5 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
7 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
10 *) Fix bug in string printing code: if *any* escaping is enabled we must
11 escape the escape character (backslash) or the resulting string is
15 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
17 *) Disable code workaround for ancient and obsolete Netscape browsers
18 and servers: an attacker can use it in a ciphersuite downgrade attack.
19 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
22 *) Fixed J-PAKE implementation error, originally discovered by
23 Sebastien Martini, further info and confirmation from Stefan
24 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
27 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
29 *) Fix extension code to avoid race conditions which can result in a buffer
30 overrun vulnerability: resumed sessions must not be modified as they can
31 be shared by multiple threads. CVE-2010-3864
34 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
37 *) Don't reencode certificate when calculating signature: cache and use
38 the original encoding instead. This makes signature verification of
39 some broken encodings work correctly.
42 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
43 is also one of the inputs.
44 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
46 *) Don't repeatedly append PBE algorithms to table if they already exist.
47 Sort table on each new add. This effectively makes the table read only
48 after all algorithms are added and subsequent calls to PKCS12_pbe_add
52 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
54 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
57 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
58 access or freeing data twice (CVE-2010-0742)
59 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
61 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
62 common in certificates and some applications which only call
63 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
67 Reduce copying into .apps and .test in makevms.com
68 Don't try to use blank CA certificate in CA.com
69 Allow use of C files from original directories in maketests.com
70 [Steven M. Schweda" <sms@antinode.info>]
72 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
74 *) When rejecting SSL/TLS records due to an incorrect version number, never
75 update s->server with a new major version number. As of
76 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
77 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
78 the previous behavior could result in a read attempt at NULL when
79 receiving specific incorrect SSL/TLS records once record payload
80 protection is active. (CVE-2010-0740)
81 [Bodo Moeller, Adam Langley <agl@chromium.org>]
83 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
84 could be crashed if the relevant tables were not present (e.g. chrooted).
85 [Tomas Hoger <thoger@redhat.com>]
87 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
89 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
90 [Martin Olsson, Neel Mehta]
92 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
93 accommodate for stack sorting, always a write lock!).
96 *) On some versions of WIN32 Heap32Next is very slow. This can cause
97 excessive delays in the RAND_poll(): over a minute. As a workaround
98 include a time check in the inner Heap32Next loop too.
101 *) The code that handled flushing of data in SSL/TLS originally used the
102 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
103 the problem outlined in PR#1949. The fix suggested there however can
104 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
105 of Apache). So instead simplify the code to flush unconditionally.
106 This should be fine since flushing with no data to flush is a no op.
109 *) Handle TLS versions 2.0 and later properly and correctly use the
110 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
111 off ancient servers have a habit of sticking around for a while...
114 *) Modify compression code so it frees up structures without using the
115 ex_data callbacks. This works around a problem where some applications
116 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
117 restarting) then use compression (e.g. SSL with compression) later.
118 This results in significant per-connection memory leaks and
119 has caused some security issues including CVE-2008-1678 and
123 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
124 change when encrypting or decrypting.
127 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
128 connect and renegotiate with servers which do not support RI.
129 Until RI is more widely deployed this option is enabled by default.
132 *) Add "missing" ssl ctrls to clear options and mode.
135 *) If client attempts to renegotiate and doesn't support RI respond with
136 a no_renegotiation alert as required by RFC5746. Some renegotiating
137 TLS clients will continue a connection gracefully when they receive
138 the alert. Unfortunately OpenSSL mishandled this alert and would hang
139 waiting for a server hello which it will never receive. Now we treat a
140 received no_renegotiation alert as a fatal error. This is because
141 applications requesting a renegotiation might well expect it to succeed
142 and would have no code in place to handle the server denying it so the
143 only safe thing to do is to terminate the connection.
146 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
147 peer supports secure renegotiation and 0 otherwise. Print out peer
148 renegotiation support in s_client/s_server.
151 *) Replace the highly broken and deprecated SPKAC certification method with
152 the updated NID creation version. This should correctly handle UTF8.
