 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
3381 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
3382 control of the error stack.
3385 *) Add support for STORE in ENGINE.
3388 *) Add the STORE type. The intention is to provide a common interface
3389 to certificate and key stores, be they simple file-based stores, or
3390 HSM-type store, or LDAP stores, or...
3391 NOTE: The code is currently UNTESTED and isn't really used anywhere.
3394 *) Add a generic structure called OPENSSL_ITEM. This can be used to
3395 pass a list of arguments to any function as well as provide a way
3396 for a function to pass data back to the caller.
3399 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
3400 works like BUF_strdup() but can be used to duplicate a portion of
3401 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
3405 *) Add the function sk_find_ex() which works like sk_find(), but will
3406 return an index to an element even if an exact match couldn't be
3407 found. The index is guaranteed to point at the element where the
3408 searched-for key would be inserted to preserve sorting order.
3411 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
3412 takes an extra flags argument for optional functionality. Currently,
3413 the following flags are defined:
3415 OBJ_BSEARCH_VALUE_ON_NOMATCH
3416 This one gets OBJ_bsearch_ex() to return a pointer to the first
3417 element where the comparing function returns a negative or zero
3420 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
3421 This one gets OBJ_bsearch_ex() to return a pointer to the first
3422 element where the comparing function returns zero. This is useful
3423 if there are more than one element where the comparing function
3427 *) Make it possible to create self-signed certificates with 'openssl ca'
3428 in such a way that the self-signed certificate becomes part of the
3429 CA database and uses the same mechanisms for serial number generation
3430 as all other certificate signing. The new flag '-selfsign' enables
3431 this functionality. Adapt CA.sh and CA.pl.in.
3434 *) Add functionality to check the public key of a certificate request
3435 against a given private. This is useful to check that a certificate
3436 request can be signed by that key (self-signing).
3439 *) Make it possible to have multiple active certificates with the same
3440 subject in the CA index file. This is done only if the keyword
3441 'unique_subject' is set to 'no' in the main CA section (default
3442 if 'CA_default') of the configuration file. The value is saved
3443 with the database itself in a separate index attribute file,
3444 named like the index file with '.attr' appended to the name.
3447 *) Generate muti valued AVAs using '+' notation in config files for
3451 *) Support for nameConstraints certificate extension.
3454 *) Support for policyConstraints certificate extension.
3457 *) Support for policyMappings certificate extension.
3460 *) Make sure the default DSA_METHOD implementation only uses its
3461 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3462 and change its own handlers to be NULL so as to remove unnecessary
3463 indirection. This lets alternative implementations fallback to the
3464 default implementation more easily.
3467 *) Support for directoryName in GeneralName related extensions
3471 *) Make it possible to link applications using Makefile.shared.
3472 Make that possible even when linking against static libraries!
3475 *) Support for single pass processing for S/MIME signing. This now
3476 means that S/MIME signing can be done from a pipe, in addition
3477 cleartext signing (multipart/signed type) is effectively streaming
3478 and the signed data does not need to be all held in memory.
3480 This is done with a new flag PKCS7_STREAM. When this flag is set
3481 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3482 is done after the data is output (and digests calculated) in
3483 SMIME_write_PKCS7().
3486 *) Add full support for -rpath/-R, both in shared libraries and
3487 applications, at least on the platforms where it's known how
3491 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3492 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3493 will now compute a table of multiples of the generator that
3494 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3495 faster (notably in the case of a single point multiplication,
3496 scalar * generator).
3497 [Nils Larsch, Bodo Moeller]
3499 *) IPv6 support for certificate extensions. The various extensions
3500 which use the IP:a.b.c.d can now take IPv6 addresses using the
3501 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3505 *) Added an ENGINE that implements RSA by performing private key
3506 exponentiations with the GMP library. The conversions to and from
3507 GMP's mpz_t format aren't optimised nor are any montgomery forms
3508 cached, and on x86 it appears OpenSSL's own performance has caught up.
3509 However there are likely to be other architectures where GMP could
3510 provide a boost. This ENGINE is not built in by default, but it can be
3511 specified at Configure time and should be accompanied by the necessary
3512 linker additions, eg;
3513 ./config -DOPENSSL_USE_GMP -lgmp
3516 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3517 testing availability of engines with "-t" - the old behaviour is
3518 produced by increasing the feature's verbosity with "-tt".
3521 *) ECDSA routines: under certain error conditions uninitialized BN objects
3522 could be freed. Solution: make sure initialization is performed early
3523 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3527 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3528 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3529 software implementations. For DSA and DH, parameter generation can
3530 also be overriden by providing the appropriate method callbacks.
3533 *) Change the "progress" mechanism used in key-generation and
3534 primality testing to functions that take a new BN_GENCB pointer in
3535 place of callback/argument pairs. The new API functions have "_ex"
3536 postfixes and the older functions are reimplemented as wrappers for
3537 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3538 declarations of the old functions to help (graceful) attempts to
3539 migrate to the new functions. Also, the new key-generation API
3540 functions operate on a caller-supplied key-structure and return
3541 success/failure rather than returning a key or NULL - this is to
3542 help make "keygen" another member function of RSA_METHOD etc.
3544 Example for using the new callback interface:
3546 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3550 BN_GENCB_set(&my_cb, my_callback, my_arg);
3552 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3553 /* For the meaning of a, b in calls to my_callback(), see the
3554 * documentation of the function that calls the callback.
3555 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3556 * my_callback should return 1 if it wants BN_is_prime_ex()
3557 * to continue, or 0 to stop.
3562 *) Change the ZLIB compression method to be stateful, and make it
3563 available to TLS with the number defined in
3564 draft-ietf-tls-compression-04.txt.
3567 *) Add the ASN.1 structures and functions for CertificatePair, which
3568 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
3570 CertificatePair ::= SEQUENCE {
3571 forward [0] Certificate OPTIONAL,
3572 reverse [1] Certificate OPTIONAL,
3573 -- at least one of the pair shall be present -- }
3575 Also implement the PEM functions to read and write certificate
3576 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
3578 This needed to be defined, mostly for the sake of the LDAP
3579 attribute crossCertificatePair, but may prove useful elsewhere as
3583 *) Make it possible to inhibit symlinking of shared libraries in
3584 Makefile.shared, for Cygwin's sake.
3587 *) Extend the BIGNUM API by creating a function
3588 void BN_set_negative(BIGNUM *a, int neg);
3589 and a macro that behave like
3590 int BN_is_negative(const BIGNUM *a);
3592 to avoid the need to access 'a->neg' directly in applications.
3595 *) Implement fast modular reduction for pseudo-Mersenne primes
3596 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
3597 EC_GROUP_new_curve_GFp() will now automatically use this
3599 [Nils Larsch <nla@trustcenter.de>]
3601 *) Add new lock type (CRYPTO_LOCK_BN).
3604 *) Change the ENGINE framework to automatically load engines
3605 dynamically from specific directories unless they could be
3606 found to already be built in or loaded. Move all the
3607 current engines except for the cryptodev one to a new
3609 The engines in engines/ are built as shared libraries if
3610 the "shared" options was given to ./Configure or ./config.
3611 Otherwise, they are inserted in libcrypto.a.
3612 /usr/local/ssl/engines is the default directory for dynamic
3613 engines, but that can be overriden at configure time through
3614 the usual use of --prefix and/or --openssldir, and at run
3615 time with the environment variable OPENSSL_ENGINES.
3616 [Geoff Thorpe and Richard Levitte]
3618 *) Add Makefile.shared, a helper makefile to build shared
3619 libraries. Addapt Makefile.org.
3622 *) Add version info to Win32 DLLs.
3623 [Peter 'Luna' Runestig" <peter@runestig.com>]
3625 *) Add new 'medium level' PKCS#12 API. Certificates and keys
3626 can be added using this API to created arbitrary PKCS#12
3627 files while avoiding the low level API.
3629 New options to PKCS12_create(), key or cert can be NULL and
3630 will then be omitted from the output file. The encryption
3631 algorithm NIDs can be set to -1 for no encryption, the mac
3632 iteration count can be set to 0 to omit the mac.
3634 Enhance pkcs12 utility by making the -nokeys and -nocerts
3635 options work when creating a PKCS#12 file. New option -nomac
3636 to omit the mac, NONE can be set for an encryption algorithm.
3637 New code is modified to use the enhanced PKCS12_create()
3638 instead of the low level API.
3641 *) Extend ASN1 encoder to support indefinite length constructed
3642 encoding. This can output sequences tags and octet strings in
3643 this form. Modify pk7_asn1.c to support indefinite length
3644 encoding. This is experimental and needs additional code to
3645 be useful, such as an ASN1 bio and some enhanced streaming
3648 Extend template encode functionality so that tagging is passed
3649 down to the template encoder.
3652 *) Let 'openssl req' fail if an argument to '-newkey' is not
3653 recognized instead of using RSA as a default.
3656 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
3657 As these are not official, they are not included in "ALL";
3658 the "ECCdraft" ciphersuite group alias can be used to select them.
3659 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
3661 *) Add ECDH engine support.
3662 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
3664 *) Add ECDH in new directory crypto/ecdh/.
3665 [Douglas Stebila (Sun Microsystems Laboratories)]
3667 *) Let BN_rand_range() abort with an error after 100 iterations
3668 without success (which indicates a broken PRNG).
3671 *) Change BN_mod_sqrt() so that it verifies that the input value
3672 is really the square of the return value. (Previously,
3673 BN_mod_sqrt would show GIGO behaviour.)
3676 *) Add named elliptic curves over binary fields from X9.62, SECG,
3677 and WAP/WTLS; add OIDs that were still missing.
3679 [Sheueling Chang Shantz and Douglas Stebila
3680 (Sun Microsystems Laboratories)]
3682 *) Extend the EC library for elliptic curves over binary fields
3683 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
3686 EC_GF2m_simple_method
3690 EC_GROUP_new_curve_GF2m
3691 EC_GROUP_set_curve_GF2m
3692 EC_GROUP_get_curve_GF2m
3693 EC_POINT_set_affine_coordinates_GF2m
3694 EC_POINT_get_affine_coordinates_GF2m
3695 EC_POINT_set_compressed_coordinates_GF2m
3697 Point compression for binary fields is disabled by default for
3698 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
3701 As binary polynomials are represented as BIGNUMs, various members
3702 of the EC_GROUP and EC_POINT data structures can be shared
3703 between the implementations for prime fields and binary fields;
3704 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
3705 are essentially identical to their ..._GFp counterparts.
3706 (For simplicity, the '..._GFp' prefix has been dropped from
3707 various internal method names.)
3709 An internal 'field_div' method (similar to 'field_mul' and
3710 'field_sqr') has been added; this is used only for binary fields.
3712 [Sheueling Chang Shantz and Douglas Stebila
3713 (Sun Microsystems Laboratories)]
3715 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
3716 through methods ('mul', 'precompute_mult').
3718 The generic implementations (now internally called 'ec_wNAF_mul'
3719 and 'ec_wNAF_precomputed_mult') remain the default if these
3720 methods are undefined.
3722 [Sheueling Chang Shantz and Douglas Stebila
3723 (Sun Microsystems Laboratories)]
3725 *) New function EC_GROUP_get_degree, which is defined through
3726 EC_METHOD. For curves over prime fields, this returns the bit
3727 length of the modulus.
3729 [Sheueling Chang Shantz and Douglas Stebila
3730 (Sun Microsystems Laboratories)]
3732 *) New functions EC_GROUP_dup, EC_POINT_dup.
3733 (These simply call ..._new and ..._copy).
3735 [Sheueling Chang Shantz and Douglas Stebila
3736 (Sun Microsystems Laboratories)]
3738 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3739 Polynomials are represented as BIGNUMs (where the sign bit is not
3740 used) in the following functions [macros]:
3743 BN_GF2m_sub [= BN_GF2m_add]
3744 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3745 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3746 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3748 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3749 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3750 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3751 BN_GF2m_cmp [= BN_ucmp]
3753 (Note that only the 'mod' functions are actually for fields GF(2^m).
3754 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3756 For some functions, an the irreducible polynomial defining a
3757 field can be given as an 'unsigned int[]' with strictly
3758 decreasing elements giving the indices of those bits that are set;
3759 i.e., p[] represents the polynomial
3760 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3762 p[0] > p[1] > ... > p[k] = 0.
3763 This applies to the following functions:
3768 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3769 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3771 BN_GF2m_mod_sqrt_arr
3772 BN_GF2m_mod_solve_quad_arr
3776 Conversion can be performed by the following functions:
3781 bntest.c has additional tests for binary polynomial arithmetic.
3783 Two implementations for BN_GF2m_mod_div() are available.
3784 The default algorithm simply uses BN_GF2m_mod_inv() and
3785 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3786 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3787 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3789 [Sheueling Chang Shantz and Douglas Stebila
3790 (Sun Microsystems Laboratories)]
3792 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3793 functionality is disabled at compile-time.
3794 [Douglas Stebila <douglas.stebila@sun.com>]
3796 *) Change default behaviour of 'openssl asn1parse' so that more
3797 information is visible when viewing, e.g., a certificate:
3799 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3800 mode the content of non-printable OCTET STRINGs is output in a
3801 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3802 avoid the appearance of a printable string.
3803 [Nils Larsch <nla@trustcenter.de>]
3805 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3807 EC_GROUP_set_asn1_flag()
3808 EC_GROUP_get_asn1_flag()
3809 EC_GROUP_set_point_conversion_form()
3810 EC_GROUP_get_point_conversion_form()
3811 These control ASN1 encoding details:
3812 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3813 has been set to OPENSSL_EC_NAMED_CURVE.
3814 - Points are encoded in uncompressed form by default; options for
3815 asn1_for are as for point2oct, namely
3816 POINT_CONVERSION_COMPRESSED
3817 POINT_CONVERSION_UNCOMPRESSED
3818 POINT_CONVERSION_HYBRID
3820 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3823 EC_GROUP_get0_seed()
3824 EC_GROUP_get_seed_len()
3825 This is used only for ASN1 purposes (so far).
3826 [Nils Larsch <nla@trustcenter.de>]
3828 *) Add 'field_type' member to EC_METHOD, which holds the NID
3829 of the appropriate field type OID. The new function
3830 EC_METHOD_get_field_type() returns this value.
3831 [Nils Larsch <nla@trustcenter.de>]
3836 EC_POINT_point2hex()
3837 EC_POINT_hex2point()
3838 providing useful interfaces to EC_POINT_point2oct() and
3839 EC_POINT_oct2point().
3840 [Nils Larsch <nla@trustcenter.de>]
3842 *) Change internals of the EC library so that the functions
3843 EC_GROUP_set_generator()
3844 EC_GROUP_get_generator()
3845 EC_GROUP_get_order()
3846 EC_GROUP_get_cofactor()
3847 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3848 to methods, which would lead to unnecessary code duplication when
3849 adding different types of curves.
3850 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3852 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3853 arithmetic, and such that modified wNAFs are generated
3854 (which avoid length expansion in many cases).
3857 *) Add a function EC_GROUP_check_discriminant() (defined via
3858 EC_METHOD) that verifies that the curve discriminant is non-zero.
3860 Add a function EC_GROUP_check() that makes some sanity tests
3861 on a EC_GROUP, its generator and order. This includes
3862 EC_GROUP_check_discriminant().
3863 [Nils Larsch <nla@trustcenter.de>]
3865 *) Add ECDSA in new directory crypto/ecdsa/.
3867 Add applications 'openssl ecparam' and 'openssl ecdsa'
3868 (these are based on 'openssl dsaparam' and 'openssl dsa').
3870 ECDSA support is also included in various other files across the
3871 library. Most notably,
3872 - 'openssl req' now has a '-newkey ecdsa:file' option;
3873 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3874 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3875 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3876 them suitable for ECDSA where domain parameters must be
3877 extracted before the specific public key;
3878 - ECDSA engine support has been added.
3879 [Nils Larsch <nla@trustcenter.de>]
3881 *) Include some named elliptic curves, and add OIDs from X9.62,
3882 SECG, and WAP/WTLS. Each curve can be obtained from the new
3884 EC_GROUP_new_by_curve_name(),
3885 and the list of available named curves can be obtained with
3886 EC_get_builtin_curves().
3887 Also add a 'curve_name' member to EC_GROUP objects, which can be
3889 EC_GROUP_set_curve_name()
3890 EC_GROUP_get_curve_name()
3891 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3893 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3894 was actually never needed) and in BN_mul(). The removal in BN_mul()
3895 required a small change in bn_mul_part_recursive() and the addition
3896 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3897 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3898 bn_sub_words() and bn_add_words() except they take arrays with
3902 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3904 *) Cleanse PEM buffers before freeing them since they may contain
3906 [Benjamin Bennett <ben@psc.edu>]
3908 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3909 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3910 authentication-only ciphersuites.
3913 *) Since AES128 and AES256 share a single mask bit in the logic of
3914 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3915 kludge to work properly if AES128 is available and AES256 isn't.
3918 *) Expand security boundary to match 1.1.1 module.
3921 *) Remove redundant features: hash file source, editing of test vectors
3922 modify fipsld to use external fips_premain.c signature.
3925 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3926 run algorithm test programs.
3929 *) Make algorithm test programs more tolerant of whitespace.
3932 *) Have SSL/TLS server implementation tolerate "mismatched" record
3933 protocol version while receiving ClientHello even if the
3934 ClientHello is fragmented. (The server can't insist on the
3935 particular protocol version it has chosen before the ServerHello
3936 message has informed the client about his choice.)
3939 *) Load error codes if they are not already present instead of using a
3940 static variable. This allows them to be cleanly unloaded and reloaded.
3943 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3945 *) Introduce limits to prevent malicious keys being able to
3946 cause a denial of service. (CVE-2006-2940)
3947 [Steve Henson, Bodo Moeller]
3949 *) Fix ASN.1 parsing of certain invalid structures that can result
3950 in a denial of service. (CVE-2006-2937) [Steve Henson]
3952 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3953 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3955 *) Fix SSL client code which could crash if connecting to a
3956 malicious SSLv2 server. (CVE-2006-4343)
3957 [Tavis Ormandy and Will Drewry, Google Security Team]
3959 *) Change ciphersuite string processing so that an explicit
3960 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3961 will no longer include "AES128-SHA"), and any other similar
3962 ciphersuite (same bitmap) from *other* protocol versions (so that
3963 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3964 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3965 changes from 0.9.8b and 0.9.8d.
3968 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3970 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3971 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3973 *) Change the Unix randomness entropy gathering to use poll() when
3974 possible instead of select(), since the latter has some
3975 undesirable limitations.
3976 [Darryl Miles via Richard Levitte and Bodo Moeller]
3978 *) Disable rogue ciphersuites:
3980 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3981 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3982 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3984 The latter two were purportedly from
3985 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3988 Also deactive the remaining ciphersuites from
3989 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3990 unofficial, and the ID has long expired.
3993 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3994 dual-core machines) and other potential thread-safety issues.
3997 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3999 *) Adapt fipsld and the build system to link against the validated FIPS
4000 module in FIPS mode.
4003 *) Fixes for VC++ 2005 build under Windows.
4006 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4007 from a Windows bash shell such as MSYS. It is autodetected from the
4008 "config" script when run from a VC++ environment. Modify standard VC++
4009 build to use fipscanister.o from the GNU make build.
4012 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4014 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4015 The value now differs depending on if you build for FIPS or not.
4016 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4017 safely run with a non-FIPSed libcrypto, as it may crash because of
4018 the difference induced by this change.
4021 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4023 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4024 (part of SSL_OP_ALL). This option used to disable the
4025 countermeasure against man-in-the-middle protocol-version
4026 rollback in the SSL 2.0 server implementation, which is a bad
4027 idea. (CVE-2005-2969)
4029 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4030 for Information Security, National Institute of Advanced Industrial
4031 Science and Technology [AIST], Japan)]
4033 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4034 mainly for FIPS compliance and not fully integrated at this stage.
4037 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4038 the exponentiation using a fixed-length exponent. (Otherwise,
4039 the information leaked through timing could expose the secret key
4040 after many signatures; cf. Bleichenbacher's attack on DSA with
4044 *) Make a new fixed-window mod_exp implementation the default for
4045 RSA, DSA, and DH private-key operations so that the sequence of
4046 squares and multiplies and the memory access pattern are
4047 independent of the particular secret key. This will mitigate
4048 cache-timing and potential related attacks.
4050 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4051 and this is automatically used by BN_mod_exp_mont() if the new flag
4052 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4053 will use this BN flag for private exponents unless the flag
4054 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4055 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4057 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4059 *) Change the client implementation for SSLv23_method() and
4060 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4061 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4062 (Previously, the SSL 2.0 backwards compatible Client Hello
4063 message format would be used even with SSL_OP_NO_SSLv2.)
4066 *) Add support for smime-type MIME parameter in S/MIME messages which some
4070 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4071 a threadsafe manner. Modify rsa code to use new function and add calls
4072 to dsa and dh code (which had race conditions before).
4075 *) Include the fixed error library code in the C error file definitions
4076 instead of fixing them up at runtime. This keeps the error code
4077 structures constant.
4080 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4082 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4085 *) Fixes for newer kerberos headers. NB: the casts are needed because
4086 the 'length' field is signed on one version and unsigned on another
4087 with no (?) obvious way to tell the difference, without these VC++
4088 complains. Also the "definition" of FAR (blank) is no longer included
4089 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4090 some needed definitions.
4093 *) Undo Cygwin change.
4096 *) Added support for proxy certificates according to RFC 3820.
4097 Because they may be a security thread to unaware applications,
4098 they must be explicitely allowed in run-time. See
4099 docs/HOWTO/proxy_certificates.txt for further information.
4102 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4104 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4105 server and client random values. Previously
4106 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4107 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4109 This change has negligible security impact because:
4111 1. Server and client random values still have 24 bytes of pseudo random
4114 2. Server and client random values are sent in the clear in the initial
4117 3. The master secret is derived using the premaster secret (48 bytes in
4118 size for static RSA ciphersuites) as well as client server and random
4121 The OpenSSL team would like to thank the UK NISCC for bringing this issue
4124 [Stephen Henson, reported by UK NISCC]
4126 *) Use Windows randomness collection on Cygwin.
4129 *) Fix hang in EGD/PRNGD query when communication socket is closed
4130 prematurely by EGD/PRNGD.
4131 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
4133 *) Prompt for pass phrases when appropriate for PKCS12 input format.
4136 *) Back-port of selected performance improvements from development
4137 branch, as well as improved support for PowerPC platforms.
4140 *) Add lots of checks for memory allocation failure, error codes to indicate
4141 failure and freeing up memory if a failure occurs.
4142 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
4144 *) Add new -passin argument to dgst.
4147 *) Perform some character comparisons of different types in X509_NAME_cmp:
4148 this is needed for some certificates that reencode DNs into UTF8Strings
4149 (in violation of RFC3280) and can't or wont issue name rollover
4153 *) Make an explicit check during certificate validation to see that
4154 the CA setting in each certificate on the chain is correct. As a
4155 side effect always do the following basic checks on extensions,
4156 not just when there's an associated purpose to the check:
4158 - if there is an unhandled critical extension (unless the user
4159 has chosen to ignore this fault)
4160 - if the path length has been exceeded (if one is set at all)
4161 - that certain extensions fit the associated purpose (if one has
4165 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
4167 *) Avoid a race condition when CRLs are checked in a multi threaded
4168 environment. This would happen due to the reordering of the revoked
4169 entries during signature checking and serial number lookup. Now the
4170 encoding is cached and the serial number sort performed under a lock.
4171 Add new STACK function sk_is_sorted().
4174 *) Add Delta CRL to the extension code.
4177 *) Various fixes to s3_pkt.c so alerts are sent properly.
4178 [David Holmes <d.holmes@f5.com>]
4180 *) Reduce the chances of duplicate issuer name and serial numbers (in
4181 violation of RFC3280) using the OpenSSL certificate creation utilities.
4182 This is done by creating a random 64 bit value for the initial serial
4183 number when a serial number file is created or when a self signed
4184 certificate is created using 'openssl req -x509'. The initial serial
4185 number file is created using 'openssl x509 -next_serial' in CA.pl
4186 rather than being initialized to 1.
4189 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
4191 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4192 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4193 [Joe Orton, Steve Henson]
4195 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
4197 [Joe Orton, Steve Henson]
4199 *) Make it possible to have multiple active certificates with the same
4200 subject in the CA index file. This is done only if the keyword
4201 'unique_subject' is set to 'no' in the main CA section (default
4202 if 'CA_default') of the configuration file. The value is saved
4203 with the database itself in a separate index attribute file,
4204 named like the index file with '.attr' appended to the name.
4207 *) X509 verify fixes. Disable broken certificate workarounds when
4208 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
4209 keyUsage extension present. Don't accept CRLs with unhandled critical
4210 extensions: since verify currently doesn't process CRL extensions this
4211 rejects a CRL with *any* critical extensions. Add new verify error codes
4215 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
4216 A clarification of RFC2560 will require the use of OCTET STRINGs and
4217 some implementations cannot handle the current raw format. Since OpenSSL
4218 copies and compares OCSP nonces as opaque blobs without any attempt at
4219 parsing them this should not create any compatibility issues.
4222 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
4223 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
4224 this HMAC (and other) operations are several times slower than OpenSSL
4228 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
4229 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4231 *) Use the correct content when signing type "other".
4234 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
4236 *) Fix various bugs revealed by running the NISCC test suite:
4238 Stop out of bounds reads in the ASN1 code when presented with
4239 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4241 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
4243 If verify callback ignores invalid public key errors don't try to check
4244 certificate signature with the NULL public key.
4248 *) New -ignore_err option in ocsp application to stop the server
4249 exiting on the first error in a request.
4252 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4253 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4257 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4258 extra data after the compression methods not only for TLS 1.0
4259 but also for SSL 3.0 (as required by the specification).
4260 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4262 *) Change X509_certificate_type() to mark the key as exported/exportable
4263 when it's 512 *bits* long, not 512 bytes.
4266 *) Change AES_cbc_encrypt() so it outputs exact multiple of
4267 blocks during encryption.
4270 *) Various fixes to base64 BIO and non blocking I/O. On write
4271 flushes were not handled properly if the BIO retried. On read
4272 data was not being buffered properly and had various logic bugs.
4273 This also affects blocking I/O when the data being decoded is a
4277 *) Various S/MIME bugfixes and compatibility changes:
4278 output correct application/pkcs7 MIME type if
4279 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
4280 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
4281 of files as .eml work). Correctly handle very long lines in MIME
4285 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
4287 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4288 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4289 a protocol version number mismatch like a decryption error
4290 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4293 *) Turn on RSA blinding by default in the default implementation
4294 to avoid a timing attack. Applications that don't want it can call
4295 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4296 They would be ill-advised to do so in most cases.