155 *) Implement RFC5746. Re-enable renegotiation but require the extension
156 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
157 turns out to be a bad idea. It has been replaced by
158 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
159 SSL_CTX_set_options(). This is really not recommended unless you
160 know what you are doing.
161 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
163 *) Fixes to stateless session resumption handling. Use initial_ctx when
164 issuing and attempting to decrypt tickets in case it has changed during
165 servername handling. Use a non-zero length session ID when attempting
166 stateless session resumption: this makes it possible to determine if
167 a resumption has occurred immediately after receiving server hello
168 (several places in OpenSSL subtly assume this) instead of later in
172 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
173 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
174 fixes for a few places where the return code is not checked
176 [Julia Lawall <julia@diku.dk>]
178 *) Add --strict-warnings option to Configure script to include devteam
179 warnings in other configurations.
182 *) Add support for --libdir option and LIBDIR variable in makefiles. This
183 makes it possible to install openssl libraries in locations which
184 have names other than "lib", for example "/usr/lib64" which some
186 [Steve Henson, based on patch from Jeremy Utley]
188 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
189 X690 8.9.12 and can produce some misleading textual output of OIDs.
190 [Steve Henson, reported by Dan Kaminsky]
192 *) Delete MD2 from algorithm tables. This follows the recommendation in
193 several standards that it is not used in new applications due to
194 several cryptographic weaknesses. For binary compatibility reasons
195 the MD2 API is still compiled in by default.
198 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
202 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
203 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
205 [Guenter <lists@gknw.net>]
207 *) Fix the server certificate chain building code to use X509_verify_cert(),
208 it used to have an ad-hoc builder which was unable to cope with anything
209 other than a simple chain.
210 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
212 *) Don't check self signed certificate signatures in X509_verify_cert()
213 by default (a flag can override this): it just wastes time without
214 adding any security. As a useful side effect self signed root CAs
215 with non-FIPS digests are now usable in FIPS mode.
218 *) In dtls1_process_out_of_seq_message() the check if the current message
219 is already buffered was missing. For every new message was memory
220 allocated, allowing an attacker to perform an denial of service attack
221 with sending out of seq handshake messages until there is no memory
222 left. Additionally every future messege was buffered, even if the
223 sequence number made no sense and would be part of another handshake.
224 So only messages with sequence numbers less than 10 in advance will be
225 buffered. (CVE-2009-1378)
226 [Robin Seggelmann, discovered by Daniel Mentz]
228 *) Records are buffered if they arrive with a future epoch to be
229 processed after finishing the corresponding handshake. There is
230 currently no limitation to this buffer allowing an attacker to perform
231 a DOS attack with sending records with future epochs until there is no
232 memory left. This patch adds the pqueue_size() function to detemine
233 the size of a buffer and limits the record buffer to 100 entries.
235 [Robin Seggelmann, discovered by Daniel Mentz]
237 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
238 parent structure is freed. (CVE-2009-1379)
241 *) Handle non-blocking I/O properly in SSL_shutdown() call.
242 [Darryl Miles <darryl-mailinglists@netbauds.net>]
245 [Ilya O. <vrghost@gmail.com>]
247 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
249 *) Disable renegotiation completely - this fixes a severe security
250 problem (CVE-2009-3555) at the cost of breaking all
251 renegotiation. Renegotiation can be re-enabled by setting
252 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
253 run-time. This is really not recommended unless you know what
257 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
259 *) Don't set val to NULL when freeing up structures, it is freed up by
260 underlying code. If sizeof(void *) > sizeof(long) this can result in
261 zeroing past the valid field. (CVE-2009-0789)
262 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
264 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
265 checked correctly. This would allow some invalid signed attributes to
266 appear to verify correctly. (CVE-2009-0591)
267 [Ivan Nestlerode <inestlerode@us.ibm.com>]
269 *) Reject UniversalString and BMPString types with invalid lengths. This
270 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
271 a legal length. (CVE-2009-0590)
274 *) Set S/MIME signing as the default purpose rather than setting it
275 unconditionally. This allows applications to override it at the store
279 *) Permit restricted recursion of ASN1 strings. This is needed in practice
280 to handle some structures.
283 *) Improve efficiency of mem_gets: don't search whole buffer each time
285 [Jeremy Shapiro <jnshapir@us.ibm.com>]
287 *) New -hex option for openssl rand.
290 *) Print out UTF8String and NumericString when parsing ASN1.
293 *) Support NumericString type for name components.
296 *) Allow CC in the environment to override the automatically chosen
297 compiler. Note that nothing is done to ensure flags work with the
301 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
303 *) Properly check EVP_VerifyFinal() and similar return values
305 [Ben Laurie, Bodo Moeller, Google Security Team]
307 *) Enable TLS extensions by default.
310 *) Allow the CHIL engine to be loaded, whether the application is
311 multithreaded or not. (This does not release the developer from the
312 obligation to set up the dynamic locking callbacks.)
313 [Sander Temme <sander@temme.net>]
315 *) Use correct exit code if there is an error in dgst command.
316 [Steve Henson; problem pointed out by Roland Dirlewanger]
318 *) Tweak Configure so that you need to say "experimental-jpake" to enable
319 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
322 *) Add experimental JPAKE support, including demo authentication in
323 s_client and s_server.
326 *) Set the comparison function in v3_addr_canonize().
327 [Rob Austein <sra@hactrn.net>]
329 *) Add support for XMPP STARTTLS in s_client.
330 [Philip Paeps <philip@freebsd.org>]
332 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
333 to ensure that even with this option, only ciphersuites in the
334 server's preference list will be accepted. (Note that the option
335 applies only when resuming a session, so the earlier behavior was
336 just about the algorithm choice for symmetric cryptography.)
339 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
341 *) Fix NULL pointer dereference if a DTLS server received
342 ChangeCipherSpec as first record (CVE-2009-1386).
345 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
346 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
349 *) The fix in 0.9.8c that supposedly got rid of unsafe
350 double-checked locking was incomplete for RSA blinding,
351 addressing just one layer of what turns out to have been
352 doubly unsafe triple-checked locking.
354 So now fix this for real by retiring the MONT_HELPER macro
355 in crypto/rsa/rsa_eay.c.
357 [Bodo Moeller; problem pointed out by Marius Schilder]
359 *) Various precautionary measures:
361 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
363 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
364 (NB: This would require knowledge of the secret session ticket key
365 to exploit, in which case you'd be SOL either way.)
367 - Change bn_nist.c so that it will properly handle input BIGNUMs
368 outside the expected range.
370 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
373 [Neel Mehta, Bodo Moeller]
375 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
376 the load fails. Useful for distros.
377 [Ben Laurie and the FreeBSD team]
379 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
382 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
385 *) Expand ENGINE to support engine supplied SSL client certificate functions.
387 This work was sponsored by Logica.
390 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
391 keystores. Support for SSL/TLS client authentication too.
392 Not compiled unless enable-capieng specified to Configure.
394 This work was sponsored by Logica.
397 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
398 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
399 attribute creation routines such as certifcate requests and PKCS#12
403 Changes between 0.9.8g and 0.9.8h [28 May 2008]
405 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
406 handshake which could lead to a cilent crash as found using the
407 Codenomicon TLS test suite (CVE-2008-1672)
408 [Steve Henson, Mark Cox]
410 *) Fix double free in TLS server name extensions which could lead to
411 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
414 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
416 Clear the error queue to ensure that error entries left from
417 older function calls do not interfere with the correct operation.
418 [Lutz Jaenicke, Erik de Castro Lopo]
420 *) Remove root CA certificates of commercial CAs:
422 The OpenSSL project does not recommend any specific CA and does not
423 have any policy with respect to including or excluding any CA.
424 Therefore it does not make any sense to ship an arbitrary selection
425 of root CA certificates with the OpenSSL software.
428 *) RSA OAEP patches to fix two separate invalid memory reads.