4297 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4299 *) Change RSA blinding code so that it works when the PRNG is not
4300 seeded (in this case, the secret RSA exponent is abused as
4301 an unpredictable seed -- if it is not unpredictable, there
4302 is no point in blinding anyway). Make RSA blinding thread-safe
4303 by remembering the creator's thread ID in rsa->blinding and
4304 having all other threads use local one-time blinding factors
4305 (this requires more computation than sharing rsa->blinding, but
4306 avoids excessive locking; and if an RSA object is not shared
4307 between threads, blinding will still be very fast).
4310 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
4311 ENGINE as defaults for all supported algorithms irrespective of
4312 the 'flags' parameter. 'flags' is now honoured, so applications
4313 should make sure they are passing it correctly.
4316 *) Target "mingw" now allows native Windows code to be generated in
4317 the Cygwin environment as well as with the MinGW compiler.
4320 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
4322 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4323 via timing by performing a MAC computation even if incorrrect
4324 block cipher padding has been found. This is a countermeasure
4325 against active attacks where the attacker has to distinguish
4326 between bad padding and a MAC verification error. (CVE-2003-0078)
4328 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4329 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4330 Martin Vuagnoux (EPFL, Ilion)]
4332 *) Make the no-err option work as intended. The intention with no-err
4333 is not to have the whole error stack handling routines removed from
4334 libcrypto, it's only intended to remove all the function name and
4335 reason texts, thereby removing some of the footprint that may not
4336 be interesting if those errors aren't displayed anyway.
4338 NOTE: it's still possible for any application or module to have it's
4339 own set of error texts inserted. The routines are there, just not
4340 used by default when no-err is given.
4343 *) Add support for FreeBSD on IA64.
4344 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
4346 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
4347 Kerberos function mit_des_cbc_cksum(). Before this change,
4348 the value returned by DES_cbc_cksum() was like the one from
4349 mit_des_cbc_cksum(), except the bytes were swapped.
4350 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
4352 *) Allow an application to disable the automatic SSL chain building.
4353 Before this a rather primitive chain build was always performed in
4354 ssl3_output_cert_chain(): an application had no way to send the
4355 correct chain if the automatic operation produced an incorrect result.
4357 Now the chain builder is disabled if either:
4359 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
4361 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
4363 The reasoning behind this is that an application would not want the
4364 auto chain building to take place if extra chain certificates are
4365 present and it might also want a means of sending no additional
4366 certificates (for example the chain has two certificates and the
4370 *) Add the possibility to build without the ENGINE framework.
4371 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4373 *) Under Win32 gmtime() can return NULL: check return value in
4374 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
4377 *) DSA routines: under certain error conditions uninitialized BN objects
4378 could be freed. Solution: make sure initialization is performed early
4379 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
4380 Nils Larsch <nla@trustcenter.de> via PR#459)
4383 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
4384 checked on reconnect on the client side, therefore session resumption
4385 could still fail with a "ssl session id is different" error. This
4386 behaviour is masked when SSL_OP_ALL is used due to
4387 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
4388 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4389 followup to PR #377.
4392 *) IA-32 assembler support enhancements: unified ELF targets, support
4393 for SCO/Caldera platforms, fix for Cygwin shared build.
4396 *) Add support for FreeBSD on sparc64. As a consequence, support for
4397 FreeBSD on non-x86 processors is separate from x86 processors on
4398 the config script, much like the NetBSD support.
4399 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
4401 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
4403 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
4406 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
4407 code (06) was taken as the first octet of the session ID and the last
4408 octet was ignored consequently. As a result SSLv2 client side session
4409 caching could not have worked due to the session ID mismatch between
4411 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4415 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
4416 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
4420 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
4421 seems that in spite of existing for more than a year, many application
4422 author have done nothing to provide the necessary callbacks, which
4423 means that this particular engine will not work properly anywhere.
4424 This is a very unfortunate situation which forces us, in the name
4425 of usability, to give the hw_ncipher.c a static lock, which is part
4427 NOTE: This is for the 0.9.7 series ONLY. This hack will never
4428 appear in 0.9.8 or later. We EXPECT application authors to have
4429 dealt properly with this when 0.9.8 is released (unless we actually
4430 make such changes in the libcrypto locking code that changes will
4431 have to be made anyway).
4434 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
4435 octets have been read, EOF or an error occurs. Without this change
4436 some truncated ASN1 structures will not produce an error.
4439 *) Disable Heimdal support, since it hasn't been fully implemented.
4440 Still give the possibility to force the use of Heimdal, but with
4441 warnings and a request that patches get sent to openssl-dev.
4444 *) Add the VC-CE target, introduce the WINCE sysname, and add
4445 INSTALL.WCE and appropriate conditionals to make it build.
4446 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4448 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4449 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4450 edit numbers of the version.
4451 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4453 *) Introduce safe string copy and catenation functions
4454 (BUF_strlcpy() and BUF_strlcat()).
4455 [Ben Laurie (CHATS) and Richard Levitte]
4457 *) Avoid using fixed-size buffers for one-line DNs.
4458 [Ben Laurie (CHATS)]
4460 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4461 resizing buffers containing secrets, and use where appropriate.
4462 [Ben Laurie (CHATS)]
4464 *) Avoid using fixed size buffers for configuration file location.
4465 [Ben Laurie (CHATS)]
4467 *) Avoid filename truncation for various CA files.
4468 [Ben Laurie (CHATS)]
4470 *) Use sizeof in preference to magic numbers.
4471 [Ben Laurie (CHATS)]
4473 *) Avoid filename truncation in cert requests.
4474 [Ben Laurie (CHATS)]
4476 *) Add assertions to check for (supposedly impossible) buffer
4478 [Ben Laurie (CHATS)]
4480 *) Don't cache truncated DNS entries in the local cache (this could
4481 potentially lead to a spoofing attack).
4482 [Ben Laurie (CHATS)]
4484 *) Fix various buffers to be large enough for hex/decimal
4485 representations in a platform independent manner.
4486 [Ben Laurie (CHATS)]
4488 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4489 resizing buffers containing secrets, and use where appropriate.
4490 [Ben Laurie (CHATS)]
4492 *) Add BIO_indent() to avoid much slightly worrying code to do
4494 [Ben Laurie (CHATS)]
4496 *) Convert sprintf()/BIO_puts() to BIO_printf().
4497 [Ben Laurie (CHATS)]
4499 *) buffer_gets() could terminate with the buffer only half
4501 [Ben Laurie (CHATS)]
4503 *) Add assertions to prevent user-supplied crypto functions from
4504 overflowing internal buffers by having large block sizes, etc.
4505 [Ben Laurie (CHATS)]
4507 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4509 [Ben Laurie (CHATS)]
4511 *) Eliminate unused copy of key in RC4.
4512 [Ben Laurie (CHATS)]
4514 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4515 [Ben Laurie (CHATS)]
4517 *) Fix off-by-one error in EGD path.
4518 [Ben Laurie (CHATS)]
4520 *) If RANDFILE path is too long, ignore instead of truncating.
4521 [Ben Laurie (CHATS)]
4523 *) Eliminate unused and incorrectly sized X.509 structure
4525 [Ben Laurie (CHATS)]
4527 *) Eliminate unused and dangerous function knumber().
4528 [Ben Laurie (CHATS)]
4530 *) Eliminate unused and dangerous structure, KSSL_ERR.
4531 [Ben Laurie (CHATS)]
4533 *) Protect against overlong session ID context length in an encoded
4534 session object. Since these are local, this does not appear to be
4536 [Ben Laurie (CHATS)]
4538 *) Change from security patch (see 0.9.6e below) that did not affect
4539 the 0.9.6 release series:
4541 Remote buffer overflow in SSL3 protocol - an attacker could
4542 supply an oversized master key in Kerberos-enabled versions.
4544 [Ben Laurie (CHATS)]
4546 *) Change the SSL kerb5 codes to match RFC 2712.
4549 *) Make -nameopt work fully for req and add -reqopt switch.
4550 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4552 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4553 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4555 *) Make sure tests can be performed even if the corresponding algorithms
4556 have been removed entirely. This was also the last step to make
4557 OpenSSL compilable with DJGPP under all reasonable conditions.
4558 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
4560 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
4561 to allow version independent disabling of normally unselected ciphers,
4562 which may be activated as a side-effect of selecting a single cipher.
4564 (E.g., cipher list string "RSA" enables ciphersuites that are left
4565 out of "ALL" because they do not provide symmetric encryption.
4566 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
4567 [Lutz Jaenicke, Bodo Moeller]
4569 *) Add appropriate support for separate platform-dependent build
4570 directories. The recommended way to make a platform-dependent
4571 build directory is the following (tested on Linux), maybe with
4574 # Place yourself outside of the OpenSSL source tree. In
4575 # this example, the environment variable OPENSSL_SOURCE
4576 # is assumed to contain the absolute OpenSSL source directory.
4577 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
4578 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
4579 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
4580 mkdir -p `dirname $F`
4581 ln -s $OPENSSL_SOURCE/$F $F
4584 To be absolutely sure not to disturb the source tree, a "make clean"
4585 is a good thing. If it isn't successfull, don't worry about it,
4586 it probably means the source directory is very clean.
4589 *) Make sure any ENGINE control commands make local copies of string
4590 pointers passed to them whenever necessary. Otherwise it is possible
4591 the caller may have overwritten (or deallocated) the original string
4592 data when a later ENGINE operation tries to use the stored values.
4593 [Götz Babin-Ebell <babinebell@trustcenter.de>]
4595 *) Improve diagnostics in file reading and command-line digests.
4596 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
4598 *) Add AES modes CFB and OFB to the object database. Correct an
4599 error in AES-CFB decryption.
4602 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
4603 allows existing EVP_CIPHER_CTX structures to be reused after
4604 calling EVP_*Final(). This behaviour is used by encryption
4605 BIOs and some applications. This has the side effect that
4606 applications must explicitly clean up cipher contexts with
4607 EVP_CIPHER_CTX_cleanup() or they will leak memory.
4610 *) Check the values of dna and dnb in bn_mul_recursive before calling
4611 bn_mul_comba (a non zero value means the a or b arrays do not contain
4612 n2 elements) and fallback to bn_mul_normal if either is not zero.
4615 *) Fix escaping of non-ASCII characters when using the -subj option
4616 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
4619 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
4620 form for "surname", serialNumber has no short form.
4621 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
4622 therefore remove "mail" short name for "internet 7".
4623 The OID for unique identifiers in X509 certificates is
4624 x500UniqueIdentifier, not uniqueIdentifier.
4625 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
4628 *) Add an "init" command to the ENGINE config module and auto initialize
4629 ENGINEs. Without any "init" command the ENGINE will be initialized
4630 after all ctrl commands have been executed on it. If init=1 the
4631 ENGINE is initailized at that point (ctrls before that point are run
4632 on the uninitialized ENGINE and after on the initialized one). If
4633 init=0 then the ENGINE will not be iniatialized at all.
4636 *) Fix the 'app_verify_callback' interface so that the user-defined
4637 argument is actually passed to the callback: In the
4638 SSL_CTX_set_cert_verify_callback() prototype, the callback
4639 declaration has been changed from
4642 int (*cb)(X509_STORE_CTX *,void *);
4643 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
4644 i=s->ctx->app_verify_callback(&ctx)
4645 has been changed into
4646 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
4648 To update applications using SSL_CTX_set_cert_verify_callback(),
4649 a dummy argument can be added to their callback functions.
4650 [D. K. Smetters <smetters@parc.xerox.com>]
4652 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
4653 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
4655 *) Add and OPENSSL_LOAD_CONF define which will cause
4656 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
4657 This allows older applications to transparently support certain
4658 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
4659 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
4660 load the config file and OPENSSL_add_all_algorithms_conf() which will
4661 always load it have also been added.
4664 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
4665 Adjust NIDs and EVP layer.
4666 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4668 *) Config modules support in openssl utility.
4670 Most commands now load modules from the config file,
4671 though in a few (such as version) this isn't done
4672 because it couldn't be used for anything.
4674 In the case of ca and req the config file used is
4675 the same as the utility itself: that is the -config
4676 command line option can be used to specify an
4680 *) Move default behaviour from OPENSSL_config(). If appname is NULL
4681 use "openssl_conf" if filename is NULL use default openssl config file.
4684 *) Add an argument to OPENSSL_config() to allow the use of an alternative
4685 config section name. Add a new flag to tolerate a missing config file
4686 and move code to CONF_modules_load_file().
4689 *) Support for crypto accelerator cards from Accelerated Encryption
4690 Processing, www.aep.ie. (Use engine 'aep')
4691 The support was copied from 0.9.6c [engine] and adapted/corrected
4692 to work with the new engine framework.
4693 [AEP Inc. and Richard Levitte]
4695 *) Support for SureWare crypto accelerator cards from Baltimore
4696 Technologies. (Use engine 'sureware')
4697 The support was copied from 0.9.6c [engine] and adapted
4698 to work with the new engine framework.
4701 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
4702 make the newer ENGINE framework commands for the CHIL engine work.
4703 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
4705 *) Make it possible to produce shared libraries on ReliantUNIX.
4706 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
4708 *) Add the configuration target debug-linux-ppro.
4709 Make 'openssl rsa' use the general key loading routines
4710 implemented in apps.c, and make those routines able to
4711 handle the key format FORMAT_NETSCAPE and the variant
4713 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4715 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4716 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4718 *) Add -keyform to rsautl, and document -engine.
4719 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
4721 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
4722 BIO_R_NO_SUCH_FILE error code rather than the generic
4723 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
4726 *) Add new functions
4728 ERR_peek_last_error_line
4729 ERR_peek_last_error_line_data.
4730 These are similar to
4733 ERR_peek_error_line_data,
4734 but report on the latest error recorded rather than the first one
4735 still in the error queue.
4736 [Ben Laurie, Bodo Moeller]
4738 *) default_algorithms option in ENGINE config module. This allows things
4740 default_algorithms = ALL
4741 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4744 *) Prelminary ENGINE config module.
4747 *) New experimental application configuration code.
4750 *) Change the AES code to follow the same name structure as all other
4751 symmetric ciphers, and behave the same way. Move everything to
4752 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4753 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4755 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4756 [Ben Laurie and Theo de Raadt]
4758 *) Add option to output public keys in req command.
4759 [Massimiliano Pala madwolf@openca.org]
4761 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4762 (up to about 10% better than before for P-192 and P-224).
4765 *) New functions/macros
4767 SSL_CTX_set_msg_callback(ctx, cb)
4768 SSL_CTX_set_msg_callback_arg(ctx, arg)
4769 SSL_set_msg_callback(ssl, cb)
4770 SSL_set_msg_callback_arg(ssl, arg)
4772 to request calling a callback function
4774 void cb(int write_p, int version, int content_type,
4775 const void *buf, size_t len, SSL *ssl, void *arg)
4777 whenever a protocol message has been completely received
4778 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4779 protocol version according to which the SSL library interprets
4780 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4781 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4782 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4783 specification (change_cipher_spec(20), alert(21), handshake(22)).
4784 'buf' and 'len' point to the actual message, 'ssl' to the
4785 SSL object, and 'arg' is the application-defined value set by
4786 SSL[_CTX]_set_msg_callback_arg().
4788 'openssl s_client' and 'openssl s_server' have new '-msg' options
4789 to enable a callback that displays all protocol messages.
4792 *) Change the shared library support so shared libraries are built as
4793 soon as the corresponding static library is finished, and thereby get
4794 openssl and the test programs linked against the shared library.
4795 This still only happens when the keyword "shard" has been given to
4796 the configuration scripts.
4798 NOTE: shared library support is still an experimental thing, and
4799 backward binary compatibility is still not guaranteed.
4800 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4802 *) Add support for Subject Information Access extension.
4803 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4805 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4806 additional bytes when new memory had to be allocated, not just
4807 when reusing an existing buffer.
4810 *) New command line and configuration option 'utf8' for the req command.
4811 This allows field values to be specified as UTF8 strings.
4814 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4815 runs for the former and machine-readable output for the latter.
4818 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4819 of the e-mail address in the DN (i.e., it will go into a certificate
4820 extension only). The new configuration file option 'email_in_dn = no'
4821 has the same effect.
4822 [Massimiliano Pala madwolf@openca.org]
4824 *) Change all functions with names starting with des_ to be starting
4825 with DES_ instead. Add wrappers that are compatible with libdes,
4826 but are named _ossl_old_des_*. Finally, add macros that map the
4827 des_* symbols to the corresponding _ossl_old_des_* if libdes
4828 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4829 desired, the des_* symbols will be mapped to DES_*, with one
4832 Since we provide two compatibility mappings, the user needs to
4833 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4834 compatibility is desired. The default (i.e., when that macro
4835 isn't defined) is OpenSSL 0.9.6c compatibility.
4837 There are also macros that enable and disable the support of old
4838 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4839 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4840 are defined, the default will apply: to support the old des routines.
4842 In either case, one must include openssl/des.h to get the correct
4843 definitions. Do not try to just include openssl/des_old.h, that
4846 NOTE: This is a major break of an old API into a new one. Software
4847 authors are encouraged to switch to the DES_ style functions. Some
4848 time in the future, des_old.h and the libdes compatibility functions
4849 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4850 default), and then completely removed.
4853 *) Test for certificates which contain unsupported critical extensions.
4854 If such a certificate is found during a verify operation it is
4855 rejected by default: this behaviour can be overridden by either
4856 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4857 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4858 X509_supported_extension() has also been added which returns 1 if a
4859 particular extension is supported.
4862 *) Modify the behaviour of EVP cipher functions in similar way to digests
4863 to retain compatibility with existing code.
4866 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4867 compatibility with existing code. In particular the 'ctx' parameter does
4868 not have to be to be initialized before the call to EVP_DigestInit() and
4869 it is tidied up after a call to EVP_DigestFinal(). New function
4870 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4871 EVP_MD_CTX_copy() changed to not require the destination to be
4872 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4873 requires the destination to be valid.
4875 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4876 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4879 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4880 so that complete 'Handshake' protocol structures are kept in memory
4881 instead of overwriting 'msg_type' and 'length' with 'body' data.
4884 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4885 [Massimo Santin via Richard Levitte]
4887 *) Major restructuring to the underlying ENGINE code. This includes
4888 reduction of linker bloat, separation of pure "ENGINE" manipulation
4889 (initialisation, etc) from functionality dealing with implementations
4890 of specific crypto iterfaces. This change also introduces integrated
4891 support for symmetric ciphers and digest implementations - so ENGINEs
4892 can now accelerate these by providing EVP_CIPHER and EVP_MD
4893 implementations of their own. This is detailed in crypto/engine/README
4894 as it couldn't be adequately described here. However, there are a few
4895 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4896 were changed in the original introduction of ENGINE code have now
4897 reverted back - the hooking from this code to ENGINE is now a good
4898 deal more passive and at run-time, operations deal directly with
4899 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4900 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4901 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4902 they were not being used by the framework as there is no concept of a
4903 BIGNUM_METHOD and they could not be generalised to the new
4904 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4905 ENGINE_cpy() has been removed as it cannot be consistently defined in
4909 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4912 *) Change mkdef.pl to sort symbols that get the same entry number,
4913 and make sure the automatically generated functions ERR_load_*
4914 become part of libeay.num as well.
4917 *) New function SSL_renegotiate_pending(). This returns true once
4918 renegotiation has been requested (either SSL_renegotiate() call
4919 or HelloRequest/ClientHello receveived from the peer) and becomes
4920 false once a handshake has been completed.
4921 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4922 sends a HelloRequest, but does not ensure that a handshake takes
4923 place. SSL_renegotiate_pending() is useful for checking if the
4924 client has followed the request.)
4927 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4928 By default, clients may request session resumption even during
4929 renegotiation (if session ID contexts permit); with this option,
4930 session resumption is possible only in the first handshake.
4932 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4933 more bits available for options that should not be part of
4934 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4937 *) Add some demos for certificate and certificate request creation.
4940 *) Make maximum certificate chain size accepted from the peer application
4941 settable (SSL*_get/set_max_cert_list()), as proposed by
4942 "Douglas E. Engert" <deengert@anl.gov>.
4945 *) Add support for shared libraries for Unixware-7
4946 (Boyd Lynn Gerber <gerberb@zenez.com>).
4949 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4950 be done prior to destruction. Use this to unload error strings from
4951 ENGINEs that load their own error strings. NB: This adds two new API
4952 functions to "get" and "set" this destroy handler in an ENGINE.
4955 *) Alter all existing ENGINE implementations (except "openssl" and
4956 "openbsd") to dynamically instantiate their own error strings. This
4957 makes them more flexible to be built both as statically-linked ENGINEs
4958 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4959 Also, add stub code to each that makes building them as self-contained
4960 shared-libraries easier (see README.ENGINE).
4963 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4964 implementations into applications that are completely implemented in
4965 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4966 commands that can be used to configure what shared-library to load and
4967 to control aspects of the way it is handled. Also, made an update to
4968 the README.ENGINE file that brings its information up-to-date and
4969 provides some information and instructions on the "dynamic" ENGINE
4970 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4973 *) Make it possible to unload ranges of ERR strings with a new
4974 "ERR_unload_strings" function.
4977 *) Add a copy() function to EVP_MD.
4980 *) Make EVP_MD routines take a context pointer instead of just the
4981 md_data void pointer.
4984 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4985 that the digest can only process a single chunk of data
4986 (typically because it is provided by a piece of
4987 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4988 is only going to provide a single chunk of data, and hence the
4989 framework needn't accumulate the data for oneshot drivers.
4992 *) As with "ERR", make it possible to replace the underlying "ex_data"
4993 functions. This change also alters the storage and management of global
4994 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4995 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4996 index counters. The API functions that use this state have been changed
4997 to take a "class_index" rather than pointers to the class's local STACK
4998 and counter, and there is now an API function to dynamically create new
4999 classes. This centralisation allows us to (a) plug a lot of the
5000 thread-safety problems that existed, and (b) makes it possible to clean
5001 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5002 such data would previously have always leaked in application code and
5003 workarounds were in place to make the memory debugging turn a blind eye
5004 to it. Application code that doesn't use this new function will still
5005 leak as before, but their memory debugging output will announce it now
5006 rather than letting it slide.
5008 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5009 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5010 has a return value to indicate success or failure.
5013 *) Make it possible to replace the underlying "ERR" functions such that the
5014 global state (2 LHASH tables and 2 locks) is only used by the "default"
5015 implementation. This change also adds two functions to "get" and "set"
5016 the implementation prior to it being automatically set the first time
5017 any other ERR function takes place. Ie. an application can call "get",
5018 pass the return value to a module it has just loaded, and that module
5019 can call its own "set" function using that value. This means the
5020 module's "ERR" operations will use (and modify) the error state in the
5021 application and not in its own statically linked copy of OpenSSL code.
5024 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5025 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5026 the operation, and provides a more encapsulated way for external code
5027 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5028 to use these functions rather than manually incrementing the counts.
5030 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5033 *) Add EVP test program.
5036 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5039 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5040 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5041 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5042 These allow a CRL to be built without having to access X509_CRL fields
5043 directly. Modify 'ca' application to use new functions.
5046 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5047 bug workarounds. Rollback attack detection is a security feature.
5048 The problem will only arise on OpenSSL servers when TLSv1 is not
5049 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5050 Software authors not wanting to support TLSv1 will have special reasons
5051 for their choice and can explicitly enable this option.
5052 [Bodo Moeller, Lutz Jaenicke]
5054 *) Rationalise EVP so it can be extended: don't include a union of
5055 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5056 (similar to those existing for EVP_CIPHER_CTX).
5061 EVP_MD_CTX_init(&md); /* new function call */
5062 EVP_DigestInit(&md, EVP_sha1());
5063 EVP_DigestUpdate(&md, in, len);
5064 EVP_DigestFinal(&md, out, NULL);
5065 EVP_MD_CTX_cleanup(&md); /* new function call */
5069 *) Make DES key schedule conform to the usual scheme, as well as
5070 correcting its structure. This means that calls to DES functions
5071 now have to pass a pointer to a des_key_schedule instead of a
5072 plain des_key_schedule (which was actually always a pointer
5075 des_key_schedule ks;
5077 des_set_key_checked(..., &ks);
5078 des_ncbc_encrypt(..., &ks, ...);
5080 (Note that a later change renames 'des_...' into 'DES_...'.)
5083 *) Initial reduction of linker bloat: the use of some functions, such as
5084 PEM causes large amounts of unused functions to be linked in due to
5085 poor organisation. For example pem_all.c contains every PEM function
5086 which has a knock on effect of linking in large amounts of (unused)
5087 ASN1 code. Grouping together similar functions and splitting unrelated
5088 functions prevents this.
5091 *) Cleanup of EVP macros.
5094 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5095 correct _ecb suffix.
5098 *) Add initial OCSP responder support to ocsp application. The
5099 revocation information is handled using the text based index
5100 use by the ca application. The responder can either handle
5101 requests generated internally, supplied in files (for example
5102 via a CGI script) or using an internal minimal server.
5105 *) Add configuration choices to get zlib compression for TLS.
5108 *) Changes to Kerberos SSL for RFC 2712 compliance:
5109 1. Implemented real KerberosWrapper, instead of just using
5110 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5111 2. Implemented optional authenticator field of KerberosWrapper.
5113 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5114 and authenticator structs; see crypto/krb5/.
5116 Generalized Kerberos calls to support multiple Kerberos libraries.
5117 [Vern Staats <staatsvr@asc.hpc.mil>,
5118 Jeffrey Altman <jaltman@columbia.edu>
5119 via Richard Levitte]
5121 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
5122 already does with RSA. testdsa.h now has 'priv_key/pub_key'
5123 values for each of the key sizes rather than having just
5124 parameters (and 'speed' generating keys each time).
5127 *) Speed up EVP routines.
5130 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
5131 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
5132 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
5133 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
5135 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
5136 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
5137 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
5140 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
5142 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
5145 *) Added the OS2-EMX target.
5146 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
5148 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
5149 to support NCONF routines in extension code. New function CONF_set_nconf()
5150 to allow functions which take an NCONF to also handle the old LHASH
5151 structure: this means that the old CONF compatible routines can be
5152 retained (in particular wrt extensions) without having to duplicate the
5153 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
5156 *) Enhance the general user interface with mechanisms for inner control
5157 and with possibilities to have yes/no kind of prompts.
5160 *) Change all calls to low level digest routines in the library and
5161 applications to use EVP. Add missing calls to HMAC_cleanup() and
5162 don't assume HMAC_CTX can be copied using memcpy().
5163 [Verdon Walker <VWalker@novell.com>, Steve Henson]
5165 *) Add the possibility to control engines through control names but with
5166 arbitrary arguments instead of just a string.
5167 Change the key loaders to take a UI_METHOD instead of a callback
5168 function pointer. NOTE: this breaks binary compatibility with earlier
5169 versions of OpenSSL [engine].
5170 Adapt the nCipher code for these new conditions and add a card insertion
5174 *) Enhance the general user interface with mechanisms to better support
5175 dialog box interfaces, application-defined prompts, the possibility
5176 to use defaults (for example default passwords from somewhere else)
5177 and interrupts/cancellations.