429 The first one involves inputs when 'lzero' is greater than
430 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
431 before the beginning of from). The second one involves inputs where
432 the 'db' section contains nothing but zeroes (there is a one-byte
433 invalid read after the end of 'db').
434 [Ivan Nestlerode <inestlerode@us.ibm.com>]
436 *) Partial backport from 0.9.9-dev:
438 Introduce bn_mul_mont (dedicated Montgomery multiplication
439 procedure) as a candidate for BIGNUM assembler implementation.
440 While 0.9.9-dev uses assembler for various architectures, only
441 x86_64 is available by default here in the 0.9.8 branch, and
442 32-bit x86 is available through a compile-time setting.
444 To try the 32-bit x86 assembler implementation, use Configure
445 option "enable-montasm" (which exists only for this backport).
447 As "enable-montasm" for 32-bit x86 disclaims code stability
448 anyway, in this constellation we activate additional code
449 backported from 0.9.9-dev for further performance improvements,
450 namely BN_from_montgomery_word. (To enable this otherwise,
451 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
453 [Andy Polyakov (backport partially by Bodo Moeller)]
455 *) Add TLS session ticket callback. This allows an application to set
456 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
457 values. This is useful for key rollover for example where several key
458 sets may exist with different names.
461 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
462 This was broken until now in 0.9.8 releases, such that the only way
463 a registered ENGINE could be used (assuming it initialises
464 successfully on the host) was to explicitly set it as the default
465 for the relevant algorithms. This is in contradiction with 0.9.7
466 behaviour and the documentation. With this fix, when an ENGINE is
467 registered into a given algorithm's table of implementations, the
468 'uptodate' flag is reset so that auto-discovery will be used next
469 time a new context for that algorithm attempts to select an
471 [Ian Lister (tweaked by Geoff Thorpe)]
473 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
474 implemention in the following ways:
476 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
479 Lack of BER streaming support means one pass streaming processing is
480 only supported if data is detached: setting the streaming flag is
481 ignored for embedded content.
483 CMS support is disabled by default and must be explicitly enabled
484 with the enable-cms configuration option.
487 *) Update the GMP engine glue to do direct copies between BIGNUM and
488 mpz_t when openssl and GMP use the same limb size. Otherwise the
489 existing "conversion via a text string export" trick is still used.
490 [Paul Sheer <paulsheer@gmail.com>]
492 *) Zlib compression BIO. This is a filter BIO which compressed and
493 uncompresses any data passed through it.
496 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
497 RFC3394 compatible AES key wrapping.
500 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
501 sets string data without copying. X509_ALGOR_set0() and
502 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
503 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
504 from an X509_ATTRIBUTE structure optionally checking it occurs only
505 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
509 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
510 to get the expected BN_FLG_CONSTTIME behavior.
511 [Bodo Moeller (Google)]
515 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
516 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
517 - added some more tests to do_tests.pl
518 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
519 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
520 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
521 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
522 - various changes to netware.pl to enable gcc-cross builds on Win32
524 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
525 - various changes to fix missing prototype warnings
526 - fixed x86nasm.pl to create correct asm files for NASM COFF output
527 - added AES, WHIRLPOOL and CPUID assembler code to build files
528 - added missing AES assembler make rules to mk1mf.pl
529 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
530 [Guenter Knauf <eflash@gmx.net>]
532 *) Implement certificate status request TLS extension defined in RFC3546.
533 A client can set the appropriate parameters and receive the encoded
534 OCSP response via a callback. A server can query the supplied parameters
535 and set the encoded OCSP response in the callback. Add simplified examples
536 to s_client and s_server.
539 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
542 + Binary incompatibility of ssl_ctx_st structure
543 + DTLS interoperation with non-compliant servers
544 + Don't call get_session_cb() without proposed session
545 + Fix ia64 assembler code
546 [Andy Polyakov, Steve Henson]
548 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
550 *) DTLS Handshake overhaul. There were longstanding issues with
551 OpenSSL DTLS implementation, which were making it impossible for
552 RFC 4347 compliant client to communicate with OpenSSL server.
553 Unfortunately just fixing these incompatibilities would "cut off"
554 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
555 server keeps tolerating non RFC compliant syntax. The opposite is
556 not true, 0.9.8f client can not communicate with earlier server.
557 This update even addresses CVE-2007-4995.
560 *) Changes to avoid need for function casts in OpenSSL: some compilers
561 (gcc 4.2 and later) reject their use.
562 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
565 *) Add RFC4507 support to OpenSSL. This includes the corrections in
566 RFC4507bis. The encrypted ticket format is an encrypted encoded
567 SSL_SESSION structure, that way new session features are automatically
570 If a client application caches session in an SSL_SESSION structure
571 support is transparent because tickets are now stored in the encoded
574 The SSL_CTX structure automatically generates keys for ticket
575 protection in servers so again support should be possible
576 with no application modification.
578 If a client or server wishes to disable RFC4507 support then the option
579 SSL_OP_NO_TICKET can be set.
581 Add a TLS extension debugging callback to allow the contents of any client
582 or server extensions to be examined.
584 This work was sponsored by Google.
587 *) Add initial support for TLS extensions, specifically for the server_name
588 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
589 have new members for a host name. The SSL data structure has an
590 additional member SSL_CTX *initial_ctx so that new sessions can be
591 stored in that context to allow for session resumption, even after the
592 SSL has been switched to a new SSL_CTX in reaction to a client's
593 server_name extension.
595 New functions (subject to change):
598 SSL_get_servername_type()
601 New CTRL codes and macros (subject to change):
603 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
604 - SSL_CTX_set_tlsext_servername_callback()
605 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
606 - SSL_CTX_set_tlsext_servername_arg()
607 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
609 openssl s_client has a new '-servername ...' option.
611 openssl s_server has new options '-servername_host ...', '-cert2 ...',
612 '-key2 ...', '-servername_fatal' (subject to change). This allows
613 testing the HostName extension for a specific single host name ('-cert'
614 and '-key' remain fallbacks for handshakes without HostName
615 negotiation). If the unrecogninzed_name alert has to be sent, this by
616 default is a warning; it becomes fatal with the '-servername_fatal'
619 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
621 *) Add AES and SSE2 assembly language support to VC++ build.
624 *) Mitigate attack on final subtraction in Montgomery reduction.
627 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
628 (which previously caused an internal error).
631 *) Squeeze another 10% out of IGE mode when in != out.
634 *) AES IGE mode speedup.
635 [Dean Gaudet (Google)]
637 *) Add the Korean symmetric 128-bit cipher SEED (see
638 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
639 add SEED ciphersuites from RFC 4162:
641 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
642 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
643 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
644 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
646 To minimize changes between patchlevels in the OpenSSL 0.9.8
647 series, SEED remains excluded from compilation unless OpenSSL
648 is configured with 'enable-seed'.
651 *) Mitigate branch prediction attacks, which can be practical if a
652 single processor is shared, allowing a spy process to extract
653 information. For detailed background information, see
654 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
655 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
656 and Necessary Software Countermeasures"). The core of the change
657 are new versions BN_div_no_branch() and
658 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
659 respectively, which are slower, but avoid the security-relevant
660 conditional branches. These are automatically called by BN_div()
661 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
662 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
663 remove a conditional branch.
665 BN_FLG_CONSTTIME is the new name for the previous
666 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
667 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
668 in the exponent causes BN_mod_exp_mont() to use the alternative
669 implementation in BN_mod_exp_mont_consttime().) The old name
670 remains as a deprecated alias.
672 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
673 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
674 constant-time implementations for more than just exponentiation.
675 Here too the old name is kept as a deprecated alias.
677 BN_BLINDING_new() will now use BN_dup() for the modulus so that
678 the BN_BLINDING structure gets an independent copy of the
679 modulus. This means that the previous "BIGNUM *m" argument to
680 BN_BLINDING_new() and to BN_BLINDING_create_param() now
681 essentially becomes "const BIGNUM *m", although we can't actually
682 change this in the header file before 0.9.9. It allows
683 RSA_setup_blinding() to use BN_with_flags() on the modulus to
684 enable BN_FLG_CONSTTIME.