5180 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
5181 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
5184 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
5185 tidy up some unnecessarily weird code in 'sk_new()').
5186 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
5188 *) Change the key loading routines for ENGINEs to use the same kind
5189 callback (pem_password_cb) as all other routines that need this
5193 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
5194 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
5195 than this minimum value is recommended.
5198 *) New random seeder for OpenVMS, using the system process statistics
5199 that are easily reachable.
5202 *) Windows apparently can't transparently handle global
5203 variables defined in DLLs. Initialisations such as:
5205 const ASN1_ITEM *it = &ASN1_INTEGER_it;
5207 wont compile. This is used by the any applications that need to
5208 declare their own ASN1 modules. This was fixed by adding the option
5209 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
5210 needed for static libraries under Win32.
5213 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
5214 setting of purpose and trust fields. New X509_STORE trust and
5215 purpose functions and tidy up setting in other SSL functions.
5218 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
5219 structure. These are inherited by X509_STORE_CTX when it is
5220 initialised. This allows various defaults to be set in the
5221 X509_STORE structure (such as flags for CRL checking and custom
5222 purpose or trust settings) for functions which only use X509_STORE_CTX
5223 internally such as S/MIME.
5225 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
5226 trust settings if they are not set in X509_STORE. This allows X509_STORE
5227 purposes and trust (in S/MIME for example) to override any set by default.
5229 Add command line options for CRL checking to smime, s_client and s_server
5233 *) Initial CRL based revocation checking. If the CRL checking flag(s)
5234 are set then the CRL is looked up in the X509_STORE structure and
5235 its validity and signature checked, then if the certificate is found
5236 in the CRL the verify fails with a revoked error.
5238 Various new CRL related callbacks added to X509_STORE_CTX structure.
5240 Command line options added to 'verify' application to support this.
5242 This needs some additional work, such as being able to handle multiple
5243 CRLs with different times, extension based lookup (rather than just
5244 by subject name) and ultimately more complete V2 CRL extension
5248 *) Add a general user interface API (crypto/ui/). This is designed
5249 to replace things like des_read_password and friends (backward
5250 compatibility functions using this new API are provided).
5251 The purpose is to remove prompting functions from the DES code
5252 section as well as provide for prompting through dialog boxes in
5253 a window system and the like.
5256 *) Add "ex_data" support to ENGINE so implementations can add state at a
5257 per-structure level rather than having to store it globally.
5260 *) Make it possible for ENGINE structures to be copied when retrieved by
5261 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
5262 This causes the "original" ENGINE structure to act like a template,
5263 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
5264 operational state can be localised to each ENGINE structure, despite the
5265 fact they all share the same "methods". New ENGINE structures returned in
5266 this case have no functional references and the return value is the single
5267 structural reference. This matches the single structural reference returned
5268 by ENGINE_by_id() normally, when it is incremented on the pre-existing
5272 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
5273 needs to match any other type at all we need to manually clear the
5277 *) Changes to the "openssl engine" utility to include;
5278 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
5279 about an ENGINE's available control commands.
5280 - executing control commands from command line arguments using the
5281 '-pre' and '-post' switches. '-post' is only used if '-t' is
5282 specified and the ENGINE is successfully initialised. The syntax for
5283 the individual commands are colon-separated, for example;
5284 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
5287 *) New dynamic control command support for ENGINEs. ENGINEs can now
5288 declare their own commands (numbers), names (strings), descriptions,
5289 and input types for run-time discovery by calling applications. A
5290 subset of these commands are implicitly classed as "executable"
5291 depending on their input type, and only these can be invoked through
5292 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
5293 can be based on user input, config files, etc). The distinction is
5294 that "executable" commands cannot return anything other than a boolean
5295 result and can only support numeric or string input, whereas some
5296 discoverable commands may only be for direct use through
5297 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
5298 pointers, or other custom uses. The "executable" commands are to
5299 support parameterisations of ENGINE behaviour that can be
5300 unambiguously defined by ENGINEs and used consistently across any
5301 OpenSSL-based application. Commands have been added to all the
5302 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
5303 control over shared-library paths without source code alterations.
5306 *) Changed all ENGINE implementations to dynamically allocate their
5307 ENGINEs rather than declaring them statically. Apart from this being
5308 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
5309 this also allows the implementations to compile without using the
5310 internal engine_int.h header.
5313 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
5314 'const' value. Any code that should be able to modify a RAND_METHOD
5315 should already have non-const pointers to it (ie. they should only
5316 modify their own ones).
5319 *) Made a variety of little tweaks to the ENGINE code.
5320 - "atalla" and "ubsec" string definitions were moved from header files
5321 to C code. "nuron" string definitions were placed in variables
5322 rather than hard-coded - allowing parameterisation of these values
5323 later on via ctrl() commands.
5324 - Removed unused "#if 0"'d code.
5325 - Fixed engine list iteration code so it uses ENGINE_free() to release
5326 structural references.
5327 - Constified the RAND_METHOD element of ENGINE structures.
5328 - Constified various get/set functions as appropriate and added
5329 missing functions (including a catch-all ENGINE_cpy that duplicates
5330 all ENGINE values onto a new ENGINE except reference counts/state).
5331 - Removed NULL parameter checks in get/set functions. Setting a method
5332 or function to NULL is a way of cancelling out a previously set
5333 value. Passing a NULL ENGINE parameter is just plain stupid anyway
5334 and doesn't justify the extra error symbols and code.
5335 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
5336 flags from engine_int.h to engine.h.
5337 - Changed prototypes for ENGINE handler functions (init(), finish(),
5338 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
5341 *) Implement binary inversion algorithm for BN_mod_inverse in addition
5342 to the algorithm using long division. The binary algorithm can be
5343 used only if the modulus is odd. On 32-bit systems, it is faster
5344 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
5345 roughly 5-15% for 256-bit moduli), so we use it only for moduli
5346 up to 450 bits. In 64-bit environments, the binary algorithm
5347 appears to be advantageous for much longer moduli; here we use it
5348 for moduli up to 2048 bits.
5351 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
5352 could not support the combine flag in choice fields.
5355 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
5356 extensions from a certificate request to the certificate.
5359 *) Allow multiple 'certopt' and 'nameopt' options to be separated
5360 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
5361 file: this allows the display of the certificate about to be
5362 signed to be customised, to allow certain fields to be included
5363 or excluded and extension details. The old system didn't display
5364 multicharacter strings properly, omitted fields not in the policy
5365 and couldn't display additional details such as extensions.
5368 *) Function EC_POINTs_mul for multiple scalar multiplication
5369 of an arbitrary number of elliptic curve points
5370 \sum scalars[i]*points[i],
5371 optionally including the generator defined for the EC_GROUP:
5372 scalar*generator + \sum scalars[i]*points[i].
5374 EC_POINT_mul is a simple wrapper function for the typical case
5375 that the point list has just one item (besides the optional
5379 *) First EC_METHODs for curves over GF(p):
5381 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
5382 operations and provides various method functions that can also
5383 operate with faster implementations of modular arithmetic.
5385 EC_GFp_mont_method() reuses most functions that are part of
5386 EC_GFp_simple_method, but uses Montgomery arithmetic.
5388 [Bodo Moeller; point addition and point doubling
5389 implementation directly derived from source code provided by
5390 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
5392 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
5393 crypto/ec/ec_lib.c):
5395 Curves are EC_GROUP objects (with an optional group generator)
5396 based on EC_METHODs that are built into the library.
5398 Points are EC_POINT objects based on EC_GROUP objects.
5400 Most of the framework would be able to handle curves over arbitrary
5401 finite fields, but as there are no obvious types for fields other
5402 than GF(p), some functions are limited to that for now.
5405 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
5406 that the file contains a complete HTTP response.
5409 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
5410 change the def and num file printf format specifier from "%-40sXXX"
5411 to "%-39s XXX". The latter will always guarantee a space after the
5412 field while the former will cause them to run together if the field
5413 is 40 of more characters long.
5416 *) Constify the cipher and digest 'method' functions and structures
5417 and modify related functions to take constant EVP_MD and EVP_CIPHER
5421 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
5422 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
5425 *) Modify EVP_Digest*() routines so they now return values. Although the
5426 internal software routines can never fail additional hardware versions
5430 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
5432 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
5433 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
5436 ERR_R_NESTED_ASN1_ERROR
5438 ERR_R_MISSING_ASN1_EOS
5439 were 4 .. 9, conflicting with
5440 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5442 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5443 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5445 Add new error code 'ERR_R_INTERNAL_ERROR'.
5448 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5452 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5453 sets the subject name for a new request or supersedes the
5454 subject name in a given request. Formats that can be parsed are
5455 'CN=Some Name, OU=myOU, C=IT'
5457 'CN=Some Name/OU=myOU/C=IT'.
5459 Add options '-batch' and '-verbose' to 'openssl req'.
5460 [Massimiliano Pala <madwolf@hackmasters.net>]
5462 *) Introduce the possibility to access global variables through
5463 functions on platform were that's the best way to handle exporting
5464 global variables in shared libraries. To enable this functionality,
5465 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5466 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5467 is normally done by Configure or something similar).
5469 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5470 in the source file (foo.c) like this:
5472 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5473 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5475 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5476 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5478 OPENSSL_DECLARE_GLOBAL(int,foo);
5479 #define foo OPENSSL_GLOBAL_REF(foo)
5480 OPENSSL_DECLARE_GLOBAL(double,bar);
5481 #define bar OPENSSL_GLOBAL_REF(bar)
5483 The #defines are very important, and therefore so is including the
5484 header file everywhere where the defined globals are used.
5486 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5487 of ASN.1 items, but that structure is a bit different.
5489 The largest change is in util/mkdef.pl which has been enhanced with
5490 better and easier to understand logic to choose which symbols should
5491 go into the Windows .def files as well as a number of fixes and code
5492 cleanup (among others, algorithm keywords are now sorted
5493 lexicographically to avoid constant rewrites).
5496 *) In BN_div() keep a copy of the sign of 'num' before writing the
5497 result to 'rm' because if rm==num the value will be overwritten
5498 and produce the wrong result if 'num' is negative: this caused
5499 problems with BN_mod() and BN_nnmod().
5502 *) Function OCSP_request_verify(). This checks the signature on an
5503 OCSP request and verifies the signer certificate. The signer
5504 certificate is just checked for a generic purpose and OCSP request
5508 *) Add OCSP_check_validity() function to check the validity of OCSP
5509 responses. OCSP responses are prepared in real time and may only
5510 be a few seconds old. Simply checking that the current time lies
5511 between thisUpdate and nextUpdate max reject otherwise valid responses
5512 caused by either OCSP responder or client clock inaccuracy. Instead
5513 we allow thisUpdate and nextUpdate to fall within a certain period of
5514 the current time. The age of the response can also optionally be
5515 checked. Two new options -validity_period and -status_age added to
5519 *) If signature or public key algorithm is unrecognized print out its
5520 OID rather that just UNKNOWN.
5523 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5524 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5525 ID to be generated from the issuer certificate alone which can then be
5526 passed to OCSP_id_issuer_cmp().
5529 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5530 ASN1 modules to export functions returning ASN1_ITEM pointers
5531 instead of the ASN1_ITEM structures themselves. This adds several
5532 new macros which allow the underlying ASN1 function/structure to
5533 be accessed transparently. As a result code should not use ASN1_ITEM
5534 references directly (such as &X509_it) but instead use the relevant
5535 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5536 use of the new ASN1 code on platforms where exporting structures
5537 is problematical (for example in shared libraries) but exporting
5538 functions returning pointers to structures is not.
5541 *) Add support for overriding the generation of SSL/TLS session IDs.
5542 These callbacks can be registered either in an SSL_CTX or per SSL.
5543 The purpose of this is to allow applications to control, if they wish,
5544 the arbitrary values chosen for use as session IDs, particularly as it
5545 can be useful for session caching in multiple-server environments. A
5546 command-line switch for testing this (and any client code that wishes
5547 to use such a feature) has been added to "s_server".
5548 [Geoff Thorpe, Lutz Jaenicke]
5550 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5551 of the form '#if defined(...) || defined(...) || ...' and
5552 '#if !defined(...) && !defined(...) && ...'. This also avoids
5553 the growing number of special cases it was previously handling.
5556 *) Make all configuration macros available for application by making
5557 sure they are available in opensslconf.h, by giving them names starting
5558 with "OPENSSL_" to avoid conflicts with other packages and by making
5559 sure e_os2.h will cover all platform-specific cases together with
5561 Additionally, it is now possible to define configuration/platform-
5562 specific names (called "system identities"). In the C code, these
5563 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
5564 macro with the name beginning with "OPENSSL_SYS_", which is determined
5565 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
5569 *) New option -set_serial to 'req' and 'x509' this allows the serial
5570 number to use to be specified on the command line. Previously self
5571 signed certificates were hard coded with serial number 0 and the
5572 CA options of 'x509' had to use a serial number in a file which was
5576 *) New options to 'ca' utility to support V2 CRL entry extensions.
5577 Currently CRL reason, invalidity date and hold instruction are
5578 supported. Add new CRL extensions to V3 code and some new objects.
5581 *) New function EVP_CIPHER_CTX_set_padding() this is used to
5582 disable standard block padding (aka PKCS#5 padding) in the EVP
5583 API, which was previously mandatory. This means that the data is
5584 not padded in any way and so the total length much be a multiple
5585 of the block size, otherwise an error occurs.
5588 *) Initial (incomplete) OCSP SSL support.
5591 *) New function OCSP_parse_url(). This splits up a URL into its host,
5592 port and path components: primarily to parse OCSP URLs. New -url
5593 option to ocsp utility.
5596 *) New nonce behavior. The return value of OCSP_check_nonce() now
5597 reflects the various checks performed. Applications can decide
5598 whether to tolerate certain situations such as an absent nonce
5599 in a response when one was present in a request: the ocsp application
5600 just prints out a warning. New function OCSP_add1_basic_nonce()
5601 this is to allow responders to include a nonce in a response even if
5602 the request is nonce-less.
5605 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
5606 skipped when using openssl x509 multiple times on a single input file,
5607 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
5610 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
5611 set string type: to handle setting ASN1_TIME structures. Fix ca
5612 utility to correctly initialize revocation date of CRLs.
5615 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
5616 the clients preferred ciphersuites and rather use its own preferences.
5617 Should help to work around M$ SGC (Server Gated Cryptography) bug in
5618 Internet Explorer by ensuring unchanged hash method during stepup.
5619 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
5622 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
5623 to aes and add a new 'exist' option to print out symbols that don't
5627 *) Additional options to ocsp utility to allow flags to be set and
5628 additional certificates supplied.
5631 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
5632 OCSP client a number of certificate to only verify the response
5636 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
5637 handle the new API. Currently only ECB, CBC modes supported. Add new
5640 Add TLS AES ciphersuites as described in RFC3268, "Advanced
5641 Encryption Standard (AES) Ciphersuites for Transport Layer
5642 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
5643 not enabled by default and were not part of the "ALL" ciphersuite
5644 alias because they were not yet official; they could be
5645 explicitly requested by specifying the "AESdraft" ciphersuite
5646 group alias. In the final release of OpenSSL 0.9.7, the group
5647 alias is called "AES" and is part of "ALL".)
5648 [Ben Laurie, Steve Henson, Bodo Moeller]
5650 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
5651 request to response.
5654 *) Functions for OCSP responders. OCSP_request_onereq_count(),
5655 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
5656 extract information from a certificate request. OCSP_response_create()
5657 creates a response and optionally adds a basic response structure.
5658 OCSP_basic_add1_status() adds a complete single response to a basic
5659 response and returns the OCSP_SINGLERESP structure just added (to allow
5660 extensions to be included for example). OCSP_basic_add1_cert() adds a
5661 certificate to a basic response and OCSP_basic_sign() signs a basic
5662 response with various flags. New helper functions ASN1_TIME_check()
5663 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
5664 (converts ASN1_TIME to GeneralizedTime).
5667 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
5668 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
5669 structure from a certificate. X509_pubkey_digest() digests the public_key
5670 contents: this is used in various key identifiers.
5673 *) Make sk_sort() tolerate a NULL argument.
5674 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
5676 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
5677 passed by the function are trusted implicitly. If any of them signed the
5678 response then it is assumed to be valid and is not verified.
5681 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
5682 to data. This was previously part of the PKCS7 ASN1 code. This
5683 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
5684 [Steve Henson, reported by Kenneth R. Robinette
5685 <support@securenetterm.com>]
5687 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
5688 routines: without these tracing memory leaks is very painful.
5689 Fix leaks in PKCS12 and PKCS7 routines.
5692 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
5693 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
5694 effectively meant GeneralizedTime would never be used. Now it
5695 is initialised to -1 but X509_time_adj() now has to check the value
5696 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
5697 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
5698 [Steve Henson, reported by Kenneth R. Robinette
5699 <support@securenetterm.com>]
5701 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
5702 result in a zero length in the ASN1_INTEGER structure which was
5703 not consistent with the structure when d2i_ASN1_INTEGER() was used
5704 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
5705 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
5706 where it did not print out a minus for negative ASN1_INTEGER.
5709 *) Add summary printout to ocsp utility. The various functions which
5710 convert status values to strings have been renamed to:
5711 OCSP_response_status_str(), OCSP_cert_status_str() and
5712 OCSP_crl_reason_str() and are no longer static. New options
5713 to verify nonce values and to disable verification. OCSP response
5714 printout format cleaned up.
5717 *) Add additional OCSP certificate checks. These are those specified
5718 in RFC2560. This consists of two separate checks: the CA of the
5719 certificate being checked must either be the OCSP signer certificate
5720 or the issuer of the OCSP signer certificate. In the latter case the
5721 OCSP signer certificate must contain the OCSP signing extended key
5722 usage. This check is performed by attempting to match the OCSP
5723 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
5724 in the OCSP_CERTID structures of the response.
5727 *) Initial OCSP certificate verification added to OCSP_basic_verify()
5728 and related routines. This uses the standard OpenSSL certificate
5729 verify routines to perform initial checks (just CA validity) and
5730 to obtain the certificate chain. Then additional checks will be
5731 performed on the chain. Currently the root CA is checked to see
5732 if it is explicitly trusted for OCSP signing. This is used to set
5733 a root CA as a global signing root: that is any certificate that
5734 chains to that CA is an acceptable OCSP signing certificate.
5737 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5738 extensions from a separate configuration file.
5739 As when reading extensions from the main configuration file,
5740 the '-extensions ...' option may be used for specifying the
5742 [Massimiliano Pala <madwolf@comune.modena.it>]
5744 *) New OCSP utility. Allows OCSP requests to be generated or
5745 read. The request can be sent to a responder and the output
5746 parsed, outputed or printed in text form. Not complete yet:
5747 still needs to check the OCSP response validity.
5750 *) New subcommands for 'openssl ca':
5751 'openssl ca -status <serial>' prints the status of the cert with
5752 the given serial number (according to the index file).
5753 'openssl ca -updatedb' updates the expiry status of certificates
5755 [Massimiliano Pala <madwolf@comune.modena.it>]
5757 *) New '-newreq-nodes' command option to CA.pl. This is like
5758 '-newreq', but calls 'openssl req' with the '-nodes' option
5759 so that the resulting key is not encrypted.
5760 [Damien Miller <djm@mindrot.org>]
5762 *) New configuration for the GNU Hurd.
5763 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5765 *) Initial code to implement OCSP basic response verify. This
5766 is currently incomplete. Currently just finds the signer's
5767 certificate and verifies the signature on the response.
5770 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5771 value of OPENSSLDIR. This is available via the new '-d' option
5772 to 'openssl version', and is also included in 'openssl version -a'.
5775 *) Allowing defining memory allocation callbacks that will be given
5776 file name and line number information in additional arguments
5777 (a const char* and an int). The basic functionality remains, as
5778 well as the original possibility to just replace malloc(),
5779 realloc() and free() by functions that do not know about these
5780 additional arguments. To register and find out the current
5781 settings for extended allocation functions, the following
5782 functions are provided:
5784 CRYPTO_set_mem_ex_functions
5785 CRYPTO_set_locked_mem_ex_functions
5786 CRYPTO_get_mem_ex_functions
5787 CRYPTO_get_locked_mem_ex_functions
5789 These work the same way as CRYPTO_set_mem_functions and friends.
5790 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5791 extended allocation function is enabled.
5792 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5793 a conventional allocation function is enabled.
5794 [Richard Levitte, Bodo Moeller]
5796 *) Finish off removing the remaining LHASH function pointer casts.
5797 There should no longer be any prototype-casting required when using
5798 the LHASH abstraction, and any casts that remain are "bugs". See
5799 the callback types and macros at the head of lhash.h for details
5800 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5803 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5804 If /dev/[u]random devices are not available or do not return enough
5805 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5807 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5808 /etc/entropy will be queried once each in this sequence, quering stops
5809 when enough entropy was collected without querying more sockets.
5812 *) Change the Unix RAND_poll() variant to be able to poll several
5813 random devices, as specified by DEVRANDOM, until a sufficient amount
5814 of data has been collected. We spend at most 10 ms on each file
5815 (select timeout) and read in non-blocking mode. DEVRANDOM now
5816 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5817 (previously it was just the string "/dev/urandom"), so on typical
5818 platforms the 10 ms delay will never occur.
5819 Also separate out the Unix variant to its own file, rand_unix.c.
5820 For VMS, there's a currently-empty rand_vms.c.
5823 *) Move OCSP client related routines to ocsp_cl.c. These
5824 provide utility functions which an application needing
5825 to issue a request to an OCSP responder and analyse the
5826 response will typically need: as opposed to those which an
5827 OCSP responder itself would need which will be added later.
5829 OCSP_request_sign() signs an OCSP request with an API similar
5830 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5831 response. OCSP_response_get1_basic() extracts basic response
5832 from response. OCSP_resp_find_status(): finds and extracts status
5833 information from an OCSP_CERTID structure (which will be created
5834 when the request structure is built). These are built from lower
5835 level functions which work on OCSP_SINGLERESP structures but
5836 wont normally be used unless the application wishes to examine
5837 extensions in the OCSP response for example.
5839 Replace nonce routines with a pair of functions.
5840 OCSP_request_add1_nonce() adds a nonce value and optionally
5841 generates a random value. OCSP_check_nonce() checks the
5842 validity of the nonce in an OCSP response.
5845 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5846 This doesn't copy the supplied OCSP_CERTID and avoids the
5847 need to free up the newly created id. Change return type
5848 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5849 This can then be used to add extensions to the request.
5850 Deleted OCSP_request_new(), since most of its functionality
5851 is now in OCSP_REQUEST_new() (and the case insensitive name
5852 clash) apart from the ability to set the request name which
5853 will be added elsewhere.
5856 *) Update OCSP API. Remove obsolete extensions argument from
5857 various functions. Extensions are now handled using the new
5858 OCSP extension code. New simple OCSP HTTP function which
5859 can be used to send requests and parse the response.
5862 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5863 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5864 uses the special reorder version of SET OF to sort the attributes
5865 and reorder them to match the encoded order. This resolves a long
5866 standing problem: a verify on a PKCS7 structure just after signing
5867 it used to fail because the attribute order did not match the
5868 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5869 it uses the received order. This is necessary to tolerate some broken
5870 software that does not order SET OF. This is handled by encoding
5871 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5872 to produce the required SET OF.
5875 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5876 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5877 files to get correct declarations of the ASN.1 item variables.
5880 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5881 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5882 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5883 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5884 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5885 ASN1_ITEM and no wrapper functions.
5888 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5889 replace the old function pointer based I/O routines. Change most of
5890 the *_d2i_bio() and *_d2i_fp() functions to use these.
5893 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5894 lines, recognice more "algorithms" that can be deselected, and make
5895 it complain about algorithm deselection that isn't recognised.
5898 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5899 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5900 to use new functions. Add NO_ASN1_OLD which can be set to remove
5901 some old style ASN1 functions: this can be used to determine if old
5902 code will still work when these eventually go away.
5905 *) New extension functions for OCSP structures, these follow the
5906 same conventions as certificates and CRLs.
5909 *) New function X509V3_add1_i2d(). This automatically encodes and
5910 adds an extension. Its behaviour can be customised with various
5911 flags to append, replace or delete. Various wrappers added for
5912 certifcates and CRLs.
5915 *) Fix to avoid calling the underlying ASN1 print routine when
5916 an extension cannot be parsed. Correct a typo in the
5917 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5920 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5921 entries for variables.
5924 *) Add functionality to apps/openssl.c for detecting locking
5925 problems: As the program is single-threaded, all we have
5926 to do is register a locking callback using an array for
5927 storing which locks are currently held by the program.
5930 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5931 SSL_get_ex_data_X509_STORE_idx(), which is used in
5932 ssl_verify_cert_chain() and thus can be called at any time
5933 during TLS/SSL handshakes so that thread-safety is essential.
5934 Unfortunately, the ex_data design is not at all suited
5935 for multi-threaded use, so it probably should be abolished.
5938 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5939 [Broadcom, tweaked and integrated by Geoff Thorpe]
5941 *) Move common extension printing code to new function
5942 X509V3_print_extensions(). Reorganise OCSP print routines and
5943 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5946 *) New function X509_signature_print() to remove duplication in some
5950 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5951 set (this was treated exactly the same as SET OF previously). This
5952 is used to reorder the STACK representing the structure to match the
5953 encoding. This will be used to get round a problem where a PKCS7
5954 structure which was signed could not be verified because the STACK
5955 order did not reflect the encoded order.
5958 *) Reimplement the OCSP ASN1 module using the new code.
5961 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5962 for its ASN1 operations. The old style function pointers still exist
5963 for now but they will eventually go away.
5966 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5967 completely replaces the old ASN1 functionality with a table driven
5968 encoder and decoder which interprets an ASN1_ITEM structure describing
5969 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5970 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5971 has also been converted to the new form.
5974 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5975 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5976 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5977 for negative moduli.
5980 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5981 of not touching the result's sign bit.
5984 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5988 *) Changed the LHASH code to use prototypes for callbacks, and created
5989 macros to declare and implement thin (optionally static) functions
5990 that provide type-safety and avoid function pointer casting for the
5991 type-specific callbacks.
5994 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5996 [Veers Staats <staatsvr@asc.hpc.mil>,
5997 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5999 *) Reformat the FAQ so the different questions and answers can be divided
6000 in sections depending on the subject.
6003 *) Have the zlib compression code load ZLIB.DLL dynamically under
6007 *) New function BN_mod_sqrt for computing square roots modulo a prime
6008 (using the probabilistic Tonelli-Shanks algorithm unless
6009 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6010 be handled deterministically).
6011 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6013 *) Make BN_mod_inverse faster by explicitly handling small quotients
6014 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6015 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6018 *) New function BN_kronecker.