686 [Matthew D Wood (Intel Corp)]
688 *) In the SSL/TLS server implementation, be strict about session ID
689 context matching (which matters if an application uses a single
690 external cache for different purposes). Previously,
691 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
692 set. This did ensure strict client verification, but meant that,
693 with applications using a single external cache for quite
694 different requirements, clients could circumvent ciphersuite
695 restrictions for a given session ID context by starting a session
696 in a different context.
699 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
700 a ciphersuite string such as "DEFAULT:RSA" cannot enable
701 authentication-only ciphersuites.
704 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
705 not complete and could lead to a possible single byte overflow
706 (CVE-2007-5135) [Ben Laurie]
708 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
710 *) Since AES128 and AES256 (and similarly Camellia128 and
711 Camellia256) share a single mask bit in the logic of
712 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
713 kludge to work properly if AES128 is available and AES256 isn't
714 (or if Camellia128 is available and Camellia256 isn't).
717 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
718 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
719 When a point or a seed is encoded in a BIT STRING, we need to
720 prevent the removal of trailing zero bits to get the proper DER
721 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
722 of a NamedBitList, for which trailing 0 bits need to be removed.)
725 *) Have SSL/TLS server implementation tolerate "mismatched" record
726 protocol version while receiving ClientHello even if the
727 ClientHello is fragmented. (The server can't insist on the
728 particular protocol version it has chosen before the ServerHello
729 message has informed the client about his choice.)
732 *) Add RFC 3779 support.
733 [Rob Austein for ARIN, Ben Laurie]
735 *) Load error codes if they are not already present instead of using a
736 static variable. This allows them to be cleanly unloaded and reloaded.
737 Improve header file function name parsing.
740 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
741 or CAPABILITY handshake as required by RFCs.
744 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
746 *) Introduce limits to prevent malicious keys being able to
747 cause a denial of service. (CVE-2006-2940)
748 [Steve Henson, Bodo Moeller]
750 *) Fix ASN.1 parsing of certain invalid structures that can result
751 in a denial of service. (CVE-2006-2937) [Steve Henson]
753 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
754 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
756 *) Fix SSL client code which could crash if connecting to a
757 malicious SSLv2 server. (CVE-2006-4343)
758 [Tavis Ormandy and Will Drewry, Google Security Team]
760 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
761 match only those. Before that, "AES256-SHA" would be interpreted
762 as a pattern and match "AES128-SHA" too (since AES128-SHA got
763 the same strength classification in 0.9.7h) as we currently only
764 have a single AES bit in the ciphersuite description bitmap.
765 That change, however, also applied to ciphersuite strings such as
766 "RC4-MD5" that intentionally matched multiple ciphersuites --
767 namely, SSL 2.0 ciphersuites in addition to the more common ones
768 from SSL 3.0/TLS 1.0.
770 So we change the selection algorithm again: Naming an explicit
771 ciphersuite selects this one ciphersuite, and any other similar
772 ciphersuite (same bitmap) from *other* protocol versions.
773 Thus, "RC4-MD5" again will properly select both the SSL 2.0
774 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
776 Since SSL 2.0 does not have any ciphersuites for which the
777 128/256 bit distinction would be relevant, this works for now.
778 The proper fix will be to use different bits for AES128 and
779 AES256, which would have avoided the problems from the beginning;
780 however, bits are scarce, so we can only do this in a new release
781 (not just a patchlevel) when we can change the SSL_CIPHER
782 definition to split the single 'unsigned long mask' bitmap into
783 multiple values to extend the available space.
787 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
789 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
790 (CVE-2006-4339) [Ben Laurie and Google Security Team]
792 *) Add AES IGE and biIGE modes.
795 *) Change the Unix randomness entropy gathering to use poll() when
796 possible instead of select(), since the latter has some
797 undesirable limitations.
798 [Darryl Miles via Richard Levitte and Bodo Moeller]
800 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
801 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
802 cannot be implicitly activated as part of, e.g., the "AES" alias.