6021 *) Fix BN_gcd so that it works on negative inputs; the result is
6022 positive unless both parameters are zero.
6023 Previously something reasonably close to an infinite loop was
6024 possible because numbers could be growing instead of shrinking
6025 in the implementation of Euclid's algorithm.
6028 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6029 sign of the number in question.
6031 Fix BN_is_word(a,w) to work correctly for w == 0.
6033 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6034 because its test if the absolute value of 'a' equals 'w'.
6035 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6036 it exists mostly for use in the implementations of BN_is_zero(),
6037 BN_is_one(), and BN_is_word().
6040 *) New function BN_swap.
6043 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6044 the exponentiation functions are more likely to produce reasonable
6045 results on negative inputs.
6048 *) Change BN_mod_mul so that the result is always non-negative.
6049 Previously, it could be negative if one of the factors was negative;
6050 I don't think anyone really wanted that behaviour.
6053 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6054 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6055 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6056 and add new functions:
6065 BN_mod_lshift1_quick
6069 These functions always generate non-negative results.
6071 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6072 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6074 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6075 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6076 be reduced modulo m.
6077 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6080 The following entry accidentily appeared in the CHANGES file
6081 distributed with OpenSSL 0.9.7. The modifications described in
6082 it do *not* apply to OpenSSL 0.9.7.
6084 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6085 was actually never needed) and in BN_mul(). The removal in BN_mul()
6086 required a small change in bn_mul_part_recursive() and the addition
6087 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6088 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6089 bn_sub_words() and bn_add_words() except they take arrays with
6094 *) In 'openssl passwd', verify passwords read from the terminal
6095 unless the '-salt' option is used (which usually means that
6096 verification would just waste user's time since the resulting
6097 hash is going to be compared with some given password hash)
6098 or the new '-noverify' option is used.
6100 This is an incompatible change, but it does not affect
6101 non-interactive use of 'openssl passwd' (passwords on the command
6102 line, '-stdin' option, '-in ...' option) and thus should not
6106 *) Remove all references to RSAref, since there's no more need for it.
6109 *) Make DSO load along a path given through an environment variable
6110 (SHLIB_PATH) with shl_load().
6113 *) Constify the ENGINE code as a result of BIGNUM constification.
6114 Also constify the RSA code and most things related to it. In a
6115 few places, most notable in the depth of the ASN.1 code, ugly
6116 casts back to non-const were required (to be solved at a later
6120 *) Make it so the openssl application has all engines loaded by default.
6123 *) Constify the BIGNUM routines a little more.
6126 *) Add the following functions:
6128 ENGINE_load_cswift()
6130 ENGINE_load_atalla()
6132 ENGINE_load_builtin_engines()
6134 That way, an application can itself choose if external engines that
6135 are built-in in OpenSSL shall ever be used or not. The benefit is
6136 that applications won't have to be linked with libdl or other dso
6137 libraries unless it's really needed.
6139 Changed 'openssl engine' to load all engines on demand.
6140 Changed the engine header files to avoid the duplication of some
6141 declarations (they differed!).
6144 *) 'openssl engine' can now list capabilities.
6147 *) Better error reporting in 'openssl engine'.
6150 *) Never call load_dh_param(NULL) in s_server.
6153 *) Add engine application. It can currently list engines by name and
6154 identity, and test if they are actually available.
6157 *) Improve RPM specification file by forcing symbolic linking and making
6158 sure the installed documentation is also owned by root.root.
6159 [Damien Miller <djm@mindrot.org>]
6161 *) Give the OpenSSL applications more possibilities to make use of
6162 keys (public as well as private) handled by engines.
6165 *) Add OCSP code that comes from CertCo.
6168 *) Add VMS support for the Rijndael code.
6171 *) Added untested support for Nuron crypto accelerator.
6174 *) Add support for external cryptographic devices. This code was
6175 previously distributed separately as the "engine" branch.
6176 [Geoff Thorpe, Richard Levitte]
6178 *) Rework the filename-translation in the DSO code. It is now possible to
6179 have far greater control over how a "name" is turned into a filename
6180 depending on the operating environment and any oddities about the
6181 different shared library filenames on each system.
6184 *) Support threads on FreeBSD-elf in Configure.
6187 *) Fix for SHA1 assembly problem with MASM: it produces
6188 warnings about corrupt line number information when assembling
6189 with debugging information. This is caused by the overlapping
6191 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
6194 NCONF_get_number() has no error checking at all. As a replacement,
6195 NCONF_get_number_e() is defined (_e for "error checking") and is
6196 promoted strongly. The old NCONF_get_number is kept around for
6197 binary backward compatibility.
6198 Make it possible for methods to load from something other than a BIO,
6199 by providing a function pointer that is given a name instead of a BIO.
6200 For example, this could be used to load configuration data from an
6204 *) Fix for non blocking accept BIOs. Added new I/O special reason
6205 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
6206 with non blocking I/O was not possible because no retry code was
6207 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
6211 *) Added the beginnings of Rijndael support.
6214 *) Fix for bug in DirectoryString mask setting. Add support for
6215 X509_NAME_print_ex() in 'req' and X509_print_ex() function
6216 to allow certificate printing to more controllable, additional
6217 'certopt' option to 'x509' to allow new printing options to be
6221 *) Clean old EAY MD5 hack from e_os.h.
6224 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
6226 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6227 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6228 [Joe Orton, Steve Henson]
6230 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
6232 *) Fix additional bug revealed by the NISCC test suite:
6234 Stop bug triggering large recursion when presented with
6235 certain ASN.1 tags (CVE-2003-0851)
6238 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
6240 *) Fix various bugs revealed by running the NISCC test suite:
6242 Stop out of bounds reads in the ASN1 code when presented with
6243 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6245 If verify callback ignores invalid public key errors don't try to check
6246 certificate signature with the NULL public key.
6250 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6251 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6255 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6256 extra data after the compression methods not only for TLS 1.0
6257 but also for SSL 3.0 (as required by the specification).
6258 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6260 *) Change X509_certificate_type() to mark the key as exported/exportable
6261 when it's 512 *bits* long, not 512 bytes.
6264 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
6266 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6267 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6268 a protocol version number mismatch like a decryption error
6269 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6272 *) Turn on RSA blinding by default in the default implementation
6273 to avoid a timing attack. Applications that don't want it can call
6274 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6275 They would be ill-advised to do so in most cases.
6276 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6278 *) Change RSA blinding code so that it works when the PRNG is not
6279 seeded (in this case, the secret RSA exponent is abused as
6280 an unpredictable seed -- if it is not unpredictable, there
6281 is no point in blinding anyway). Make RSA blinding thread-safe
6282 by remembering the creator's thread ID in rsa->blinding and
6283 having all other threads use local one-time blinding factors
6284 (this requires more computation than sharing rsa->blinding, but
6285 avoids excessive locking; and if an RSA object is not shared
6286 between threads, blinding will still be very fast).
6289 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
6291 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6292 via timing by performing a MAC computation even if incorrrect
6293 block cipher padding has been found. This is a countermeasure
6294 against active attacks where the attacker has to distinguish
6295 between bad padding and a MAC verification error. (CVE-2003-0078)
6297 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6298 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6299 Martin Vuagnoux (EPFL, Ilion)]
6301 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
6303 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
6304 memory from it's contents. This is done with a counter that will
6305 place alternating values in each byte. This can be used to solve
6306 two issues: 1) the removal of calls to memset() by highly optimizing
6307 compilers, and 2) cleansing with other values than 0, since those can
6308 be read through on certain media, for example a swap space on disk.
6311 *) Bugfix: client side session caching did not work with external caching,
6312 because the session->cipher setting was not restored when reloading
6313 from the external cache. This problem was masked, when
6314 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
6315 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
6318 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
6319 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
6320 [Zeev Lieber <zeev-l@yahoo.com>]
6322 *) Undo an undocumented change introduced in 0.9.6e which caused
6323 repeated calls to OpenSSL_add_all_ciphers() and
6324 OpenSSL_add_all_digests() to be ignored, even after calling
6328 *) Change the default configuration reader to deal with last line not
6329 being properly terminated.
6332 *) Change X509_NAME_cmp() so it applies the special rules on handling
6333 DN values that are of type PrintableString, as well as RDNs of type
6334 emailAddress where the value has the type ia5String.
6335 [stefank@valicert.com via Richard Levitte]
6337 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
6338 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
6339 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
6340 the bitwise-OR of the two for use by the majority of applications
6341 wanting this behaviour, and update the docs. The documented
6342 behaviour and actual behaviour were inconsistent and had been
6343 changing anyway, so this is more a bug-fix than a behavioural
6345 [Geoff Thorpe, diagnosed by Nadav Har'El]
6347 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
6348 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
6351 *) Fix initialization code race conditions in
6352 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
6353 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
6354 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
6355 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
6356 ssl2_get_cipher_by_char(),
6357 ssl3_get_cipher_by_char().
6358 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
6360 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
6361 the cached sessions are flushed, as the remove_cb() might use ex_data
6362 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
6363 (see [openssl.org #212]).
6364 [Geoff Thorpe, Lutz Jaenicke]
6366 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
6367 length, instead of the encoding length to d2i_ASN1_OBJECT.
6370 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
6372 *) [In 0.9.6g-engine release:]
6373 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
6374 [Lynn Gazis <lgazis@rainbow.com>]
6376 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
6378 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
6379 and get fix the header length calculation.
6380 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
6381 Alon Kantor <alonk@checkpoint.com> (and others),
6384 *) Use proper error handling instead of 'assertions' in buffer
6385 overflow checks added in 0.9.6e. This prevents DoS (the
6386 assertions could call abort()).
6387 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
6389 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
6391 *) Add various sanity checks to asn1_get_length() to reject
6392 the ASN1 length bytes if they exceed sizeof(long), will appear
6393 negative or the content length exceeds the length of the
6395 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6397 *) Fix cipher selection routines: ciphers without encryption had no flags
6398 for the cipher strength set and where therefore not handled correctly
6399 by the selection routines (PR #130).
6402 *) Fix EVP_dsa_sha macro.
6406 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
6407 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
6408 that was added in OpenSSL 0.9.6d.
6410 As the countermeasure turned out to be incompatible with some
6411 broken SSL implementations, the new option is part of SSL_OP_ALL.
6412 SSL_OP_ALL is usually employed when compatibility with weird SSL
6413 implementations is desired (e.g. '-bugs' option to 's_client' and
6414 's_server'), so the new option is automatically set in many
6418 *) Changes in security patch:
6420 Changes marked "(CHATS)" were sponsored by the Defense Advanced
6421 Research Projects Agency (DARPA) and Air Force Research Laboratory,
6422 Air Force Materiel Command, USAF, under agreement number
6425 *) Add various sanity checks to asn1_get_length() to reject
6426 the ASN1 length bytes if they exceed sizeof(long), will appear
6427 negative or the content length exceeds the length of the
6428 supplied buffer. (CVE-2002-0659)
6429 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6431 *) Assertions for various potential buffer overflows, not known to
6433 [Ben Laurie (CHATS)]
6435 *) Various temporary buffers to hold ASCII versions of integers were
6436 too small for 64 bit platforms. (CVE-2002-0655)
6437 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6439 *) Remote buffer overflow in SSL3 protocol - an attacker could
6440 supply an oversized session ID to a client. (CVE-2002-0656)
6441 [Ben Laurie (CHATS)]
6443 *) Remote buffer overflow in SSL2 protocol - an attacker could
6444 supply an oversized client master key. (CVE-2002-0656)
6445 [Ben Laurie (CHATS)]
6447 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6449 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6450 encoded as NULL) with id-dsa-with-sha1.
6451 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6453 *) Check various X509_...() return values in apps/req.c.
6454 [Nils Larsch <nla@trustcenter.de>]
6456 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6457 an end-of-file condition would erronously be flagged, when the CRLF
6458 was just at the end of a processed block. The bug was discovered when
6459 processing data through a buffering memory BIO handing the data to a
6460 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6461 <ptsekov@syntrex.com> and Nedelcho Stanev.
6464 *) Implement a countermeasure against a vulnerability recently found
6465 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6466 before application data chunks to avoid the use of known IVs
6467 with data potentially chosen by the attacker.
6470 *) Fix length checks in ssl3_get_client_hello().
6473 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6474 to prevent ssl3_read_internal() from incorrectly assuming that
6475 ssl3_read_bytes() found application data while handshake
6476 processing was enabled when in fact s->s3->in_read_app_data was
6477 merely automatically cleared during the initial handshake.
6478 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6480 *) Fix object definitions for Private and Enterprise: they were not
6481 recognized in their shortname (=lowercase) representation. Extend
6482 obj_dat.pl to issue an error when using undefined keywords instead
6483 of silently ignoring the problem (Svenning Sorensen
6484 <sss@sss.dnsalias.net>).
6487 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6488 generators, i.e. generators other than 2 and 5. (Previously, the
6489 code did not properly initialise the 'add' and 'rem' values to
6490 BN_generate_prime().)
6492 In the new general case, we do not insist that 'generator' is
6493 actually a primitive root: This requirement is rather pointless;
6494 a generator of the order-q subgroup is just as good, if not
6498 *) Map new X509 verification errors to alerts. Discovered and submitted by
6499 Tom Wu <tom@arcot.com>.
6502 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6503 returning non-zero before the data has been completely received
6504 when using non-blocking I/O.
6505 [Bodo Moeller; problem pointed out by John Hughes]
6507 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6508 [Ben Laurie, Lutz Jaenicke]
6510 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6511 Yoram Zahavi <YoramZ@gilian.com>).
6514 *) Add information about CygWin 1.3 and on, and preserve proper
6515 configuration for the versions before that.
6516 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6518 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6519 check whether we deal with a copy of a session and do not delete from
6520 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6521 <izhar@checkpoint.com>.
6524 *) Do not store session data into the internal session cache, if it
6525 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6526 flag is set). Proposed by Aslam <aslam@funk.com>.
6529 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6533 *) [In 0.9.6d-engine release:]
6534 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6535 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6537 *) Add the configuration target linux-s390x.
6538 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6540 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6541 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6542 variable as an indication that a ClientHello message has been
6543 received. As the flag value will be lost between multiple
6544 invocations of ssl3_accept when using non-blocking I/O, the
6545 function may not be aware that a handshake has actually taken
6546 place, thus preventing a new session from being added to the
6549 To avoid this problem, we now set s->new_session to 2 instead of
6550 using a local variable.
6551 [Lutz Jaenicke, Bodo Moeller]
6553 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6554 if the SSL_R_LENGTH_MISMATCH error is detected.
6555 [Geoff Thorpe, Bodo Moeller]
6557 *) New 'shared_ldflag' column in Configure platform table.
6560 *) Fix EVP_CIPHER_mode macro.
6561 ["Dan S. Camper" <dan@bti.net>]
6563 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
6564 type, we must throw them away by setting rr->length to 0.
6565 [D P Chang <dpc@qualys.com>]
6567 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
6569 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
6570 <Dominikus.Scherkl@biodata.com>. (The previous implementation
6571 worked incorrectly for those cases where range = 10..._2 and
6572 3*range is two bits longer than range.)
6575 *) Only add signing time to PKCS7 structures if it is not already
6579 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
6580 OBJ_ld_ce should be OBJ_id_ce.
6581 Also some ip-pda OIDs in crypto/objects/objects.txt were
6582 incorrect (cf. RFC 3039).
6583 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
6585 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
6586 returns early because it has nothing to do.
6587 [Andy Schneider <andy.schneider@bjss.co.uk>]
6589 *) [In 0.9.6c-engine release:]
6590 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
6591 [Andy Schneider <andy.schneider@bjss.co.uk>]
6593 *) [In 0.9.6c-engine release:]
6594 Add support for Cryptographic Appliance's keyserver technology.
6595 (Use engine 'keyclient')
6596 [Cryptographic Appliances and Geoff Thorpe]
6598 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
6599 is called via tools/c89.sh because arguments have to be
6600 rearranged (all '-L' options must appear before the first object
6602 [Richard Shapiro <rshapiro@abinitio.com>]
6604 *) [In 0.9.6c-engine release:]
6605 Add support for Broadcom crypto accelerator cards, backported
6607 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
6609 *) [In 0.9.6c-engine release:]
6610 Add support for SureWare crypto accelerator cards from
6611 Baltimore Technologies. (Use engine 'sureware')
6612 [Baltimore Technologies and Mark Cox]
6614 *) [In 0.9.6c-engine release:]
6615 Add support for crypto accelerator cards from Accelerated
6616 Encryption Processing, www.aep.ie. (Use engine 'aep')
6617 [AEP Inc. and Mark Cox]
6619 *) Add a configuration entry for gcc on UnixWare.
6620 [Gary Benson <gbenson@redhat.com>]
6622 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
6623 messages are stored in a single piece (fixed-length part and
6624 variable-length part combined) and fix various bugs found on the way.
6627 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
6628 instead. BIO_gethostbyname() does not know what timeouts are
6629 appropriate, so entries would stay in cache even when they have
6631 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
6633 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
6634 faced with a pathologically small ClientHello fragment that does
6635 not contain client_version: Instead of aborting with an error,
6636 simply choose the highest available protocol version (i.e.,
6637 TLS 1.0 unless it is disabled). In practice, ClientHello
6638 messages are never sent like this, but this change gives us
6639 strictly correct behaviour at least for TLS.
6642 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
6643 never resets s->method to s->ctx->method when called from within
6644 one of the SSL handshake functions.
6645 [Bodo Moeller; problem pointed out by Niko Baric]
6647 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
6648 (sent using the client's version number) if client_version is
6649 smaller than the protocol version in use. Also change
6650 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
6651 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
6652 the client will at least see that alert.
6655 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
6659 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
6660 client receives HelloRequest while in a handshake.
6661 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
6663 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
6664 should end in 'break', not 'goto end' which circuments various
6665 cleanups done in state SSL_ST_OK. But session related stuff
6666 must be disabled for SSL_ST_OK in the case that we just sent a
6669 Also avoid some overhead by not calling ssl_init_wbio_buffer()
6670 before just sending a HelloRequest.
6671 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
6673 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
6674 reveal whether illegal block cipher padding was found or a MAC
6675 verification error occured. (Neither SSLerr() codes nor alerts
6676 are directly visible to potential attackers, but the information
6677 may leak via logfiles.)
6679 Similar changes are not required for the SSL 2.0 implementation
6680 because the number of padding bytes is sent in clear for SSL 2.0,
6681 and the extra bytes are just ignored. However ssl/s2_pkt.c
6682 failed to verify that the purported number of padding bytes is in
6686 *) Add OpenUNIX-8 support including shared libraries
6687 (Boyd Lynn Gerber <gerberb@zenez.com>).
6690 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
6691 'wristwatch attack' using huge encoding parameters (cf.
6692 James H. Manger's CRYPTO 2001 paper). Note that the
6693 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
6694 encoding parameters and hence was not vulnerable.
6697 *) BN_sqr() bug fix.
6698 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
6700 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
6701 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
6702 followed by modular reduction.
6703 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
6705 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
6706 equivalent based on BN_pseudo_rand() instead of BN_rand().
6709 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
6710 This function was broken, as the check for a new client hello message
6711 to handle SGC did not allow these large messages.
6712 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
6715 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
6718 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
6719 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
6722 *) Rework the configuration and shared library support for Tru64 Unix.
6723 The configuration part makes use of modern compiler features and
6724 still retains old compiler behavior for those that run older versions
6725 of the OS. The shared library support part includes a variant that
6726 uses the RPATH feature, and is available through the special
6727 configuration target "alpha-cc-rpath", which will never be selected
6729 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
6731 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
6732 with the same message size as in ssl3_get_certificate_request().
6733 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
6734 messages might inadvertently be reject as too long.
6735 [Petr Lampa <lampa@fee.vutbr.cz>]
6737 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6740 *) Modified SSL library such that the verify_callback that has been set
6741 specificly for an SSL object with SSL_set_verify() is actually being
6742 used. Before the change, a verify_callback set with this function was
6743 ignored and the verify_callback() set in the SSL_CTX at the time of
6744 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6745 to allow the necessary settings.
6748 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6749 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6750 done automatically (in contradiction to the requirements of the C
6751 standard). This made problems when used from OpenSSH.
6754 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6755 dh->length and always used
6757 BN_rand_range(priv_key, dh->p).
6759 BN_rand_range() is not necessary for Diffie-Hellman, and this
6760 specific range makes Diffie-Hellman unnecessarily inefficient if
6761 dh->length (recommended exponent length) is much smaller than the
6762 length of dh->p. We could use BN_rand_range() if the order of
6763 the subgroup was stored in the DH structure, but we only have
6768 BN_rand(priv_key, l, ...)
6770 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6776 RSA_eay_public_encrypt
6777 RSA_eay_private_decrypt
6778 RSA_eay_private_encrypt (signing)
6779 RSA_eay_public_decrypt (signature verification)
6781 (default implementations for RSA_public_encrypt,
6782 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6783 always reject numbers >= n.
6786 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6787 to synchronize access to 'locking_thread'. This is necessary on
6788 systems where access to 'locking_thread' (an 'unsigned long'
6789 variable) is not atomic.
6792 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6793 *before* setting the 'crypto_lock_rand' flag. The previous code had
6794 a race condition if 0 is a valid thread ID.
6795 [Travis Vitek <vitek@roguewave.com>]
6797 *) Add support for shared libraries under Irix.
6798 [Albert Chin-A-Young <china@thewrittenword.com>]
6800 *) Add configuration option to build on Linux on both big-endian and
6802 [Ralf Baechle <ralf@uni-koblenz.de>]
6804 *) Add the possibility to create shared libraries on HP-UX.
6807 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6809 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6810 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6811 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6812 PRNG state recovery was possible based on the output of
6813 one PRNG request appropriately sized to gain knowledge on
6814 'md' followed by enough consecutive 1-byte PRNG requests
6815 to traverse all of 'state'.
6817 1. When updating 'md_local' (the current thread's copy of 'md')
6818 during PRNG output generation, hash all of the previous
6819 'md_local' value, not just the half used for PRNG output.
6821 2. Make the number of bytes from 'state' included into the hash
6822 independent from the number of PRNG bytes requested.
6824 The first measure alone would be sufficient to avoid
6825 Markku-Juhani's attack. (Actually it had never occurred
6826 to me that the half of 'md_local' used for chaining was the
6827 half from which PRNG output bytes were taken -- I had always
6828 assumed that the secret half would be used.) The second
6829 measure makes sure that additional data from 'state' is never
6830 mixed into 'md_local' in small portions; this heuristically
6831 further strengthens the PRNG.
6834 *) Fix crypto/bn/asm/mips3.s.
6837 *) When only the key is given to "enc", the IV is undefined. Print out
6838 an error message in this case.
6841 *) Handle special case when X509_NAME is empty in X509 printing routines.
6844 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6845 positive and less than q.
6848 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6849 used: it isn't thread safe and the add_lock_callback should handle
6851 [Paul Rose <Paul.Rose@bridge.com>]
6853 *) Verify that incoming data obeys the block size in
6854 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6858 [Ulf Möller, Bodo Möller]
6860 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6861 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6862 when fixing the server behaviour for backwards-compatible 'client
6863 hello' messages. (Note that the attack is impractical against
6864 SSL 3.0 and TLS 1.0 anyway because length and version checking
6865 means that the probability of guessing a valid ciphertext is
6866 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6869 Before 0.9.5, the countermeasure (hide the error by generating a
6870 random 'decryption result') did not work properly because
6871 ERR_clear_error() was missing, meaning that SSL_get_error() would
6872 detect the supposedly ignored error.
6874 Both problems are now fixed.
6877 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6878 (previously it was 1024).
6881 *) Fix for compatibility mode trust settings: ignore trust settings
6882 unless some valid trust or reject settings are present.
6885 *) Fix for blowfish EVP: its a variable length cipher.
6888 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6889 parameters in DSA public key structures and return an error in the
6890 DSA routines if parameters are absent.
6893 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6894 in the current directory if neither $RANDFILE nor $HOME was set.
6895 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6896 caused some confusion to Windows users who haven't defined $HOME.
6897 Thus RAND_file_name() is changed again: e_os.h can define a
6898 DEFAULT_HOME, which will be used if $HOME is not set.
6899 For Windows, we use "C:"; on other platforms, we still require
6900 environment variables.
6902 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6903 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6904 having multiple threads call RAND_poll() concurrently.
6907 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6908 combination of a flag and a thread ID variable.
6909 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6910 flag), *other* threads can enter ssleay_add_bytes without obeying
6911 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6912 that they do not hold after the first thread unsets add_do_not_lock).
6915 *) Change bctest again: '-x' expressions are not available in all
6919 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6921 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6922 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6924 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6925 the default extension for executables, if any. Also, make the perl
6926 scripts that use symlink() to test if it really exists and use "cp"
6927 if it doesn't. All this made OpenSSL compilable and installable in
6931 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6932 If SEQUENCE is length is indefinite just set c->slen to the total
6933 amount of data available.
6934 [Steve Henson, reported by shige@FreeBSD.org]
6935 [This change does not apply to 0.9.7.]
6937 *) Change bctest to avoid here-documents inside command substitution
6938 (workaround for FreeBSD /bin/sh bug).
6939 For compatibility with Ultrix, avoid shell functions (introduced
6940 in the bctest version that searches along $PATH).
6943 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6944 with des_encrypt() defined on some operating systems, like Solaris
6948 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6949 On the Importance of Eliminating Errors in Cryptographic
6950 Computations, J. Cryptology 14 (2001) 2, 101-119,
6951 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6954 *) MIPS assembler BIGNUM division bug fix.
6957 *) Disabled incorrect Alpha assembler code.
6960 *) Fix PKCS#7 decode routines so they correctly update the length
6961 after reading an EOC for the EXPLICIT tag.
6963 [This change does not apply to 0.9.7.]
6965 *) Fix bug in PKCS#12 key generation routines. This was triggered
6966 if a 3DES key was generated with a 0 initial byte. Include
6967 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6968 (but broken) behaviour.
6971 *) Enhance bctest to search for a working bc along $PATH and print
6973 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6975 *) Fix memory leaks in err.c: free err_data string if necessary;
6976 don't write to the wrong index in ERR_set_error_data.
6979 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6983 *) Replace rdtsc with _emit statements for VC++ version 5.
6984 [Jeremy Cooper <jeremy@baymoo.org>]
6986 *) Make it possible to reuse SSLv2 sessions.
6989 *) In copy_email() check for >= 0 as a return value for
6990 X509_NAME_get_index_by_NID() since 0 is a valid index.
6991 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6993 *) Avoid coredump with unsupported or invalid public keys by checking if
6994 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6995 PKCS7_verify() fails with non detached data.
6998 *) Don't use getenv in library functions when run as setuid/setgid.
6999 New function OPENSSL_issetugid().
7002 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7003 due to incorrect handling of multi-threading:
7005 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7007 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7009 3. Count how many times MemCheck_off() has been called so that
7010 nested use can be treated correctly. This also avoids
7011 inband-signalling in the previous code (which relied on the
7012 assumption that thread ID 0 is impossible).
7015 *) Add "-rand" option also to s_client and s_server.
7018 *) Fix CPU detection on Irix 6.x.
7019 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7020 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7022 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7025 [This change does not apply to 0.9.7.]
7027 *) Use the cached encoding of an X509_NAME structure rather than
7028 copying it. This is apparently the reason for the libsafe "errors"
7029 but the code is actually correct.
7032 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7033 Bleichenbacher's DSA attack.
7034 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7035 to be set and top=0 forces the highest bit to be set; top=-1 is new
7036 and leaves the highest bit random.
7037 [Ulf Moeller, Bodo Moeller]
7039 *) In the NCONF_...-based implementations for CONF_... queries
7040 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7041 a temporary CONF structure with the data component set to NULL
7042 (which gives segmentation faults in lh_retrieve).
7043 Instead, use NULL for the CONF pointer in CONF_get_string and
7044 CONF_get_number (which may use environment variables) and directly
7045 return NULL from CONF_get_section.
7048 *) Fix potential buffer overrun for EBCDIC.
7051 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7052 keyUsage if basicConstraints absent for a CA.
7055 *) Make SMIME_write_PKCS7() write mail header values with a format that
7056 is more generally accepted (no spaces before the semicolon), since
7057 some programs can't parse those values properly otherwise. Also make
7058 sure BIO's that break lines after each write do not create invalid
7062 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7063 macros previously used would not encode an empty SEQUENCE OF
7064 and break the signature.
7066 [This change does not apply to 0.9.7.]
7068 *) Zero the premaster secret after deriving the master secret in
7072 *) Add some EVP_add_digest_alias registrations (as found in
7073 OpenSSL_add_all_digests()) to SSL_library_init()
7074 aka OpenSSL_add_ssl_algorithms(). This provides improved
7075 compatibility with peers using X.509 certificates
7076 with unconventional AlgorithmIdentifier OIDs.
7079 *) Fix for Irix with NO_ASM.
7080 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7082 *) ./config script fixes.
7083 [Ulf Moeller, Richard Levitte]
7085 *) Fix 'openssl passwd -1'.
7088 *) Change PKCS12_key_gen_asc() so it can cope with non null
7089 terminated strings whose length is passed in the passlen
7090 parameter, for example from PEM callbacks. This was done
7091 by adding an extra length parameter to asc2uni().
7092 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7094 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7095 call failed, free the DSA structure.
7098 *) Fix to uni2asc() to cope with zero length Unicode strings.
7099 These are present in some PKCS#12 files.
7102 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7103 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7104 when writing a 32767 byte record.
7105 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7107 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7108 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7110 (RSA objects have a reference count access to which is protected
7111 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7112 so they are meant to be shared between threads.)
7113 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7114 "Reddie, Steven" <Steven.Reddie@ca.com>]
7116 *) Fix a deadlock in CRYPTO_mem_leaks().
7119 *) Use better test patterns in bntest.
7122 *) rand_win.c fix for Borland C.
7125 *) BN_rshift bugfix for n == 0.
7128 *) Add a 'bctest' script that checks for some known 'bc' bugs
7129 so that 'make test' does not abort just because 'bc' is broken.
7132 *) Store verify_result within SSL_SESSION also for client side to
7133 avoid potential security hole. (Re-used sessions on the client side
7134 always resulted in verify_result==X509_V_OK, not using the original
7135 result of the server certificate verification.)
7138 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
7139 SSL3_RT_APPLICATION_DATA, return 0.
7140 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
7144 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
7145 releases, have been re-implemented by renaming the previous
7146 implementations of ssl2_read and ssl3_read to ssl2_read_internal
7147 and ssl3_read_internal, respectively, and adding 'peek' parameters
7148 to them. The new ssl[23]_{read,peek} functions are calls to
7149 ssl[23]_read_internal with the 'peek' flag set appropriately.
7150 A 'peek' parameter has also been added to ssl3_read_bytes, which
7151 does the actual work for ssl3_read_internal.
7154 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
7155 the method-specific "init()" handler. Also clean up ex_data after
7156 calling the method-specific "finish()" handler. Previously, this was
7157 happening the other way round.
7160 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
7161 The previous value, 12, was not always sufficient for BN_mod_exp().
7164 *) Make sure that shared libraries get the internal name engine with
7165 the full version number and not just 0. This should mark the
7166 shared libraries as not backward compatible. Of course, this should
7167 be changed again when we can guarantee backward binary compatibility.
7170 *) Fix typo in get_cert_by_subject() in by_dir.c
7171 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
7173 *) Rework the system to generate shared libraries:
7175 - Make note of the expected extension for the shared libraries and
7176 if there is a need for symbolic links from for example libcrypto.so.0
7177 to libcrypto.so.0.9.7. There is extended info in Configure for
7180 - Make as few rebuilds of the shared libraries as possible.
7182 - Still avoid linking the OpenSSL programs with the shared libraries.
7184 - When installing, install the shared libraries separately from the
7188 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
7190 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
7191 and not in SSL_clear because the latter is also used by the
7192 accept/connect functions; previously, the settings made by
7193 SSL_set_read_ahead would be lost during the handshake.
7194 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
7196 *) Correct util/mkdef.pl to be selective about disabled algorithms.
7197 Previously, it would create entries for disableed algorithms no
7201 *) Added several new manual pages for SSL_* function.
7204 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
7206 *) In ssl23_get_client_hello, generate an error message when faced
7207 with an initial SSL 3.0/TLS record that is too small to contain the
7208 first two bytes of the ClientHello message, i.e. client_version.
7209 (Note that this is a pathologic case that probably has never happened
7210 in real life.) The previous approach was to use the version number
7211 from the record header as a substitute; but our protocol choice
7212 should not depend on that one because it is not authenticated
7213 by the Finished messages.
7216 *) More robust randomness gathering functions for Windows.
7217 [Jeffrey Altman <jaltman@columbia.edu>]
7219 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
7220 not set then we don't setup the error code for issuer check errors
7221 to avoid possibly overwriting other errors which the callback does
7222 handle. If an application does set the flag then we assume it knows
7223 what it is doing and can handle the new informational codes
7227 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
7228 a general "ANY" type, as such it should be able to decode anything
7229 including tagged types. However it didn't check the class so it would
7230 wrongly interpret tagged types in the same way as their universal
7231 counterpart and unknown types were just rejected. Changed so that the
7232 tagged and unknown types are handled in the same way as a SEQUENCE:
7233 that is the encoding is stored intact. There is also a new type
7234 "V_ASN1_OTHER" which is used when the class is not universal, in this
7235 case we have no idea what the actual type is so we just lump them all
7239 *) On VMS, stdout may very well lead to a file that is written to
7240 in a record-oriented fashion. That means that every write() will
7241 write a separate record, which will be read separately by the
7242 programs trying to read from it. This can be very confusing.
7244 The solution is to put a BIO filter in the way that will buffer
7245 text until a linefeed is reached, and then write everything a
7246 line at a time, so every record written will be an actual line,
7247 not chunks of lines and not (usually doesn't happen, but I've
7248 seen it once) several lines in one record. BIO_f_linebuffer() is
7251 Currently, it's a VMS-only method, because that's where it has
7252 been tested well enough.
7255 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
7256 it can return incorrect results.
7257 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
7258 but it was in 0.9.6-beta[12].)
7261 *) Disable the check for content being present when verifying detached
7262 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
7263 include zero length content when signing messages.
7266 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
7267 BIO_ctrl (for BIO pairs).
7270 *) Add DSO method for VMS.
7273 *) Bug fix: Montgomery multiplication could produce results with the
7277 *) Add RPM specification openssl.spec and modify it to build three
7278 packages. The default package contains applications, application
7279 documentation and run-time libraries. The devel package contains
7280 include files, static libraries and function documentation. The
7281 doc package contains the contents of the doc directory. The original
7282 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
7285 *) Add a large number of documentation files for many SSL routines.
7286 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7288 *) Add a configuration entry for Sony News 4.
7289 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
7291 *) Don't set the two most significant bits to one when generating a
7292 random number < q in the DSA library.
7295 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
7296 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
7297 the underlying transport is blocking) if a handshake took place.
7298 (The default behaviour is needed by applications such as s_client
7299 and s_server that use select() to determine when to use SSL_read;
7300 but for applications that know in advance when to expect data, it
7301 just makes things more complicated.)
7304 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
7308 *) Add a few more EBCDIC conditionals that make `req' and `x509'
7309 work better on such systems.
7310 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7312 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
7313 Update PKCS12_parse() so it copies the friendlyName and the
7314 keyid to the certificates aux info.
7317 *) Fix bug in PKCS7_verify() which caused an infinite loop
7318 if there was more than one signature.
7319 [Sven Uszpelkat <su@celocom.de>]
7321 *) Major change in util/mkdef.pl to include extra information
7322 about each symbol, as well as presentig variables as well
7323 as functions. This change means that there's n more need
7324 to rebuild the .num files when some algorithms are excluded.
7327 *) Allow the verify time to be set by an application,
7328 rather than always using the current time.
7331 *) Phase 2 verify code reorganisation. The certificate
7332 verify code now looks up an issuer certificate by a
7333 number of criteria: subject name, authority key id
7334 and key usage. It also verifies self signed certificates
7335 by the same criteria. The main comparison function is
7336 X509_check_issued() which performs these checks.
7338 Lot of changes were necessary in order to support this
7339 without completely rewriting the lookup code.
7341 Authority and subject key identifier are now cached.
7343 The LHASH 'certs' is X509_STORE has now been replaced
7344 by a STACK_OF(X509_OBJECT). This is mainly because an
7345 LHASH can't store or retrieve multiple objects with
7346 the same hash value.
7348 As a result various functions (which were all internal
7349 use only) have changed to handle the new X509_STORE
7350 structure. This will break anything that messed round
7351 with X509_STORE internally.
7353 The functions X509_STORE_add_cert() now checks for an
7354 exact match, rather than just subject name.
7356 The X509_STORE API doesn't directly support the retrieval
7357 of multiple certificates matching a given criteria, however
7358 this can be worked round by performing a lookup first
7359 (which will fill the cache with candidate certificates)
7360 and then examining the cache for matches. This is probably
7361 the best we can do without throwing out X509_LOOKUP
7362 entirely (maybe later...).
7364 The X509_VERIFY_CTX structure has been enhanced considerably.
7366 All certificate lookup operations now go via a get_issuer()
7367 callback. Although this currently uses an X509_STORE it
7368 can be replaced by custom lookups. This is a simple way
7369 to bypass the X509_STORE hackery necessary to make this
7370 work and makes it possible to use more efficient techniques
7371 in future. A very simple version which uses a simple
7372 STACK for its trusted certificate store is also provided
7373 using X509_STORE_CTX_trusted_stack().
7375 The verify_cb() and verify() callbacks now have equivalents
7376 in the X509_STORE_CTX structure.
7378 X509_STORE_CTX also has a 'flags' field which can be used
7379 to customise the verify behaviour.
7382 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
7383 excludes S/MIME capabilities.
7386 *) When a certificate request is read in keep a copy of the
7387 original encoding of the signed data and use it when outputing
7388 again. Signatures then use the original encoding rather than
7389 a decoded, encoded version which may cause problems if the
7390 request is improperly encoded.
7393 *) For consistency with other BIO_puts implementations, call
7394 buffer_write(b, ...) directly in buffer_puts instead of calling
7397 In BIO_puts, increment b->num_write as in BIO_write.
7398 [Peter.Sylvester@EdelWeb.fr]
7400 *) Fix BN_mul_word for the case where the word is 0. (We have to use
7401 BN_zero, we may not return a BIGNUM with an array consisting of
7405 *) Avoid calling abort() from within the library when problems are
7406 detected, except if preprocessor symbols have been defined
7407 (such as REF_CHECK, BN_DEBUG etc.).
7410 *) New openssl application 'rsautl'. This utility can be
7411 used for low level RSA operations. DER public key
7412 BIO/fp routines also added.
7415 *) New Configure entry and patches for compiling on QNX 4.
7416 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
7418 *) A demo state-machine implementation was sponsored by
7419 Nuron (http://www.nuron.com/) and is now available in
7420 demos/state_machine.
7423 *) New options added to the 'dgst' utility for signature
7424 generation and verification.
7427 *) Unrecognized PKCS#7 content types are now handled via a
7428 catch all ASN1_TYPE structure. This allows unsupported
7429 types to be stored as a "blob" and an application can
7430 encode and decode it manually.
7433 *) Fix various signed/unsigned issues to make a_strex.c
7435 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7437 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7438 length if passed a buffer. ASN1_INTEGER_to_BN failed
7439 if passed a NULL BN and its argument was negative.
7440 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7442 *) Modification to PKCS#7 encoding routines to output definite
7443 length encoding. Since currently the whole structures are in
7444 memory there's not real point in using indefinite length
7445 constructed encoding. However if OpenSSL is compiled with
7446 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7449 *) Added BIO_vprintf() and BIO_vsnprintf().
7452 *) Added more prefixes to parse for in the the strings written
7453 through a logging bio, to cover all the levels that are available
7454 through syslog. The prefixes are now:
7456 PANIC, EMERG, EMR => LOG_EMERG
7457 ALERT, ALR => LOG_ALERT
7458 CRIT, CRI => LOG_CRIT
7459 ERROR, ERR => LOG_ERR
7460 WARNING, WARN, WAR => LOG_WARNING
7461 NOTICE, NOTE, NOT => LOG_NOTICE
7462 INFO, INF => LOG_INFO
7463 DEBUG, DBG => LOG_DEBUG
7465 and as before, if none of those prefixes are present at the
7466 beginning of the string, LOG_ERR is chosen.
7468 On Win32, the LOG_* levels are mapped according to this:
7470 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7471 LOG_WARNING => EVENTLOG_WARNING_TYPE
7472 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7476 *) Made it possible to reconfigure with just the configuration
7477 argument "reconf" or "reconfigure". The command line arguments
7478 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7479 and are retrieved from there when reconfiguring.
7483 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7485 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7488 *) The obj_dat.pl script was messing up the sorting of object
7489 names. The reason was that it compared the quoted version
7490 of strings as a result "OCSP" > "OCSP Signing" because
7491 " > SPACE. Changed script to store unquoted versions of
7492 names and add quotes on output. It was also omitting some
7493 names from the lookup table if they were given a default
7494 value (that is if SN is missing it is given the same
7495 value as LN and vice versa), these are now added on the
7496 grounds that if an object has a name we should be able to
7497 look it up. Finally added warning output when duplicate
7498 short or long names are found.
7501 *) Changes needed for Tandem NSK.
7502 [Scott Uroff <scott@xypro.com>]
7504 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7505 RSA_padding_check_SSLv23(), special padding was never detected
7506 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7507 version rollback attacks was not effective.
7509 In s23_clnt.c, don't use special rollback-attack detection padding
7510 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7511 client; similarly, in s23_srvr.c, don't do the rollback check if
7512 SSL 2.0 is the only protocol enabled in the server.
7515 *) Make it possible to get hexdumps of unprintable data with 'openssl
7516 asn1parse'. By implication, the functions ASN1_parse_dump() and
7517 BIO_dump_indent() are added.
7520 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7521 these print out strings and name structures based on various
7522 flags including RFC2253 support and proper handling of
7523 multibyte characters. Added options to the 'x509' utility
7524 to allow the various flags to be set.
7527 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7528 Also change the functions X509_cmp_current_time() and
7529 X509_gmtime_adj() work with an ASN1_TIME structure,
7530 this will enable certificates using GeneralizedTime in validity
7531 dates to be checked.
7534 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7535 negative public key encodings) on by default,
7536 NO_NEG_PUBKEY_BUG can be set to disable it.
7539 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7540 content octets. An i2c_ASN1_OBJECT is unnecessary because
7541 the encoding can be trivially obtained from the structure.
7544 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7545 not read locks (CRYPTO_r_[un]lock).
7548 *) A first attempt at creating official support for shared
7549 libraries through configuration. I've kept it so the
7550 default is static libraries only, and the OpenSSL programs
7551 are always statically linked for now, but there are
7552 preparations for dynamic linking in place.
7553 This has been tested on Linux and Tru64.
7556 *) Randomness polling function for Win9x, as described in:
7557 Peter Gutmann, Software Generation of Practically Strong
7561 *) Fix so PRNG is seeded in req if using an already existing
7565 *) New options to smime application. -inform and -outform
7566 allow alternative formats for the S/MIME message including
7567 PEM and DER. The -content option allows the content to be
7568 specified separately. This should allow things like Netscape
7569 form signing output easier to verify.
7572 *) Fix the ASN1 encoding of tags using the 'long form'.
7575 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
7576 STRING types. These convert content octets to and from the
7577 underlying type. The actual tag and length octets are
7578 already assumed to have been read in and checked. These
7579 are needed because all other string types have virtually
7580 identical handling apart from the tag. By having versions
7581 of the ASN1 functions that just operate on content octets
7582 IMPLICIT tagging can be handled properly. It also allows
7583 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
7584 and ASN1_INTEGER are identical apart from the tag.
7587 *) Change the handling of OID objects as follows:
7589 - New object identifiers are inserted in objects.txt, following
7590 the syntax given in objects.README.
7591 - objects.pl is used to process obj_mac.num and create a new
7593 - obj_dat.pl is used to create a new obj_dat.h, using the data in
7596 This is currently kind of a hack, and the perl code in objects.pl
7597 isn't very elegant, but it works as I intended. The simplest way
7598 to check that it worked correctly is to look in obj_dat.h and
7599 check the array nid_objs and make sure the objects haven't moved
7600 around (this is important!). Additions are OK, as well as
7601 consistent name changes.
7604 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
7607 *) Addition of the command line parameter '-rand file' to 'openssl req'.
7608 The given file adds to whatever has already been seeded into the
7609 random pool through the RANDFILE configuration file option or
7610 environment variable, or the default random state file.
7613 *) mkstack.pl now sorts each macro group into lexical order.
7614 Previously the output order depended on the order the files
7615 appeared in the directory, resulting in needless rewriting
7619 *) Patches to make OpenSSL compile under Win32 again. Mostly
7620 work arounds for the VC++ problem that it treats func() as
7621 func(void). Also stripped out the parts of mkdef.pl that
7622 added extra typesafe functions: these no longer exist.
7625 *) Reorganisation of the stack code. The macros are now all
7626 collected in safestack.h . Each macro is defined in terms of
7627 a "stack macro" of the form SKM_<name>(type, a, b). The
7628 DEBUG_SAFESTACK is now handled in terms of function casts,
7629 this has the advantage of retaining type safety without the
7630 use of additional functions. If DEBUG_SAFESTACK is not defined
7631 then the non typesafe macros are used instead. Also modified the
7632 mkstack.pl script to handle the new form. Needs testing to see
7633 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
7634 the default if no major problems. Similar behaviour for ASN1_SET_OF
7635 and PKCS12_STACK_OF.
7638 *) When some versions of IIS use the 'NET' form of private key the
7639 key derivation algorithm is different. Normally MD5(password) is
7640 used as a 128 bit RC4 key. In the modified case
7641 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
7642 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
7643 as the old Netscape_RSA functions except they have an additional
7644 'sgckey' parameter which uses the modified algorithm. Also added
7645 an -sgckey command line option to the rsa utility. Thanks to
7646 Adrian Peck <bertie@ncipher.com> for posting details of the modified
7647 algorithm to openssl-dev.
7650 *) The evp_local.h macros were using 'c.##kname' which resulted in
7651 invalid expansion on some systems (SCO 5.0.5 for example).
7652 Corrected to 'c.kname'.
7653 [Phillip Porch <root@theporch.com>]
7655 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
7656 a STACK of email addresses from a certificate or request, these look
7657 in the subject name and the subject alternative name extensions and
7658 omit any duplicate addresses.
7661 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
7662 This makes DSA verification about 2 % faster.
7665 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
7666 (meaning that now 2^5 values will be precomputed, which is only 4 KB
7667 plus overhead for 1024 bit moduli).
7668 This makes exponentiations about 0.5 % faster for 1024 bit
7669 exponents (as measured by "openssl speed rsa2048").
7672 *) Rename memory handling macros to avoid conflicts with other
7674 Malloc => OPENSSL_malloc
7675 Malloc_locked => OPENSSL_malloc_locked
7676 Realloc => OPENSSL_realloc
7677 Free => OPENSSL_free
7680 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
7681 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
7684 *) CygWin32 support.
7685 [John Jarvie <jjarvie@newsguy.com>]
7687 *) The type-safe stack code has been rejigged. It is now only compiled
7688 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
7689 by default all type-specific stack functions are "#define"d back to
7690 standard stack functions. This results in more streamlined output
7691 but retains the type-safety checking possibilities of the original
7695 *) The STACK code has been cleaned up, and certain type declarations
7696 that didn't make a lot of sense have been brought in line. This has
7697 also involved a cleanup of sorts in safestack.h to more correctly
7698 map type-safe stack functions onto their plain stack counterparts.
7699 This work has also resulted in a variety of "const"ifications of
7700 lots of the code, especially "_cmp" operations which should normally
7701 be prototyped with "const" parameters anyway.
7704 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
7705 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
7706 (The PRNG state consists of two parts, the large pool 'state' and 'md',
7707 where all of 'md' is used each time the PRNG is used, but 'state'
7708 is used only indexed by a cyclic counter. As entropy may not be
7709 well distributed from the beginning, 'md' is important as a
7710 chaining variable. However, the output function chains only half
7711 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
7712 all of 'md', and seeding with STATE_SIZE dummy bytes will result
7713 in all of 'state' being rewritten, with the new values depending
7714 on virtually all of 'md'. This overcomes the 80 bit limitation.)
7717 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
7718 the handshake is continued after ssl_verify_cert_chain();
7719 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
7720 can lead to 'unexplainable' connection aborts later.
7721 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
7723 *) Major EVP API cipher revision.
7724 Add hooks for extra EVP features. This allows various cipher
7725 parameters to be set in the EVP interface. Support added for variable
7726 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
7727 setting of RC2 and RC5 parameters.
7729 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
7732 Remove lots of duplicated code from the EVP library. For example *every*
7733 cipher init() function handles the 'iv' in the same way according to the
7734 cipher mode. They also all do nothing if the 'key' parameter is NULL and
7735 for CFB and OFB modes they zero ctx->num.
7737 New functionality allows removal of S/MIME code RC2 hack.
7739 Most of the routines have the same form and so can be declared in terms
7742 By shifting this to the top level EVP_CipherInit() it can be removed from
7743 all individual ciphers. If the cipher wants to handle IVs or keys
7744 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7747 Change lots of functions like EVP_EncryptUpdate() to now return a
7748 value: although software versions of the algorithms cannot fail
7749 any installed hardware versions can.
7752 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7753 this option is set, tolerate broken clients that send the negotiated
7754 protocol version number instead of the requested protocol version
7758 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7759 i.e. non-zero for export ciphersuites, zero otherwise.
7760 Previous versions had this flag inverted, inconsistent with
7761 rsa_tmp_cb (..._TMP_RSA_CB).
7762 [Bodo Moeller; problem reported by Amit Chopra]
7764 *) Add missing DSA library text string. Work around for some IIS
7765 key files with invalid SEQUENCE encoding.
7768 *) Add a document (doc/standards.txt) that list all kinds of standards
7769 and so on that are implemented in OpenSSL.
7772 *) Enhance c_rehash script. Old version would mishandle certificates
7773 with the same subject name hash and wouldn't handle CRLs at all.
7774 Added -fingerprint option to crl utility, to support new c_rehash
7778 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7781 *) Fix for SSL server purpose checking. Server checking was
7782 rejecting certificates which had extended key usage present
7783 but no ssl client purpose.
7784 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7786 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7787 is a little unclear about how a blank password is handled.
7788 Since the password in encoded as a BMPString with terminating
7789 double NULL a zero length password would end up as just the
7790 double NULL. However no password at all is different and is
7791 handled differently in the PKCS#12 key generation code. NS
7792 treats a blank password as zero length. MSIE treats it as no
7793 password on export: but it will try both on import. We now do
7794 the same: PKCS12_parse() tries zero length and no password if
7795 the password is set to "" or NULL (NULL is now a valid password:
7796 it wasn't before) as does the pkcs12 application.
7799 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7800 perror when PEM_read_bio_X509_REQ fails, the error message must
7801 be obtained from the error queue.
7804 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7805 it in ERR_remove_state if appropriate, and change ERR_get_state
7806 accordingly to avoid race conditions (this is necessary because
7807 thread_hash is no longer constant once set).
7810 *) Bugfix for linux-elf makefile.one.
7813 *) RSA_get_default_method() will now cause a default
7814 RSA_METHOD to be chosen if one doesn't exist already.
7815 Previously this was only set during a call to RSA_new()
7816 or RSA_new_method(NULL) meaning it was possible for
7817 RSA_get_default_method() to return NULL.
7820 *) Added native name translation to the existing DSO code
7821 that will convert (if the flag to do so is set) filenames
7822 that are sufficiently small and have no path information
7823 into a canonical native form. Eg. "blah" converted to
7824 "libblah.so" or "blah.dll" etc.
7827 *) New function ERR_error_string_n(e, buf, len) which is like
7828 ERR_error_string(e, buf), but writes at most 'len' bytes
7829 including the 0 terminator. For ERR_error_string_n, 'buf'
7831 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7833 *) CONF library reworked to become more general. A new CONF
7834 configuration file reader "class" is implemented as well as a
7835 new functions (NCONF_*, for "New CONF") to handle it. The now
7836 old CONF_* functions are still there, but are reimplemented to
7837 work in terms of the new functions. Also, a set of functions
7838 to handle the internal storage of the configuration data is
7839 provided to make it easier to write new configuration file
7840 reader "classes" (I can definitely see something reading a
7841 configuration file in XML format, for example), called _CONF_*,
7842 or "the configuration storage API"...
7844 The new configuration file reading functions are:
7846 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7847 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7849 NCONF_default, NCONF_WIN32
7851 NCONF_dump_fp, NCONF_dump_bio
7853 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7854 NCONF_new creates a new CONF object. This works in the same way
7855 as other interfaces in OpenSSL, like the BIO interface.
7856 NCONF_dump_* dump the internal storage of the configuration file,
7857 which is useful for debugging. All other functions take the same
7858 arguments as the old CONF_* functions wth the exception of the
7859 first that must be a `CONF *' instead of a `LHASH *'.
7861 To make it easer to use the new classes with the old CONF_* functions,
7862 the function CONF_set_default_method is provided.
7865 *) Add '-tls1' option to 'openssl ciphers', which was already
7866 mentioned in the documentation but had not been implemented.
7867 (This option is not yet really useful because even the additional
7868 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7871 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7872 OpenSSL-based applications) load shared libraries and bind to
7873 them in a portable way.
7874 [Geoff Thorpe, with contributions from Richard Levitte]
7876 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7878 *) Make sure _lrotl and _lrotr are only used with MSVC.
7880 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7881 (the default implementation of RAND_status).
7883 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7884 to '-clrext' (= clear extensions), as intended and documented.
7885 [Bodo Moeller; inconsistency pointed out by Michael Attili
7888 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7889 was larger than the MD block size.
7890 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7892 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7893 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7894 using the passed key: if the passed key was a private key the result
7895 of X509_print(), for example, would be to print out all the private key
7899 *) des_quad_cksum() byte order bug fix.
7900 [Ulf Möller, using the problem description in krb4-0.9.7, where
7901 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7903 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7905 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7907 *) For easily testing in shell scripts whether some command
7908 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7909 returns with exit code 0 iff no command of the given name is available.
7910 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7911 the output goes to stdout and nothing is printed to stderr.
7912 Additional arguments are always ignored.
7914 Since for each cipher there is a command of the same name,
7915 the 'no-cipher' compilation switches can be tested this way.
7917 ('openssl no-XXX' is not able to detect pseudo-commands such
7918 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7921 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7924 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7925 is set; it will be thrown away anyway because each handshake creates
7927 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7928 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7929 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7930 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7933 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7934 'Q' and 'R' lose their special meanings (quit/renegotiate).
7935 This is part of what -quiet does; unlike -quiet, -ign_eof
7936 does not suppress any output.
7939 *) Add compatibility options to the purpose and trust code. The
7940 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7941 accepts a certificate or CA, this was the previous behaviour,
7942 with all the associated security issues.
7944 X509_TRUST_COMPAT is the old trust behaviour: only and
7945 automatically trust self signed roots in certificate store. A
7946 new trust setting X509_TRUST_DEFAULT is used to specify that
7947 a purpose has no associated trust setting and it should instead
7948 use the value in the default purpose.
7951 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7952 and fix a memory leak.
7955 *) In util/mkerr.pl (which implements 'make errors'), preserve
7956 reason strings from the previous version of the .c file, as
7957 the default to have only downcase letters (and digits) in
7958 automatically generated reasons codes is not always appropriate.
7961 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7962 using strerror. Previously, ERR_reason_error_string() returned
7963 library names as reason strings for SYSerr; but SYSerr is a special
7964 case where small numbers are errno values, not library numbers.
7967 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7968 converts DSA parameters into DH parameters. (When creating parameters,
7969 DSA_generate_parameters is used.)
7972 *) Include 'length' (recommended exponent length) in C code generated
7973 by 'openssl dhparam -C'.
7976 *) The second argument to set_label in perlasm was already being used
7977 so couldn't be used as a "file scope" flag. Moved to third argument
7981 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7982 instead of RAND_bytes for encryption IVs and salts.
7985 *) Include RAND_status() into RAND_METHOD instead of implementing
7986 it only for md_rand.c Otherwise replacing the PRNG by calling
7987 RAND_set_rand_method would be impossible.
7990 *) Don't let DSA_generate_key() enter an infinite loop if the random
7991 number generation fails.
7994 *) New 'rand' application for creating pseudo-random output.
7997 *) Added configuration support for Linux/IA64
7998 [Rolf Haberrecker <rolf@suse.de>]
8000 *) Assembler module support for Mingw32.
8003 *) Shared library support for HPUX (in shlib/).
8004 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8006 *) Shared library support for Solaris gcc.
8007 [Lutz Behnke <behnke@trustcenter.de>]
8009 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8011 *) PKCS7_encrypt() was adding text MIME headers twice because they
8012 were added manually and by SMIME_crlf_copy().
8015 *) In bntest.c don't call BN_rand with zero bits argument.
8016 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8018 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8019 case was implemented. This caused BN_div_recp() to fail occasionally.
8022 *) Add an optional second argument to the set_label() in the perl
8023 assembly language builder. If this argument exists and is set
8024 to 1 it signals that the assembler should use a symbol whose
8025 scope is the entire file, not just the current function. This
8026 is needed with MASM which uses the format label:: for this scope.
8027 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8029 *) Change the ASN1 types so they are typedefs by default. Before
8030 almost all types were #define'd to ASN1_STRING which was causing
8031 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8035 *) Change names of new functions to the new get1/get0 naming
8036 convention: After 'get1', the caller owns a reference count
8037 and has to call ..._free; 'get0' returns a pointer to some
8038 data structure without incrementing reference counters.
8039 (Some of the existing 'get' functions increment a reference
8040 counter, some don't.)
8041 Similarly, 'set1' and 'add1' functions increase reference
8042 counters or duplicate objects.
8045 *) Allow for the possibility of temp RSA key generation failure:
8046 the code used to assume it always worked and crashed on failure.
8049 *) Fix potential buffer overrun problem in BIO_printf().
8050 [Ulf Möller, using public domain code by Patrick Powell; problem
8051 pointed out by David Sacerdote <das33@cornell.edu>]
8053 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8054 RAND_egd() and RAND_status(). In the command line application,
8055 the EGD socket can be specified like a seed file using RANDFILE
8059 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8060 Some CAs (e.g. Verisign) distribute certificates in this form.
8063 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8064 list to exclude them. This means that no special compilation option
8065 is needed to use anonymous DH: it just needs to be included in the
8069 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8070 EVP_MD_type. The old functionality is available in a new macro called
8071 EVP_MD_md(). Change code that uses it and update docs.
8074 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8075 where the 'void *' argument is replaced by a function pointer argument.
8076 Previously 'void *' was abused to point to functions, which works on
8077 many platforms, but is not correct. As these functions are usually
8078 called by macros defined in OpenSSL header files, most source code
8079 should work without changes.
8082 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8083 sections with information on -D... compiler switches used for
8084 compiling the library so that applications can see them. To enable
8085 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8086 must be defined. E.g.,
8087 #define OPENSSL_ALGORITHM_DEFINES
8088 #include <openssl/opensslconf.h>
8089 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8090 [Richard Levitte, Ulf and Bodo Möller]
8092 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8096 *) Change the 'other' type in certificate aux info to a STACK_OF
8097 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8098 the required ASN1 format: arbitrary types determined by an OID.
8101 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8102 argument to 'req'. This is not because the function is newer or
8103 better than others it just uses the work 'NEW' in the certificate
8104 request header lines. Some software needs this.
8107 *) Reorganise password command line arguments: now passwords can be
8108 obtained from various sources. Delete the PEM_cb function and make
8109 it the default behaviour: i.e. if the callback is NULL and the
8110 usrdata argument is not NULL interpret it as a null terminated pass
8111 phrase. If usrdata and the callback are NULL then the pass phrase
8112 is prompted for as usual.
8115 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8116 the support is automatically enabled. The resulting binaries will
8117 autodetect the card and use it if present.
8118 [Ben Laurie and Compaq Inc.]
8120 *) Work around for Netscape hang bug. This sends certificate request
8121 and server done in one record. Since this is perfectly legal in the
8122 SSL/TLS protocol it isn't a "bug" option and is on by default. See
8123 the bugs/SSLv3 entry for more info.
8126 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
8129 *) Add -rand argument to smime and pkcs12 applications and read/write
8133 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
8136 *) Add command line password options to the remaining applications.
8139 *) Bug fix for BN_div_recp() for numerators with an even number of
8143 *) More tests in bntest.c, and changed test_bn output.
8146 *) ./config recognizes MacOS X now.
8149 *) Bug fix for BN_div() when the first words of num and divsor are
8150 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
8153 *) Add support for various broken PKCS#8 formats, and command line
8154 options to produce them.
8157 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
8158 get temporary BIGNUMs from a BN_CTX.
8161 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
8165 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
8166 include a #define from the old name to the new. The original intent
8167 was that statically linked binaries could for example just call
8168 SSLeay_add_all_ciphers() to just add ciphers to the table and not
8169 link with digests. This never worked becayse SSLeay_add_all_digests()
8170 and SSLeay_add_all_ciphers() were in the same source file so calling
8171 one would link with the other. They are now in separate source files.
8174 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
8177 *) Use a less unusual form of the Miller-Rabin primality test (it used
8178 a binary algorithm for exponentiation integrated into the Miller-Rabin
8179 loop, our standard modexp algorithms are faster).
8182 *) Support for the EBCDIC character set completed.
8183 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
8185 *) Source code cleanups: use const where appropriate, eliminate casts,
8186 use void * instead of char * in lhash.
8189 *) Bugfix: ssl3_send_server_key_exchange was not restartable
8190 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
8191 this the server could overwrite ephemeral keys that the client
8195 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
8196 using 50 iterations of the Rabin-Miller test.
8198 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
8199 iterations of the Rabin-Miller test as required by the appendix
8200 to FIPS PUB 186[-1]) instead of DSA_is_prime.
8201 As BN_is_prime_fasttest includes trial division, DSA parameter
8202 generation becomes much faster.
8204 This implies a change for the callback functions in DSA_is_prime
8205 and DSA_generate_parameters: The callback function is called once
8206 for each positive witness in the Rabin-Miller test, not just
8207 occasionally in the inner loop; and the parameters to the
8208 callback function now provide an iteration count for the outer
8209 loop rather than for the current invocation of the inner loop.
8210 DSA_generate_parameters additionally can call the callback
8211 function with an 'iteration count' of -1, meaning that a
8212 candidate has passed the trial division test (when q is generated
8213 from an application-provided seed, trial division is skipped).
8216 *) New function BN_is_prime_fasttest that optionally does trial
8217 division before starting the Rabin-Miller test and has
8218 an additional BN_CTX * argument (whereas BN_is_prime always
8219 has to allocate at least one BN_CTX).
8220 'callback(1, -1, cb_arg)' is called when a number has passed the
8221 trial division stage.
8224 *) Fix for bug in CRL encoding. The validity dates weren't being handled
8228 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
8231 *) New function BN_pseudo_rand().
8234 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
8235 bignum version of BN_from_montgomery() with the working code from
8236 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
8240 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
8241 made it impossible to use the same SSL_SESSION data structure in
8242 SSL2 clients in multiple threads.
8245 *) The return value of RAND_load_file() no longer counts bytes obtained
8246 by stat(). RAND_load_file(..., -1) is new and uses the complete file
8247 to seed the PRNG (previously an explicit byte count was required).
8248 [Ulf Möller, Bodo Möller]
8250 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
8251 used (char *) instead of (void *) and had casts all over the place.
8254 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
8257 *) Retain source code compatibility for BN_prime_checks macro:
8258 BN_is_prime(..., BN_prime_checks, ...) now uses
8259 BN_prime_checks_for_size to determine the appropriate number of
8260 Rabin-Miller iterations.
8263 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
8264 DH_CHECK_P_NOT_SAFE_PRIME.
8265 (Check if this is true? OpenPGP calls them "strong".)
8268 *) Merge the functionality of "dh" and "gendh" programs into a new program
8269 "dhparam". The old programs are retained for now but will handle DH keys
8270 (instead of parameters) in future.
8273 *) Make the ciphers, s_server and s_client programs check the return values
8274 when a new cipher list is set.
8277 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
8278 ciphers. Before when the 56bit ciphers were enabled the sorting was
8281 The syntax for the cipher sorting has been extended to support sorting by
8282 cipher-strength (using the strength_bits hard coded in the tables).
8283 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
8285 Fix a bug in the cipher-command parser: when supplying a cipher command
8286 string with an "undefined" symbol (neither command nor alphanumeric
8287 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
8288 an error is flagged.
8290 Due to the strength-sorting extension, the code of the
8291 ssl_create_cipher_list() function was completely rearranged. I hope that
8292 the readability was also increased :-)
8293 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8295 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
8296 for the first serial number and places 2 in the serial number file. This
8297 avoids problems when the root CA is created with serial number zero and
8298 the first user certificate has the same issuer name and serial number
8302 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
8303 the new code. Add documentation for this stuff.
8306 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
8307 X509_*() to X509at_*() on the grounds that they don't handle X509
8308 structures and behave in an analagous way to the X509v3 functions:
8309 they shouldn't be called directly but wrapper functions should be used
8312 So we also now have some wrapper functions that call the X509at functions
8313 when passed certificate requests. (TO DO: similar things can be done with
8314 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
8315 things. Some of these need some d2i or i2d and print functionality
8316 because they handle more complex structures.)
8319 *) Add missing #ifndefs that caused missing symbols when building libssl
8320 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
8321 NO_RSA in ssl/s2*.c.
8322 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
8324 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
8325 has a return value which indicates the quality of the random data
8326 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
8327 error queue. New function RAND_pseudo_bytes() generates output that is
8328 guaranteed to be unique but not unpredictable. RAND_add is like
8329 RAND_seed, but takes an extra argument for an entropy estimate
8330 (RAND_seed always assumes full entropy).
8333 *) Do more iterations of Rabin-Miller probable prime test (specifically,
8334 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
8335 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
8336 in crypto/bn/bn_prime.c for the complete table). This guarantees a
8337 false-positive rate of at most 2^-80 for random input.
8340 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
8343 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
8344 in the 0.9.5 release), this returns the chain
8345 from an X509_CTX structure with a dup of the stack and all
8346 the X509 reference counts upped: so the stack will exist
8347 after X509_CTX_cleanup() has been called. Modify pkcs12.c
8350 Also make SSL_SESSION_print() print out the verify return
8354 *) Add manpage for the pkcs12 command. Also change the default
8355 behaviour so MAC iteration counts are used unless the new
8356 -nomaciter option is used. This improves file security and
8357 only older versions of MSIE (4.0 for example) need it.
8360 *) Honor the no-xxx Configure options when creating .DEF files.
8363 *) Add PKCS#10 attributes to field table: challengePassword,
8364 unstructuredName and unstructuredAddress. These are taken from
8365 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
8366 international characters are used.
8368 More changes to X509_ATTRIBUTE code: allow the setting of types
8369 based on strings. Remove the 'loc' parameter when adding
8370 attributes because these will be a SET OF encoding which is sorted
8374 *) Initial changes to the 'req' utility to allow request generation
8375 automation. This will allow an application to just generate a template
8376 file containing all the field values and have req construct the
8379 Initial support for X509_ATTRIBUTE handling. Stacks of these are
8380 used all over the place including certificate requests and PKCS#7
8381 structures. They are currently handled manually where necessary with
8382 some primitive wrappers for PKCS#7. The new functions behave in a
8383 manner analogous to the X509 extension functions: they allow
8384 attributes to be looked up by NID and added.
8386 Later something similar to the X509V3 code would be desirable to
8387 automatically handle the encoding, decoding and printing of the
8388 more complex types. The string types like challengePassword can
8389 be handled by the string table functions.
8391 Also modified the multi byte string table handling. Now there is
8392 a 'global mask' which masks out certain types. The table itself
8393 can use the flag STABLE_NO_MASK to ignore the mask setting: this
8394 is useful when for example there is only one permissible type
8395 (as in countryName) and using the mask might result in no valid
8399 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
8400 SSL_get_peer_finished to allow applications to obtain the latest
8401 Finished messages sent to the peer or expected from the peer,
8402 respectively. (SSL_get_peer_finished is usually the Finished message
8403 actually received from the peer, otherwise the protocol will be aborted.)
8405 As the Finished message are message digests of the complete handshake
8406 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
8407 be used for external authentication procedures when the authentication
8408 provided by SSL/TLS is not desired or is not enough.
8411 *) Enhanced support for Alpha Linux is added. Now ./config checks if
8412 the host supports BWX extension and if Compaq C is present on the
8413 $PATH. Just exploiting of the BWX extension results in 20-30%
8414 performance kick for some algorithms, e.g. DES and RC4 to mention
8415 a couple. Compaq C in turn generates ~20% faster code for MD5 and
8419 *) Add support for MS "fast SGC". This is arguably a violation of the
8420 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
8421 weak crypto and after checking the certificate is SGC a second one
8422 with strong crypto. MS SGC stops the first handshake after receiving
8423 the server certificate message and sends a second client hello. Since
8424 a server will typically do all the time consuming operations before
8425 expecting any further messages from the client (server key exchange
8426 is the most expensive) there is little difference between the two.
8428 To get OpenSSL to support MS SGC we have to permit a second client
8429 hello message after we have sent server done. In addition we have to
8430 reset the MAC if we do get this second client hello.
8433 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
8434 if a DER encoded private key is RSA or DSA traditional format. Changed
8435 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
8436 format DER encoded private key. Newer code should use PKCS#8 format which
8437 has the key type encoded in the ASN1 structure. Added DER private key
8438 support to pkcs8 application.
8441 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8442 ciphersuites has been selected (as required by the SSL 3/TLS 1
8443 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8444 is set, we interpret this as a request to violate the specification
8445 (the worst that can happen is a handshake failure, and 'correct'
8446 behaviour would result in a handshake failure anyway).
8449 *) In SSL_CTX_add_session, take into account that there might be multiple
8450 SSL_SESSION structures with the same session ID (e.g. when two threads
8451 concurrently obtain them from an external cache).
8452 The internal cache can handle only one SSL_SESSION with a given ID,
8453 so if there's a conflict, we now throw out the old one to achieve
8457 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8458 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8459 some routines that use cipher OIDs: some ciphers do not have OIDs
8460 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8464 *) Simplify the trust setting structure and code. Now we just have
8465 two sequences of OIDs for trusted and rejected settings. These will
8466 typically have values the same as the extended key usage extension
8467 and any application specific purposes.
8469 The trust checking code now has a default behaviour: it will just
8470 check for an object with the same NID as the passed id. Functions can
8471 be provided to override either the default behaviour or the behaviour
8472 for a given id. SSL client, server and email already have functions
8473 in place for compatibility: they check the NID and also return "trusted"
8474 if the certificate is self signed.
8477 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8478 traditional format into an EVP_PKEY structure.
8481 *) Add a password callback function PEM_cb() which either prompts for
8482 a password if usr_data is NULL or otherwise assumes it is a null
8483 terminated password. Allow passwords to be passed on command line
8484 environment or config files in a few more utilities.
8487 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8488 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8489 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8490 Update documentation.
8493 *) Support for ASN1 "NULL" type. This could be handled before by using
8494 ASN1_TYPE but there wasn't any function that would try to read a NULL
8495 and produce an error if it couldn't. For compatibility we also have
8496 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8497 don't allocate anything because they don't need to.
8500 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8502 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8504 *) Rebuild of the memory allocation routines used by OpenSSL code and
8505 possibly others as well. The purpose is to make an interface that
8506 provide hooks so anyone can build a separate set of allocation and
8507 deallocation routines to be used by OpenSSL, for example memory
8508 pool implementations, or something else, which was previously hard
8509 since Malloc(), Realloc() and Free() were defined as macros having
8510 the values malloc, realloc and free, respectively (except for Win32
8511 compilations). The same is provided for memory debugging code.
8512 OpenSSL already comes with functionality to find memory leaks, but
8513 this gives people a chance to debug other memory problems.
8515 With these changes, a new set of functions and macros have appeared:
8517 CRYPTO_set_mem_debug_functions() [F]
8518 CRYPTO_get_mem_debug_functions() [F]
8519 CRYPTO_dbg_set_options() [F]
8520 CRYPTO_dbg_get_options() [F]
8521 CRYPTO_malloc_debug_init() [M]
8523 The memory debug functions are NULL by default, unless the library
8524 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8525 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8526 gives the standard debugging functions that come with OpenSSL) or
8527 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8528 provided by the library user) must be used. When the standard
8529 debugging functions are used, CRYPTO_dbg_set_options can be used to
8530 request additional information:
8531 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8532 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8534 Also, things like CRYPTO_set_mem_functions will always give the
8535 expected result (the new set of functions is used for allocation
8536 and deallocation) at all times, regardless of platform and compiler
8539 To finish it up, some functions that were never use in any other
8540 way than through macros have a new API and new semantic:
8543 CRYPTO_dbg_realloc()
8546 All macros of value have retained their old syntax.
8547 [Richard Levitte and Bodo Moeller]
8549 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8550 ordering of SMIMECapabilities wasn't in "strength order" and there
8551 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8555 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8556 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8557 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
8559 *) Merge in my S/MIME library for OpenSSL. This provides a simple
8560 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
8561 functionality to handle multipart/signed properly) and a utility
8562 called 'smime' to call all this stuff. This is based on code I
8563 originally wrote for Celo who have kindly allowed it to be
8564 included in OpenSSL.
8567 *) Add variants des_set_key_checked and des_set_key_unchecked of
8568 des_set_key (aka des_key_sched). Global variable des_check_key
8569 decides which of these is called by des_set_key; this way
8570 des_check_key behaves as it always did, but applications and
8571 the library itself, which was buggy for des_check_key == 1,
8572 have a cleaner way to pick the version they need.
8575 *) New function PKCS12_newpass() which changes the password of a
8579 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
8580 dynamic mix. In both cases the ids can be used as an index into the
8581 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
8582 functions so they accept a list of the field values and the
8583 application doesn't need to directly manipulate the X509_TRUST
8587 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
8591 *) Modify the way the V3 extension code looks up extensions. This now
8592 works in a similar way to the object code: we have some "standard"
8593 extensions in a static table which is searched with OBJ_bsearch()
8594 and the application can add dynamic ones if needed. The file
8595 crypto/x509v3/ext_dat.h now has the info: this file needs to be
8596 updated whenever a new extension is added to the core code and kept
8597 in ext_nid order. There is a simple program 'tabtest.c' which checks
8598 this. New extensions are not added too often so this file can readily
8599 be maintained manually.
8601 There are two big advantages in doing things this way. The extensions
8602 can be looked up immediately and no longer need to be "added" using
8603 X509V3_add_standard_extensions(): this function now does nothing.
8604 [Side note: I get *lots* of email saying the extension code doesn't
8605 work because people forget to call this function]
8606 Also no dynamic allocation is done unless new extensions are added:
8607 so if we don't add custom extensions there is no need to call
8608 X509V3_EXT_cleanup().
8611 *) Modify enc utility's salting as follows: make salting the default. Add a
8612 magic header, so unsalted files fail gracefully instead of just decrypting
8613 to garbage. This is because not salting is a big security hole, so people
8614 should be discouraged from doing it.
8617 *) Fixes and enhancements to the 'x509' utility. It allowed a message
8618 digest to be passed on the command line but it only used this
8619 parameter when signing a certificate. Modified so all relevant
8620 operations are affected by the digest parameter including the
8621 -fingerprint and -x509toreq options. Also -x509toreq choked if a
8622 DSA key was used because it didn't fix the digest.
8625 *) Initial certificate chain verify code. Currently tests the untrusted
8626 certificates for consistency with the verify purpose (which is set
8627 when the X509_STORE_CTX structure is set up) and checks the pathlength.
8629 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
8630 this is because it will reject chains with invalid extensions whereas
8631 every previous version of OpenSSL and SSLeay made no checks at all.
8633 Trust code: checks the root CA for the relevant trust settings. Trust
8634 settings have an initial value consistent with the verify purpose: e.g.
8635 if the verify purpose is for SSL client use it expects the CA to be
8636 trusted for SSL client use. However the default value can be changed to
8637 permit custom trust settings: one example of this would be to only trust
8638 certificates from a specific "secure" set of CAs.
8640 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
8641 which should be used for version portability: especially since the
8642 verify structure is likely to change more often now.
8644 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
8645 to set them. If not set then assume SSL clients will verify SSL servers
8648 Two new options to the verify program: -untrusted allows a set of
8649 untrusted certificates to be passed in and -purpose which sets the
8650 intended purpose of the certificate. If a purpose is set then the
8651 new chain verify code is used to check extension consistency.
8654 *) Support for the authority information access extension.
8657 *) Modify RSA and DSA PEM read routines to transparently handle
8658 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
8659 public keys in a format compatible with certificate
8660 SubjectPublicKeyInfo structures. Unfortunately there were already
8661 functions called *_PublicKey_* which used various odd formats so
8662 these are retained for compatibility: however the DSA variants were
8663 never in a public release so they have been deleted. Changed dsa/rsa
8664 utilities to handle the new format: note no releases ever handled public
8665 keys so we should be OK.
8667 The primary motivation for this change is to avoid the same fiasco
8668 that dogs private keys: there are several incompatible private key
8669 formats some of which are standard and some OpenSSL specific and
8670 require various evil hacks to allow partial transparent handling and
8671 even then it doesn't work with DER formats. Given the option anything
8672 other than PKCS#8 should be dumped: but the other formats have to
8673 stay in the name of compatibility.
8675 With public keys and the benefit of hindsight one standard format
8676 is used which works with EVP_PKEY, RSA or DSA structures: though
8677 it clearly returns an error if you try to read the wrong kind of key.
8679 Added a -pubkey option to the 'x509' utility to output the public key.
8680 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
8681 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
8682 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
8683 that do the same as the EVP_PKEY_assign_*() except they up the
8684 reference count of the added key (they don't "swallow" the
8688 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
8689 CRLs would fail if the file contained no certificates or no CRLs:
8690 added a new function to read in both types and return the number
8691 read: this means that if none are read it will be an error. The
8692 DER versions of the certificate and CRL reader would always fail
8693 because it isn't possible to mix certificates and CRLs in DER format
8694 without choking one or the other routine. Changed this to just read
8695 a certificate: this is the best we can do. Also modified the code
8696 in apps/verify.c to take notice of return codes: it was previously
8697 attempting to read in certificates from NULL pointers and ignoring
8698 any errors: this is one reason why the cert and CRL reader seemed
8699 to work. It doesn't check return codes from the default certificate
8700 routines: these may well fail if the certificates aren't installed.
8703 *) Code to support otherName option in GeneralName.
8706 *) First update to verify code. Change the verify utility
8707 so it warns if it is passed a self signed certificate:
8708 for consistency with the normal behaviour. X509_verify
8709 has been modified to it will now verify a self signed
8710 certificate if *exactly* the same certificate appears
8711 in the store: it was previously impossible to trust a
8712 single self signed certificate. This means that:
8713 openssl verify ss.pem
8714 now gives a warning about a self signed certificate but
8715 openssl verify -CAfile ss.pem ss.pem
8719 *) For servers, store verify_result in SSL_SESSION data structure
8720 (and add it to external session representation).
8721 This is needed when client certificate verifications fails,
8722 but an application-provided verification callback (set by
8723 SSL_CTX_set_cert_verify_callback) allows accepting the session
8724 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
8725 but returns 1): When the session is reused, we have to set
8726 ssl->verify_result to the appropriate error code to avoid
8728 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
8730 *) Fix a bug in the new PKCS#7 code: it didn't consider the
8731 case in PKCS7_dataInit() where the signed PKCS7 structure
8732 didn't contain any existing data because it was being created.
8733 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
8735 *) Add a salt to the key derivation routines in enc.c. This
8736 forms the first 8 bytes of the encrypted file. Also add a
8737 -S option to allow a salt to be input on the command line.
8740 *) New function X509_cmp(). Oddly enough there wasn't a function
8741 to compare two certificates. We do this by working out the SHA1
8742 hash and comparing that. X509_cmp() will be needed by the trust
8746 *) SSL_get1_session() is like SSL_get_session(), but increments
8747 the reference count in the SSL_SESSION returned.
8748 [Geoff Thorpe <geoff@eu.c2.net>]
8750 *) Fix for 'req': it was adding a null to request attributes.
8751 Also change the X509_LOOKUP and X509_INFO code to handle
8752 certificate auxiliary information.
8755 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8759 *) Add the possibility to add extra information to the memory leak
8760 detecting output, to form tracebacks, showing from where each
8761 allocation was originated: CRYPTO_push_info("constant string") adds
8762 the string plus current file name and line number to a per-thread
8763 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8764 is like calling CYRPTO_pop_info() until the stack is empty.
8765 Also updated memory leak detection code to be multi-thread-safe.
8768 *) Add options -text and -noout to pkcs7 utility and delete the
8769 encryption options which never did anything. Update docs.
8772 *) Add options to some of the utilities to allow the pass phrase
8773 to be included on either the command line (not recommended on
8774 OSes like Unix) or read from the environment. Update the
8775 manpages and fix a few bugs.
8778 *) Add a few manpages for some of the openssl commands.
8781 *) Fix the -revoke option in ca. It was freeing up memory twice,
8782 leaking and not finding already revoked certificates.
8785 *) Extensive changes to support certificate auxiliary information.
8786 This involves the use of X509_CERT_AUX structure and X509_AUX
8787 functions. An X509_AUX function such as PEM_read_X509_AUX()
8788 can still read in a certificate file in the usual way but it
8789 will also read in any additional "auxiliary information". By
8790 doing things this way a fair degree of compatibility can be
8791 retained: existing certificates can have this information added
8792 using the new 'x509' options.
8794 Current auxiliary information includes an "alias" and some trust
8795 settings. The trust settings will ultimately be used in enhanced
8796 certificate chain verification routines: currently a certificate
8797 can only be trusted if it is self signed and then it is trusted
8801 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8802 The problem was that one of the replacement routines had not been working
8803 since SSLeay releases. For now the offending routine has been replaced
8804 with non-optimised assembler. Even so, this now gives around 95%
8805 performance improvement for 1024 bit RSA signs.
8808 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8809 handling. Most clients have the effective key size in bits equal to
8810 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8811 A few however don't do this and instead use the size of the decrypted key
8812 to determine the RC2 key length and the AlgorithmIdentifier to determine
8813 the effective key length. In this case the effective key length can still
8814 be 40 bits but the key length can be 168 bits for example. This is fixed
8815 by manually forcing an RC2 key into the EVP_PKEY structure because the
8816 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8817 the key length and effective key length are equal.
8820 *) Add a bunch of functions that should simplify the creation of
8821 X509_NAME structures. Now you should be able to do:
8822 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8823 and have it automatically work out the correct field type and fill in
8824 the structures. The more adventurous can try:
8825 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8826 and it will (hopefully) work out the correct multibyte encoding.
8829 *) Change the 'req' utility to use the new field handling and multibyte
8830 copy routines. Before the DN field creation was handled in an ad hoc
8831 way in req, ca, and x509 which was rather broken and didn't support
8832 BMPStrings or UTF8Strings. Since some software doesn't implement
8833 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8834 using the dirstring_type option. See the new comment in the default
8835 openssl.cnf for more info.
8838 *) Make crypto/rand/md_rand.c more robust:
8839 - Assure unique random numbers after fork().
8840 - Make sure that concurrent threads access the global counter and
8841 md serializably so that we never lose entropy in them
8842 or use exactly the same state in multiple threads.
8843 Access to the large state is not always serializable because
8844 the additional locking could be a performance killer, and
8845 md should be large enough anyway.
8848 *) New file apps/app_rand.c with commonly needed functionality
8849 for handling the random seed file.
8851 Use the random seed file in some applications that previously did not:
8853 dsaparam -genkey (which also ignored its '-rand' option),
8856 x509 (when signing).
8857 Except on systems with /dev/urandom, it is crucial to have a random
8858 seed file at least for key creation, DSA signing, and for DH exchanges;
8859 for RSA signatures we could do without one.
8861 gendh and gendsa (unlike genrsa) used to read only the first byte
8862 of each file listed in the '-rand' option. The function as previously
8863 found in genrsa is now in app_rand.c and is used by all programs
8864 that support '-rand'.
8867 *) In RAND_write_file, use mode 0600 for creating files;
8868 don't just chmod when it may be too late.
8871 *) Report an error from X509_STORE_load_locations
8872 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8875 *) New function ASN1_mbstring_copy() this copies a string in either
8876 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8877 into an ASN1_STRING type. A mask of permissible types is passed
8878 and it chooses the "minimal" type to use or an error if not type
8882 *) Add function equivalents to the various macros in asn1.h. The old
8883 macros are retained with an M_ prefix. Code inside the library can
8884 use the M_ macros. External code (including the openssl utility)
8885 should *NOT* in order to be "shared library friendly".
8888 *) Add various functions that can check a certificate's extensions
8889 to see if it usable for various purposes such as SSL client,
8890 server or S/MIME and CAs of these types. This is currently
8891 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8892 verification. Also added a -purpose flag to x509 utility to
8893 print out all the purposes.
8896 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8900 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8901 for, obtain and decode and extension and obtain its critical flag.
8902 This allows all the necessary extension code to be handled in a
8903 single function call.
8906 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8907 platforms. See crypto/rc4/rc4_enc.c for further details.
8910 *) New -noout option to asn1parse. This causes no output to be produced
8911 its main use is when combined with -strparse and -out to extract data
8912 from a file (which may not be in ASN.1 format).
8915 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8916 when producing the local key id.
8917 [Richard Levitte <levitte@stacken.kth.se>]
8919 *) New option -dhparam in s_server. This allows a DH parameter file to be
8920 stated explicitly. If it is not stated then it tries the first server
8921 certificate file. The previous behaviour hard coded the filename
8925 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8926 a public key to be input or output. For example:
8927 openssl rsa -in key.pem -pubout -out pubkey.pem
8928 Also added necessary DSA public key functions to handle this.
8931 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8932 in the message. This was handled by allowing
8933 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8934 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8936 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8937 to the end of the strings whereas this didn't. This would cause problems
8938 if strings read with d2i_ASN1_bytes() were later modified.
8939 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8941 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8942 data and it contains EOF it will end up returning an error. This is
8943 caused by input 46 bytes long. The cause is due to the way base64
8944 BIOs find the start of base64 encoded data. They do this by trying a
8945 trial decode on each line until they find one that works. When they
8946 do a flag is set and it starts again knowing it can pass all the
8947 data directly through the decoder. Unfortunately it doesn't reset
8948 the context it uses. This means that if EOF is reached an attempt
8949 is made to pass two EOFs through the context and this causes the
8950 resulting error. This can also cause other problems as well. As is
8951 usual with these problems it takes *ages* to find and the fix is
8952 trivial: move one line.
8953 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8955 *) Ugly workaround to get s_client and s_server working under Windows. The
8956 old code wouldn't work because it needed to select() on sockets and the
8957 tty (for keypresses and to see if data could be written). Win32 only
8958 supports select() on sockets so we select() with a 1s timeout on the
8959 sockets and then see if any characters are waiting to be read, if none
8960 are present then we retry, we also assume we can always write data to
8961 the tty. This isn't nice because the code then blocks until we've
8962 received a complete line of data and it is effectively polling the
8963 keyboard at 1s intervals: however it's quite a bit better than not
8964 working at all :-) A dedicated Windows application might handle this
8965 with an event loop for example.
8968 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8969 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8970 will be called when RSA_sign() and RSA_verify() are used. This is useful
8971 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8972 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8973 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8974 This necessitated the support of an extra signature type NID_md5_sha1
8975 for SSL signatures and modifications to the SSL library to use it instead
8976 of calling RSA_public_decrypt() and RSA_private_encrypt().
8979 *) Add new -verify -CAfile and -CApath options to the crl program, these
8980 will lookup a CRL issuers certificate and verify the signature in a
8981 similar way to the verify program. Tidy up the crl program so it
8982 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8983 less strict. It will now permit CRL extensions even if it is not
8984 a V2 CRL: this will allow it to tolerate some broken CRLs.
8987 *) Initialize all non-automatic variables each time one of the openssl
8988 sub-programs is started (this is necessary as they may be started
8989 multiple times from the "OpenSSL>" prompt).
8990 [Lennart Bang, Bodo Moeller]
8992 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8993 removing all other RSA functionality (this is what NO_RSA does). This
8994 is so (for example) those in the US can disable those operations covered
8995 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8999 *) Non-copying interface to BIO pairs.
9000 (still largely untested)
9003 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9004 ASCII string. This was handled independently in various places before.
9007 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9008 UTF8 strings a character at a time.
9011 *) Use client_version from client hello to select the protocol
9012 (s23_srvr.c) and for RSA client key exchange verification
9013 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9016 *) Add various utility functions to handle SPKACs, these were previously
9017 handled by poking round in the structure internals. Added new function
9018 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9019 print, verify and generate SPKACs. Based on an original idea from
9020 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9023 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9026 *) Allow the config file extension section to be overwritten on the
9027 command line. Based on an original idea from Massimiliano Pala
9028 <madwolf@comune.modena.it>. The new option is called -extensions
9029 and can be applied to ca, req and x509. Also -reqexts to override
9030 the request extensions in req and -crlexts to override the crl extensions
9034 *) Add new feature to the SPKAC handling in ca. Now you can include
9035 the same field multiple times by preceding it by "XXXX." for example:
9038 this is the same syntax as used in the req config file.
9041 *) Allow certificate extensions to be added to certificate requests. These
9042 are specified in a 'req_extensions' option of the req section of the
9043 config file. They can be printed out with the -text option to req but
9044 are otherwise ignored at present.
9047 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9048 data read consists of only the final block it would not decrypted because
9049 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9050 A misplaced 'break' also meant the decrypted final block might not be
9051 copied until the next read.
9054 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9055 a few extra parameters to the DH structure: these will be useful if
9056 for example we want the value of 'q' or implement X9.42 DH.
9059 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9060 provides hooks that allow the default DSA functions or functions on a
9061 "per key" basis to be replaced. This allows hardware acceleration and
9062 hardware key storage to be handled without major modification to the
9063 library. Also added low level modexp hooks and CRYPTO_EX structure and
9064 associated functions.
9067 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9068 as "read only": it can't be written to and the buffer it points to will
9069 not be freed. Reading from a read only BIO is much more efficient than
9070 a normal memory BIO. This was added because there are several times when
9071 an area of memory needs to be read from a BIO. The previous method was
9072 to create a memory BIO and write the data to it, this results in two
9073 copies of the data and an O(n^2) reading algorithm. There is a new
9074 function BIO_new_mem_buf() which creates a read only memory BIO from
9075 an area of memory. Also modified the PKCS#7 routines to use read only
9079 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9080 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9081 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9082 but a retry condition occured while trying to read the rest.
9085 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9086 NID_pkcs7_encrypted by default: this was wrong since this should almost
9087 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9088 the encrypted data type: this is a more sensible place to put it and it
9089 allows the PKCS#12 code to be tidied up that duplicated this
9093 *) Changed obj_dat.pl script so it takes its input and output files on
9094 the command line. This should avoid shell escape redirection problems
9098 *) Initial support for certificate extension requests, these are included
9099 in things like Xenroll certificate requests. Included functions to allow
9100 extensions to be obtained and added.
9103 *) -crlf option to s_client and s_server for sending newlines as
9104 CRLF (as required by many protocols).
9107 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9109 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9110 [Ralf S. Engelschall]
9112 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9113 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9115 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
9119 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
9120 DH parameters/keys (q is lost during that conversion, but the resulting
9121 DH parameters contain its length).
9123 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
9124 much faster than DH_generate_parameters (which creates parameters
9125 where p = 2*q + 1), and also the smaller q makes DH computations
9126 much more efficient (160-bit exponentiation instead of 1024-bit
9127 exponentiation); so this provides a convenient way to support DHE
9128 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
9129 utter importance to use
9130 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9132 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9133 when such DH parameters are used, because otherwise small subgroup
9134 attacks may become possible!
9137 *) Avoid memory leak in i2d_DHparams.
9140 *) Allow the -k option to be used more than once in the enc program:
9141 this allows the same encrypted message to be read by multiple recipients.
9144 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
9145 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
9146 it will always use the numerical form of the OID, even if it has a short
9150 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
9151 method only got called if p,q,dmp1,dmq1,iqmp components were present,
9152 otherwise bn_mod_exp was called. In the case of hardware keys for example
9153 no private key components need be present and it might store extra data
9154 in the RSA structure, which cannot be accessed from bn_mod_exp.
9155 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
9156 private key operations.
9159 *) Added support for SPARC Linux.
9162 *) pem_password_cb function type incompatibly changed from
9163 typedef int pem_password_cb(char *buf, int size, int rwflag);
9165 ....(char *buf, int size, int rwflag, void *userdata);
9166 so that applications can pass data to their callbacks:
9167 The PEM[_ASN1]_{read,write}... functions and macros now take an
9168 additional void * argument, which is just handed through whenever
9169 the password callback is called.
9170 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
9172 New function SSL_CTX_set_default_passwd_cb_userdata.
9174 Compatibility note: As many C implementations push function arguments
9175 onto the stack in reverse order, the new library version is likely to
9176 interoperate with programs that have been compiled with the old
9177 pem_password_cb definition (PEM_whatever takes some data that
9178 happens to be on the stack as its last argument, and the callback
9179 just ignores this garbage); but there is no guarantee whatsoever that
9182 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
9183 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
9184 problems not only on Windows, but also on some Unix platforms.
9185 To avoid problematic command lines, these definitions are now in an
9186 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
9187 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
9190 *) MIPS III/IV assembler module is reimplemented.
9193 *) More DES library cleanups: remove references to srand/rand and
9194 delete an unused file.
9197 *) Add support for the the free Netwide assembler (NASM) under Win32,
9198 since not many people have MASM (ml) and it can be hard to obtain.
9199 This is currently experimental but it seems to work OK and pass all
9200 the tests. Check out INSTALL.W32 for info.
9203 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
9204 without temporary keys kept an extra copy of the server key,
9205 and connections with temporary keys did not free everything in case
9209 *) New function RSA_check_key and new openssl rsa option -check
9210 for verifying the consistency of RSA keys.
9211 [Ulf Moeller, Bodo Moeller]
9213 *) Various changes to make Win32 compile work:
9214 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
9215 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
9216 comparison" warnings.
9217 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
9220 *) Add a debugging option to PKCS#5 v2 key generation function: when
9221 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
9222 derived keys are printed to stderr.
9225 *) Copy the flags in ASN1_STRING_dup().
9226 [Roman E. Pavlov <pre@mo.msk.ru>]
9228 *) The x509 application mishandled signing requests containing DSA
9229 keys when the signing key was also DSA and the parameters didn't match.
9231 It was supposed to omit the parameters when they matched the signing key:
9232 the verifying software was then supposed to automatically use the CA's
9233 parameters if they were absent from the end user certificate.
9235 Omitting parameters is no longer recommended. The test was also
9236 the wrong way round! This was probably due to unusual behaviour in
9237 EVP_cmp_parameters() which returns 1 if the parameters match.
9238 This meant that parameters were omitted when they *didn't* match and
9239 the certificate was useless. Certificates signed with 'ca' didn't have
9241 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
9243 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
9244 The interface is as follows:
9245 Applications can use
9246 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
9247 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
9248 "off" is now the default.
9249 The library internally uses
9250 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
9251 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
9252 to disable memory-checking temporarily.
9254 Some inconsistent states that previously were possible (and were
9255 even the default) are now avoided.
9257 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
9258 with each memory chunk allocated; this is occasionally more helpful
9259 than just having a counter.
9261 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
9263 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
9267 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
9268 which largely parallels "options", but is for changing API behaviour,
9269 whereas "options" are about protocol behaviour.
9270 Initial "mode" flags are:
9272 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
9273 a single record has been written.
9274 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
9275 retries use the same buffer location.
9276 (But all of the contents must be
9280 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
9283 *) Fix problems with no-hmac etc.
9284 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
9286 *) New functions RSA_get_default_method(), RSA_set_method() and
9287 RSA_get_method(). These allows replacement of RSA_METHODs without having
9288 to mess around with the internals of an RSA structure.
9291 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
9292 Also really enable memory leak checks in openssl.c and in some
9294 [Chad C. Mulligan, Bodo Moeller]
9296 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
9297 up the length of negative integers. This has now been simplified to just
9298 store the length when it is first determined and use it later, rather
9299 than trying to keep track of where data is copied and updating it to
9301 [Steve Henson, reported by Brien Wheeler
9302 <bwheeler@authentica-security.com>]
9304 *) Add a new function PKCS7_signatureVerify. This allows the verification
9305 of a PKCS#7 signature but with the signing certificate passed to the
9306 function itself. This contrasts with PKCS7_dataVerify which assumes the
9307 certificate is present in the PKCS#7 structure. This isn't always the
9308 case: certificates can be omitted from a PKCS#7 structure and be
9309 distributed by "out of band" means (such as a certificate database).
9312 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
9313 function prototypes in pem.h, also change util/mkdef.pl to add the
9314 necessary function names.
9317 *) mk1mf.pl (used by Windows builds) did not properly read the
9318 options set by Configure in the top level Makefile, and Configure
9319 was not even able to write more than one option correctly.
9320 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
9323 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
9324 file to be loaded from a BIO or FILE pointer. The BIO version will
9325 for example allow memory BIOs to contain config info.
9328 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
9329 Whoever hopes to achieve shared-library compatibility across versions
9330 must use this, not the compile-time macro.
9331 (Exercise 0.9.4: Which is the minimum library version required by
9333 Note: All this applies only to multi-threaded programs, others don't
9337 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
9338 through a BIO pair triggered the default case, i.e.
9339 SSLerr(...,SSL_R_UNKNOWN_STATE).
9342 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
9343 can use the SSL library even if none of the specific BIOs is
9347 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
9348 for the encoded length.
9349 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
9351 *) Add initial documentation of the X509V3 functions.
9354 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
9355 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
9356 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
9357 secure PKCS#8 private key format with a high iteration count.
9360 *) Fix determination of Perl interpreter: A perl or perl5
9361 _directory_ in $PATH was also accepted as the interpreter.
9362 [Ralf S. Engelschall]
9364 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
9365 wrong with it but it was very old and did things like calling
9366 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
9370 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
9371 to use the new extension code.
9374 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
9375 with macros. This should make it easier to change their form, add extra
9376 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
9380 *) Add to configuration table a new entry that can specify an alternative
9381 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
9382 according to Mark Crispin <MRC@Panda.COM>.
9386 *) DES CBC did not update the IV. Weird.
9389 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
9390 Changing the behaviour of the former might break existing programs --
9391 where IV updating is needed, des_ncbc_encrypt can be used.
9394 *) When bntest is run from "make test" it drives bc to check its
9395 calculations, as well as internally checking them. If an internal check
9396 fails, it needs to cause bc to give a non-zero result or make test carries
9397 on without noticing the failure. Fixed.
9400 *) DES library cleanups.
9403 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
9404 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
9405 ciphers. NOTE: although the key derivation function has been verified
9406 against some published test vectors it has not been extensively tested
9407 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
9411 *) Instead of "mkdir -p", which is not fully portable, use new
9412 Perl script "util/mkdir-p.pl".
9415 *) Rewrite the way password based encryption (PBE) is handled. It used to
9416 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
9417 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
9418 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
9419 the 'parameter' field of the AlgorithmIdentifier is passed to the
9420 underlying key generation function so it must do its own ASN1 parsing.
9421 This has also changed the EVP_PBE_CipherInit() function which now has a
9422 'parameter' argument instead of literal salt and iteration count values
9423 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
9426 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
9427 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
9428 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
9429 KEY" because this clashed with PKCS#8 unencrypted string. Since this
9430 value was just used as a "magic string" and not used directly its
9431 value doesn't matter.
9434 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9438 *) "linux-sparc64" configuration (ultrapenguin).
9439 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9440 "linux-sparc" configuration.
9441 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9443 *) config now generates no-xxx options for missing ciphers.
9446 *) Support the EBCDIC character set (work in progress).
9447 File ebcdic.c not yet included because it has a different license.
9448 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9450 *) Support BS2000/OSD-POSIX.
9451 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9453 *) Make callbacks for key generation use void * instead of char *.
9456 *) Make S/MIME samples compile (not yet tested).
9459 *) Additional typesafe stacks.
9462 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9466 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9468 *) New configuration variant "sco5-gcc".
9470 *) Updated some demos.
9471 [Sean O Riordain, Wade Scholine]
9473 *) Add missing BIO_free at exit of pkcs12 application.
9476 *) Fix memory leak in conf.c.
9479 *) Updates for Win32 to assembler version of MD5.
9482 *) Set #! path to perl in apps/der_chop to where we found it
9483 instead of using a fixed path.
9486 *) SHA library changes for irix64-mips4-cc.
9489 *) Improvements for VMS support.
9493 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9495 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9496 This also avoids the problems with SC4.2 and unpatched SC5.
9497 [Andy Polyakov <appro@fy.chalmers.se>]
9499 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9500 These are required because of the typesafe stack would otherwise break
9501 existing code. If old code used a structure member which used to be STACK
9502 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9503 sk_num or sk_value it would produce an error because the num, data members
9504 are not present in STACK_OF. Now it just produces a warning. sk_set
9505 replaces the old method of assigning a value to sk_value
9506 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9507 that does this will no longer work (and should use sk_set instead) but
9508 this could be regarded as a "questionable" behaviour anyway.
9511 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9512 correctly handle encrypted S/MIME data.
9515 *) Change type of various DES function arguments from des_cblock
9516 (which means, in function argument declarations, pointer to char)
9517 to des_cblock * (meaning pointer to array with 8 char elements),
9518 which allows the compiler to do more typechecking; it was like
9519 that back in SSLeay, but with lots of ugly casts.
9521 Introduce new type const_des_cblock.
9524 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9525 problems: find RecipientInfo structure that matches recipient certificate
9526 and initialise the ASN1 structures properly based on passed cipher.
9529 *) Belatedly make the BN tests actually check the results.
9532 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9533 to and from BNs: it was completely broken. New compilation option
9534 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9535 key elements as negative integers.
9538 *) Reorganize and speed up MD5.
9539 [Andy Polyakov <appro@fy.chalmers.se>]
9542 [Richard Levitte <richard@levitte.org>]
9544 *) New option -out to asn1parse to allow the parsed structure to be
9545 output to a file. This is most useful when combined with the -strparse
9546 option to examine the output of things like OCTET STRINGS.
9549 *) Make SSL library a little more fool-proof by not requiring any longer
9550 that SSL_set_{accept,connect}_state be called before
9551 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9552 in many applications because usually everything *appeared* to work as
9553 intended anyway -- now it really works as intended).
9556 *) Move openssl.cnf out of lib/.
9559 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
9560 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
9561 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
9562 [Ralf S. Engelschall]
9564 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
9565 handle PKCS#7 enveloped data properly.
9566 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
9568 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
9569 copying pointers. The cert_st handling is changed by this in
9570 various ways (and thus what used to be known as ctx->default_cert
9571 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
9572 any longer when s->cert does not give us what we need).
9573 ssl_cert_instantiate becomes obsolete by this change.
9574 As soon as we've got the new code right (possibly it already is?),
9575 we have solved a couple of bugs of the earlier code where s->cert
9576 was used as if it could not have been shared with other SSL structures.
9578 Note that using the SSL API in certain dirty ways now will result
9579 in different behaviour than observed with earlier library versions:
9580 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
9581 does not influence s as it used to.
9583 In order to clean up things more thoroughly, inside SSL_SESSION
9584 we don't use CERT any longer, but a new structure SESS_CERT
9585 that holds per-session data (if available); currently, this is
9586 the peer's certificate chain and, for clients, the server's certificate
9587 and temporary key. CERT holds only those values that can have
9588 meaningful defaults in an SSL_CTX.
9591 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
9592 from the internal representation. Various PKCS#7 fixes: remove some
9593 evil casts and set the enc_dig_alg field properly based on the signing
9597 *) Allow PKCS#12 password to be set from the command line or the
9598 environment. Let 'ca' get its config file name from the environment
9599 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
9603 *) Allow certificate policies extension to use an IA5STRING for the
9604 organization field. This is contrary to the PKIX definition but
9605 VeriSign uses it and IE5 only recognises this form. Document 'x509'
9609 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
9610 without disallowing inline assembler and the like for non-pedantic builds.
9613 *) Support Borland C++ builder.
9614 [Janez Jere <jj@void.si>, modified by Ulf Möller]
9619 *) SHA-1 cleanups and performance enhancements.
9620 [Andy Polyakov <appro@fy.chalmers.se>]
9622 *) Sparc v8plus assembler for the bignum library.
9623 [Andy Polyakov <appro@fy.chalmers.se>]
9625 *) Accept any -xxx and +xxx compiler options in Configure.
9628 *) Update HPUX configuration.
9631 *) Add missing sk_<type>_unshift() function to safestack.h
9632 [Ralf S. Engelschall]
9634 *) New function SSL_CTX_use_certificate_chain_file that sets the
9635 "extra_cert"s in addition to the certificate. (This makes sense
9636 only for "PEM" format files, as chains as a whole are not
9640 *) Support verify_depth from the SSL API.
9641 x509_vfy.c had what can be considered an off-by-one-error:
9642 Its depth (which was not part of the external interface)
9643 was actually counting the number of certificates in a chain;
9644 now it really counts the depth.
9647 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
9648 instead of X509err, which often resulted in confusing error
9649 messages since the error codes are not globally unique
9650 (e.g. an alleged error in ssl3_accept when a certificate
9651 didn't match the private key).
9653 *) New function SSL_CTX_set_session_id_context that allows to set a default
9654 value (so that you don't need SSL_set_session_id_context for each
9655 connection using the SSL_CTX).
9658 *) OAEP decoding bug fix.
9661 *) Support INSTALL_PREFIX for package builders, as proposed by
9665 *) New Configure options "threads" and "no-threads". For systems
9666 where the proper compiler options are known (currently Solaris
9667 and Linux), "threads" is the default.
9670 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
9673 *) Install various scripts to $(OPENSSLDIR)/misc, not to
9674 $(INSTALLTOP)/bin -- they shouldn't clutter directories
9675 such as /usr/local/bin.
9678 *) "make linux-shared" to build shared libraries.
9679 [Niels Poppe <niels@netbox.org>]
9681 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
9684 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
9685 extension adding in x509 utility.
9688 *) Remove NOPROTO sections and error code comments.
9691 *) Partial rewrite of the DEF file generator to now parse the ANSI
9695 *) New Configure options --prefix=DIR and --openssldir=DIR.
9698 *) Complete rewrite of the error code script(s). It is all now handled
9699 by one script at the top level which handles error code gathering,
9700 header rewriting and C source file generation. It should be much better
9701 than the old method: it now uses a modified version of Ulf's parser to
9702 read the ANSI prototypes in all header files (thus the old K&R definitions
9703 aren't needed for error creation any more) and do a better job of
9704 translating function codes into names. The old 'ASN1 error code imbedded
9705 in a comment' is no longer necessary and it doesn't use .err files which
9706 have now been deleted. Also the error code call doesn't have to appear all
9707 on one line (which resulted in some large lines...).
9710 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
9713 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
9714 0 (which usually indicates a closed connection), but continue reading.
9717 *) Fix some race conditions.
9720 *) Add support for CRL distribution points extension. Add Certificate
9721 Policies and CRL distribution points documentation.
9724 *) Move the autogenerated header file parts to crypto/opensslconf.h.
9727 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
9728 8 of keying material. Merlin has also confirmed interop with this fix
9729 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
9730 [Merlin Hughes <merlin@baltimore.ie>]
9732 *) Fix lots of warnings.
9733 [Richard Levitte <levitte@stacken.kth.se>]
9735 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
9736 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9737 [Richard Levitte <levitte@stacken.kth.se>]
9739 *) Fix problems with sizeof(long) == 8.
9740 [Andy Polyakov <appro@fy.chalmers.se>]
9742 *) Change functions to ANSI C.
9745 *) Fix typos in error codes.
9746 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9748 *) Remove defunct assembler files from Configure.
9751 *) SPARC v8 assembler BIGNUM implementation.
9752 [Andy Polyakov <appro@fy.chalmers.se>]
9754 *) Support for Certificate Policies extension: both print and set.
9755 Various additions to support the r2i method this uses.
9758 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9759 return a const string when you are expecting an allocated buffer.
9762 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9763 types DirectoryString and DisplayText.
9766 *) Add code to allow r2i extensions to access the configuration database,
9767 add an LHASH database driver and add several ctx helper functions.
9770 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9771 fail when they extended the size of a BIGNUM.
9774 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9775 support typesafe stack.
9778 *) Fix typo in SSL_[gs]et_options().
9779 [Nils Frostberg <nils@medcom.se>]
9781 *) Delete various functions and files that belonged to the (now obsolete)
9782 old X509V3 handling code.
9785 *) New Configure option "rsaref".
9788 *) Don't auto-generate pem.h.
9791 *) Introduce type-safe ASN.1 SETs.
9794 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9795 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9797 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9798 that links with OpenSSL (well at least cause lots of warnings), but fear
9799 not: the conversion is trivial, and it eliminates loads of evil casts. A
9800 few STACKed things have been converted already. Feel free to convert more.
9801 In the fullness of time, I'll do away with the STACK type altogether.
9804 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9805 specified in <certfile> by updating the entry in the index.txt file.
9806 This way one no longer has to edit the index.txt file manually for
9807 revoking a certificate. The -revoke option does the gory details now.
9808 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9810 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9811 `-text' option at all and this way the `-noout -text' combination was
9812 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9813 [Ralf S. Engelschall]
9815 *) Make sure a corresponding plain text error message exists for the
9816 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9817 verify callback function determined that a certificate was revoked.
9818 [Ralf S. Engelschall]
9820 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9821 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9822 all available cipers including rc5, which was forgotten until now.
9823 In order to let the testing shell script know which algorithms
9824 are available, a new (up to now undocumented) command
9825 "openssl list-cipher-commands" is used.
9828 *) Bugfix: s_client occasionally would sleep in select() when
9829 it should have checked SSL_pending() first.
9832 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9833 the raw DSA values prior to ASN.1 encoding.
9836 *) Tweaks to Configure
9837 [Niels Poppe <niels@netbox.org>]
9839 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9843 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9846 *) New config option to avoid instructions that are illegal on the 80386.
9847 The default code is faster, but requires at least a 486.
9850 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9851 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9852 same as SSL2_VERSION anyway.
9855 *) New "-showcerts" option for s_client.
9858 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9859 application. Various cleanups and fixes.
9862 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9863 modify error routines to work internally. Add error codes and PBE init
9864 to library startup routines.
9867 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9868 packing functions to asn1 and evp. Changed function names and error
9869 codes along the way.
9872 *) PKCS12 integration: and so it begins... First of several patches to
9873 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9874 objects to objects.h
9877 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9878 and display support for Thawte strong extranet extension.
9881 *) Add LinuxPPC support.
9882 [Jeff Dubrule <igor@pobox.org>]
9884 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9885 bn_div_words in alpha.s.
9886 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9888 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9889 OAEP isn't supported when OpenSSL is built with RSAref.
9890 [Ulf Moeller <ulf@fitug.de>]
9892 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9893 so they no longer are missing under -DNOPROTO.
9894 [Soren S. Jorvang <soren@t.dk>]
9897 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9899 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9900 doesn't work when the session is reused. Coming soon!
9903 *) Fix a security hole, that allows sessions to be reused in the wrong
9904 context thus bypassing client cert protection! All software that uses
9905 client certs and session caches in multiple contexts NEEDS PATCHING to
9906 allow session reuse! A fuller solution is in the works.
9907 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9909 *) Some more source tree cleanups (removed obsolete files
9910 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9911 permission on "config" script to be executable) and a fix for the INSTALL
9913 [Ulf Moeller <ulf@fitug.de>]
9915 *) Remove some legacy and erroneous uses of malloc, free instead of
9917 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9919 *) Make rsa_oaep_test return non-zero on error.
9920 [Ulf Moeller <ulf@fitug.de>]
9922 *) Add support for native Solaris shared libraries. Configure
9923 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9924 if someone would make that last step automatic.
9925 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9927 *) ctx_size was not built with the right compiler during "make links". Fixed.
9930 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9931 except NULL ciphers". This means the default cipher list will no longer
9932 enable NULL ciphers. They need to be specifically enabled e.g. with
9933 the string "DEFAULT:eNULL".
9936 *) Fix to RSA private encryption routines: if p < q then it would
9937 occasionally produce an invalid result. This will only happen with
9938 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9941 *) Be less restrictive and allow also `perl util/perlpath.pl
9942 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9943 because this way one can also use an interpreter named `perl5' (which is
9944 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9945 installed as `perl').
9946 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9948 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9949 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9951 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9952 advapi32.lib to Win32 build and change the pem test comparision
9953 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9954 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9955 and crypto/des/ede_cbcm_enc.c.
9958 *) DES quad checksum was broken on big-endian architectures. Fixed.
9961 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9962 Win32 test batch file so it (might) work again. The Win32 test batch file
9963 is horrible: I feel ill....
9966 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9967 in e_os.h. Audit of header files to check ANSI and non ANSI
9968 sections: 10 functions were absent from non ANSI section and not exported
9969 from Windows DLLs. Fixed up libeay.num for new functions.
9972 *) Make `openssl version' output lines consistent.
9973 [Ralf S. Engelschall]
9975 *) Fix Win32 symbol export lists for BIO functions: Added
9976 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9977 to ms/libeay{16,32}.def.
9978 [Ralf S. Engelschall]
9980 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9981 fine under Unix and passes some trivial tests I've now added. But the
9982 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9983 added to make sure no one expects that this stuff really works in the
9984 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9985 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9987 [Ralf S. Engelschall]
9989 *) Fix the generation of two part addresses in perl.
9990 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9992 *) Add config entry for Linux on MIPS.
9993 [John Tobey <jtobey@channel1.com>]
9995 *) Make links whenever Configure is run, unless we are on Windoze.
9998 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9999 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10003 *) Add a useful kludge to allow package maintainers to specify compiler and
10004 other platforms details on the command line without having to patch the
10005 Configure script everytime: One now can use ``perl Configure
10006 <id>:<details>'', i.e. platform ids are allowed to have details appended
10007 to them (seperated by colons). This is treated as there would be a static
10008 pre-configured entry in Configure's %table under key <id> with value
10009 <details> and ``perl Configure <id>'' is called. So, when you want to
10010 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10011 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10012 now, which overrides the FreeBSD-elf entry on-the-fly.
10013 [Ralf S. Engelschall]
10015 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10018 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10019 on the `perl Configure ...' command line. This way one can compile
10020 OpenSSL libraries with Position Independent Code (PIC) which is needed
10021 for linking it into DSOs.
10022 [Ralf S. Engelschall]
10024 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10028 *) Cleaned up the LICENSE document: The official contact for any license
10029 questions now is the OpenSSL core team under openssl-core@openssl.org.
10030 And add a paragraph about the dual-license situation to make sure people
10031 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10032 to the OpenSSL toolkit.
10033 [Ralf S. Engelschall]
10035 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10036 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10037 Additonally cleaned up the `make links' target: Remove unnecessary
10038 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10039 to speed processing and no longer clutter the display with confusing
10040 stuff. Instead only the actually done links are displayed.
10041 [Ralf S. Engelschall]
10043 *) Permit null encryption ciphersuites, used for authentication only. It used
10044 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10045 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10049 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10050 signed attributes when verifying signatures (this would break them),
10051 the detached data encoding was wrong and public keys obtained using
10052 X509_get_pubkey() weren't freed.
10055 *) Add text documentation for the BUFFER functions. Also added a work around
10056 to a Win95 console bug. This was triggered by the password read stuff: the
10057 last character typed gets carried over to the next fread(). If you were
10058 generating a new cert request using 'req' for example then the last
10059 character of the passphrase would be CR which would then enter the first
10063 *) Added the new `Includes OpenSSL Cryptography Software' button as
10064 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10065 button and can be used by applications based on OpenSSL to show the
10066 relationship to the OpenSSL project.
10067 [Ralf S. Engelschall]
10069 *) Remove confusing variables in function signatures in files
10070 ssl/ssl_lib.c and ssl/ssl.h.
10071 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10073 *) Don't install bss_file.c under PREFIX/include/
10074 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10076 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10077 functions that return function pointers and has support for NT specific
10078 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10079 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10080 unsigned to signed types: this was killing the Win32 compile.
10083 *) Add new certificate file to stack functions,
10084 SSL_add_dir_cert_subjects_to_stack() and
10085 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10086 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10087 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10088 This means that Apache-SSL and similar packages don't have to mess around
10089 to add as many CAs as they want to the preferred list.
10092 *) Experiment with doxygen documentation. Currently only partially applied to
10094 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10095 openssl.doxy as the configuration file.
10098 *) Get rid of remaining C++-style comments which strict C compilers hate.
10099 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10101 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10102 compiled in by default: it has problems with large keys.
10105 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10106 DH private keys and/or callback functions which directly correspond to
10107 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10108 is needed for applications which have to configure certificates on a
10109 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10111 For the RSA certificate situation is makes no difference, but
10112 for the DSA certificate situation this fixes the "no shared cipher"
10113 problem where the OpenSSL cipher selection procedure failed because the
10114 temporary keys were not overtaken from the context and the API provided
10115 no way to reconfigure them.
10116 The new functions now let applications reconfigure the stuff and they
10117 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
10118 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
10119 non-public-API function ssl_cert_instantiate() is used as a helper
10120 function and also to reduce code redundancy inside ssl_rsa.c.
10121 [Ralf S. Engelschall]
10123 *) Move s_server -dcert and -dkey options out of the undocumented feature
10124 area because they are useful for the DSA situation and should be
10125 recognized by the users.
10126 [Ralf S. Engelschall]
10128 *) Fix the cipher decision scheme for export ciphers: the export bits are
10129 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
10130 SSL_EXP_MASK. So, the original variable has to be used instead of the
10131 already masked variable.
10132 [Richard Levitte <levitte@stacken.kth.se>]
10134 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
10135 [Richard Levitte <levitte@stacken.kth.se>]
10137 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
10138 from `int' to `unsigned int' because it's a length and initialized by
10139 EVP_DigestFinal() which expects an `unsigned int *'.
10140 [Richard Levitte <levitte@stacken.kth.se>]
10142 *) Don't hard-code path to Perl interpreter on shebang line of Configure
10143 script. Instead use the usual Shell->Perl transition trick.
10144 [Ralf S. Engelschall]
10146 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
10147 (in addition to RSA certificates) to match the behaviour of `openssl dsa
10148 -noout -modulus' as it's already the case for `openssl rsa -noout
10149 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
10150 currently the public key is printed (a decision which was already done by
10151 `openssl dsa -modulus' in the past) which serves a similar purpose.
10152 Additionally the NO_RSA no longer completely removes the whole -modulus
10153 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
10155 [Ralf S. Engelschall]
10157 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
10158 BIO. See the source (crypto/evp/bio_ok.c) for more info.
10159 [Arne Ansper <arne@ats.cyber.ee>]
10161 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
10162 to be added. Now both 'req' and 'ca' can use new objects defined in the
10166 *) Add cool BIO that does syslog (or event log on NT).
10167 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
10169 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
10170 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
10171 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
10172 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
10175 *) Add preliminary config info for new extension code.
10178 *) Make RSA_NO_PADDING really use no padding.
10179 [Ulf Moeller <ulf@fitug.de>]
10181 *) Generate errors when private/public key check is done.
10184 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
10185 for some CRL extensions and new objects added.
10188 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
10189 key usage extension and fuller support for authority key id.
10192 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
10193 padding method for RSA, which is recommended for new applications in PKCS
10194 #1 v2.0 (RFC 2437, October 1998).
10195 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
10196 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
10197 against Bleichbacher's attack on RSA.
10198 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
10201 *) Updates to the new SSL compression code
10202 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10204 *) Fix so that the version number in the master secret, when passed
10205 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
10206 (because the server will not accept higher), that the version number
10207 is 0x03,0x01, not 0x03,0x00
10208 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10210 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
10211 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
10212 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
10215 *) Support for RAW extensions where an arbitrary extension can be
10216 created by including its DER encoding. See apps/openssl.cnf for
10220 *) Make sure latest Perl versions don't interpret some generated C array
10221 code as Perl array code in the crypto/err/err_genc.pl script.
10222 [Lars Weber <3weber@informatik.uni-hamburg.de>]
10224 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
10225 not many people have the assembler. Various Win32 compilation fixes and
10226 update to the INSTALL.W32 file with (hopefully) more accurate Win32
10227 build instructions.
10230 *) Modify configure script 'Configure' to automatically create crypto/date.h
10231 file under Win32 and also build pem.h from pem.org. New script
10232 util/mkfiles.pl to create the MINFO file on environments that can't do a
10233 'make files': perl util/mkfiles.pl >MINFO should work.
10236 *) Major rework of DES function declarations, in the pursuit of correctness
10237 and purity. As a result, many evil casts evaporated, and some weirdness,
10238 too. You may find this causes warnings in your code. Zapping your evil
10239 casts will probably fix them. Mostly.
10242 *) Fix for a typo in asn1.h. Bug fix to object creation script
10243 obj_dat.pl. It considered a zero in an object definition to mean
10244 "end of object": none of the objects in objects.h have any zeros
10245 so it wasn't spotted.
10246 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
10248 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
10249 Masking (CBCM). In the absence of test vectors, the best I have been able
10250 to do is check that the decrypt undoes the encrypt, so far. Send me test
10251 vectors if you have them.
10254 *) Correct calculation of key length for export ciphers (too much space was
10255 allocated for null ciphers). This has not been tested!
10258 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
10259 message is now correct (it understands "crypto" and "ssl" on its
10260 command line). There is also now an "update" option. This will update
10261 the util/ssleay.num and util/libeay.num files with any new functions.
10263 perl util/mkdef.pl crypto ssl update
10264 it will update them.
10267 *) Overhauled the Perl interface (perl/*):
10268 - ported BN stuff to OpenSSL's different BN library
10269 - made the perl/ source tree CVS-aware
10270 - renamed the package from SSLeay to OpenSSL (the files still contain
10271 their history because I've copied them in the repository)
10272 - removed obsolete files (the test scripts will be replaced
10273 by better Test::Harness variants in the future)
10274 [Ralf S. Engelschall]
10276 *) First cut for a very conservative source tree cleanup:
10277 1. merge various obsolete readme texts into doc/ssleay.txt
10278 where we collect the old documents and readme texts.
10279 2. remove the first part of files where I'm already sure that we no
10280 longer need them because of three reasons: either they are just temporary
10281 files which were left by Eric or they are preserved original files where
10282 I've verified that the diff is also available in the CVS via "cvs diff
10283 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
10284 the crypto/md/ stuff).
10285 [Ralf S. Engelschall]
10287 *) More extension code. Incomplete support for subject and issuer alt
10288 name, issuer and authority key id. Change the i2v function parameters
10289 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
10290 what that's for :-) Fix to ASN1 macro which messed up
10291 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
10294 *) Preliminary support for ENUMERATED type. This is largely copied from the
10298 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
10299 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10301 *) Make sure `make rehash' target really finds the `openssl' program.
10302 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10304 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
10305 like to hear about it if this slows down other processors.
10308 *) Add CygWin32 platform information to Configure script.
10309 [Alan Batie <batie@aahz.jf.intel.com>]
10311 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
10312 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
10314 *) New program nseq to manipulate netscape certificate sequences
10317 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
10321 *) Fixes to BN code. Previously the default was to define BN_RECURSION
10322 but the BN code had some problems that would cause failures when
10323 doing certificate verification and some other functions.
10324 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10326 *) Add ASN1 and PEM code to support netscape certificate sequences.
10329 *) Add ASN1 and PEM code to support netscape certificate sequences.
10332 *) Add several PKIX and private extended key usage OIDs.
10335 *) Modify the 'ca' program to handle the new extension code. Modify
10336 openssl.cnf for new extension format, add comments.
10339 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
10340 and add a sample to openssl.cnf so req -x509 now adds appropriate
10344 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
10345 error code, add initial support to X509_print() and x509 application.
10348 *) Takes a deep breath and start addding X509 V3 extension support code. Add
10349 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
10350 stuff is currently isolated and isn't even compiled yet.
10353 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
10354 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
10355 Removed the versions check from X509 routines when loading extensions:
10356 this allows certain broken certificates that don't set the version
10357 properly to be processed.
10360 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
10361 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
10362 can still be regenerated with "make depend".
10365 *) Spelling mistake in C version of CAST-128.
10366 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
10368 *) Changes to the error generation code. The perl script err-code.pl
10369 now reads in the old error codes and retains the old numbers, only
10370 adding new ones if necessary. It also only changes the .err files if new
10371 codes are added. The makefiles have been modified to only insert errors
10372 when needed (to avoid needlessly modifying header files). This is done
10373 by only inserting errors if the .err file is newer than the auto generated
10374 C file. To rebuild all the error codes from scratch (the old behaviour)
10375 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
10376 or delete all the .err files.
10379 *) CAST-128 was incorrectly implemented for short keys. The C version has
10380 been fixed, but is untested. The assembler versions are also fixed, but
10381 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
10382 to regenerate it if needed.
10383 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
10384 Hagino <itojun@kame.net>]
10386 *) File was opened incorrectly in randfile.c.
10387 [Ulf Möller <ulf@fitug.de>]
10389 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
10390 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
10391 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
10392 al: it's just almost always a UTCTime. Note this patch adds new error
10393 codes so do a "make errors" if there are problems.
10396 *) Correct Linux 1 recognition in config.
10397 [Ulf Möller <ulf@fitug.de>]
10399 *) Remove pointless MD5 hash when using DSA keys in ca.
10400 [Anonymous <nobody@replay.com>]
10402 *) Generate an error if given an empty string as a cert directory. Also
10403 generate an error if handed NULL (previously returned 0 to indicate an
10404 error, but didn't set one).
10405 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
10407 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
10410 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
10411 parameters. This was causing a warning which killed off the Win32 compile.
10414 *) Remove C++ style comments from crypto/bn/bn_local.h.
10415 [Neil Costigan <neil.costigan@celocom.com>]
10417 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
10418 based on a text string, looking up short and long names and finally
10419 "dot" format. The "dot" format stuff didn't work. Added new function
10420 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
10421 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
10422 OID is not part of the table.
10425 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
10426 X509_LOOKUP_by_alias().
10429 *) Sort openssl functions by name.
10432 *) Get the gendsa program working (hopefully) and add it to app list. Remove
10433 encryption from sample DSA keys (in case anyone is interested the password
10437 *) Make _all_ *_free functions accept a NULL pointer.
10438 [Frans Heymans <fheymans@isaserver.be>]
10440 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10442 [Anonymous <nobody@replay.com>]
10444 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10445 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10447 *) Don't blow it for numeric -newkey arguments to apps/req.
10448 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10450 *) Temp key "for export" tests were wrong in s3_srvr.c.
10451 [Anonymous <nobody@replay.com>]
10453 *) Add prototype for temp key callback functions
10454 SSL_CTX_set_tmp_{rsa,dh}_callback().
10457 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10458 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10461 *) X509_name_add_entry() freed the wrong thing after an error.
10462 [Arne Ansper <arne@ats.cyber.ee>]
10464 *) rsa_eay.c would attempt to free a NULL context.
10465 [Arne Ansper <arne@ats.cyber.ee>]
10467 *) BIO_s_socket() had a broken should_retry() on Windoze.
10468 [Arne Ansper <arne@ats.cyber.ee>]
10470 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10471 [Arne Ansper <arne@ats.cyber.ee>]
10473 *) Make sure the already existing X509_STORE->depth variable is initialized
10474 in X509_STORE_new(), but document the fact that this variable is still
10475 unused in the certificate verification process.
10476 [Ralf S. Engelschall]
10478 *) Fix the various library and apps files to free up pkeys obtained from
10479 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10482 *) Fix reference counting in X509_PUBKEY_get(). This makes
10483 demos/maurice/example2.c work, amongst others, probably.
10484 [Steve Henson and Ben Laurie]
10486 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10487 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10488 are no longer created. This way we have a single and consistent command
10489 line interface `openssl <command>', similar to `cvs <command>'.
10490 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10492 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10493 BIT STRING wrapper always have zero unused bits.
10496 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10499 *) Make the top-level INSTALL documentation easier to understand.
10502 *) Makefiles updated to exit if an error occurs in a sub-directory
10503 make (including if user presses ^C) [Paul Sutton]
10505 *) Make Montgomery context stuff explicit in RSA data structure.
10508 *) Fix build order of pem and err to allow for generated pem.h.
10511 *) Fix renumbering bug in X509_NAME_delete_entry().
10514 *) Enhanced the err-ins.pl script so it makes the error library number
10515 global and can add a library name. This is needed for external ASN1 and
10516 other error libraries.
10519 *) Fixed sk_insert which never worked properly.
10522 *) Fix ASN1 macros so they can handle indefinite length construted
10523 EXPLICIT tags. Some non standard certificates use these: they can now
10527 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10528 into a single doc/ssleay.txt bundle. This way the information is still
10529 preserved but no longer messes up this directory. Now it's new room for
10530 the new set of documenation files.
10531 [Ralf S. Engelschall]
10533 *) SETs were incorrectly DER encoded. This was a major pain, because they
10534 shared code with SEQUENCEs, which aren't coded the same. This means that
10535 almost everything to do with SETs or SEQUENCEs has either changed name or
10536 number of arguments.
10537 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10539 *) Fix test data to work with the above.
10542 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10543 was already fixed by Eric for 0.9.1 it seems.
10544 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10546 *) Autodetect FreeBSD3.
10549 *) Fix various bugs in Configure. This affects the following platforms:
10553 unixware-2.0-pentium
10557 *) Eliminate generated files from CVS. Reorder tests to regenerate files
10558 before they are needed.
10561 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
10565 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
10567 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
10568 changed SSLeay to OpenSSL in version strings.
10569 [Ralf S. Engelschall]
10571 *) Some fixups to the top-level documents.
10574 *) Fixed the nasty bug where rsaref.h was not found under compile-time
10575 because the symlink to include/ was missing.
10576 [Ralf S. Engelschall]
10578 *) Incorporated the popular no-RSA/DSA-only patches
10579 which allow to compile a RSA-free SSLeay.
10580 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
10582 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
10583 when "ssleay" is still not found.
10584 [Ralf S. Engelschall]
10586 *) Added more platforms to Configure: Cray T3E, HPUX 11,
10587 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
10589 *) Updated the README file.
10590 [Ralf S. Engelschall]
10592 *) Added various .cvsignore files in the CVS repository subdirs
10593 to make a "cvs update" really silent.
10594 [Ralf S. Engelschall]
10596 *) Recompiled the error-definition header files and added
10597 missing symbols to the Win32 linker tables.
10598 [Ralf S. Engelschall]
10600 *) Cleaned up the top-level documents;
10601 o new files: CHANGES and LICENSE
10602 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
10603 o merged COPYRIGHT into LICENSE
10604 o removed obsolete TODO file
10605 o renamed MICROSOFT to INSTALL.W32
10606 [Ralf S. Engelschall]
10608 *) Removed dummy files from the 0.9.1b source tree:
10609 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
10610 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
10611 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
10612 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
10613 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
10614 [Ralf S. Engelschall]
10616 *) Added various platform portability fixes.
10619 *) The Genesis of the OpenSSL rpject:
10620 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
10621 Young and Tim J. Hudson created while they were working for C2Net until
10623 [The OpenSSL Project]
10626 Changes between 0.9.0b and 0.9.1b [not released]
10628 *) Updated a few CA certificates under certs/
10631 *) Changed some BIGNUM api stuff.
10634 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
10635 DGUX x86, Linux Alpha, etc.
10638 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
10639 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
10643 *) Add -strparse option to asn1pars program which parses nested
10645 [Dr Stephen Henson <shenson@bigfoot.com>]
10647 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
10650 *) DSA fix for "ca" program.
10653 *) Added "-genkey" option to "dsaparam" program.
10656 *) Added RIPE MD160 (rmd160) message digest.
10659 *) Added -a (all) option to "ssleay version" command.
10662 *) Added PLATFORM define which is the id given to Configure.
10665 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
10668 *) Extended the ASN.1 parser routines.
10671 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
10674 *) Added a BN_CTX to the BN library.
10677 *) Fixed the weak key values in DES library
10680 *) Changed API in EVP library for cipher aliases.
10683 *) Added support for RC2/64bit cipher.
10686 *) Converted the lhash library to the crypto/mem.c functions.
10689 *) Added more recognized ASN.1 object ids.
10692 *) Added more RSA padding checks for SSL/TLS.
10695 *) Added BIO proxy/filter functionality.
10698 *) Added extra_certs to SSL_CTX which can be used
10699 send extra CA certificates to the client in the CA cert chain sending
10700 process. It can be configured with SSL_CTX_add_extra_chain_cert().
10703 *) Now Fortezza is denied in the authentication phase because
10704 this is key exchange mechanism is not supported by SSLeay at all.
10707 *) Additional PKCS1 checks.
10710 *) Support the string "TLSv1" for all TLS v1 ciphers.
10713 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
10714 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
10717 *) Fixed a few memory leaks.
10720 *) Fixed various code and comment typos.
10723 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
10724 bytes sent in the client random.
10725 [Edward Bishop <ebishop@spyglass.com>]