803 However, please upgrade to OpenSSL 0.9.9[-dev] for
804 non-experimental use of the ECC ciphersuites to get TLS extension
805 support, which is required for curve and point format negotiation
806 to avoid potential handshake problems.
809 *) Disable rogue ciphersuites:
811 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
812 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
813 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
815 The latter two were purportedly from
816 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
819 Also deactivate the remaining ciphersuites from
820 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
821 unofficial, and the ID has long expired.
824 *) Fix RSA blinding Heisenbug (problems sometimes occured on
825 dual-core machines) and other potential thread-safety issues.
828 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
829 versions), which is now available for royalty-free use
830 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
831 Also, add Camellia TLS ciphersuites from RFC 4132.
833 To minimize changes between patchlevels in the OpenSSL 0.9.8
834 series, Camellia remains excluded from compilation unless OpenSSL
835 is configured with 'enable-camellia'.
838 *) Disable the padding bug check when compression is in use. The padding
839 bug check assumes the first packet is of even length, this is not
840 necessarily true if compresssion is enabled and can result in false
841 positives causing handshake failure. The actual bug test is ancient
842 code so it is hoped that implementations will either have fixed it by
843 now or any which still have the bug do not support compression.
846 Changes between 0.9.8a and 0.9.8b [04 May 2006]
848 *) When applying a cipher rule check to see if string match is an explicit
849 cipher suite and only match that one cipher suite if it is.
852 *) Link in manifests for VC++ if needed.
853 [Austin Ziegler <halostatue@gmail.com>]
855 *) Update support for ECC-based TLS ciphersuites according to
856 draft-ietf-tls-ecc-12.txt with proposed changes (but without
857 TLS extensions, which are supported starting with the 0.9.9
858 branch, not in the OpenSSL 0.9.8 branch).
861 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
862 opaque EVP_CIPHER_CTX handling.
865 *) Fixes and enhancements to zlib compression code. We now only use
866 "zlib1.dll" and use the default __cdecl calling convention on Win32
867 to conform with the standards mentioned here:
868 http://www.zlib.net/DLL_FAQ.txt
869 Static zlib linking now works on Windows and the new --with-zlib-include
870 --with-zlib-lib options to Configure can be used to supply the location
871 of the headers and library. Gracefully handle case where zlib library
875 *) Several fixes and enhancements to the OID generation code. The old code
876 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
877 handle numbers larger than ULONG_MAX, truncated printing and had a
878 non standard OBJ_obj2txt() behaviour.
881 *) Add support for building of engines under engine/ as shared libraries
882 under VC++ build system.
885 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
886 Hopefully, we will not see any false combination of paths any more.
889 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
891 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
892 (part of SSL_OP_ALL). This option used to disable the
893 countermeasure against man-in-the-middle protocol-version
894 rollback in the SSL 2.0 server implementation, which is a bad
895 idea. (CVE-2005-2969)
897 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
898 for Information Security, National Institute of Advanced Industrial
899 Science and Technology [AIST], Japan)]
901 *) Add two function to clear and return the verify parameter flags.
904 *) Keep cipherlists sorted in the source instead of sorting them at
905 runtime, thus removing the need for a lock.
908 *) Avoid some small subgroup attacks in Diffie-Hellman.
909 [Nick Mathewson and Ben Laurie]
911 *) Add functions for well-known primes.
914 *) Extended Windows CE support.
915 [Satoshi Nakamura and Andy Polyakov]
917 *) Initialize SSL_METHOD structures at compile time instead of during
918 runtime, thus removing the need for a lock.
921 *) Make PKCS7_decrypt() work even if no certificate is supplied by
922 attempting to decrypt each encrypted key in turn. Add support to
926 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
928 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
931 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
934 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
935 key into the same file any more.
938 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
941 *) Add -utf8 command line and config file option to 'ca'.
942 [Stefan <stf@udoma.org]
944 *) Removed the macro des_crypt(), as it seems to conflict with some
945 libraries. Use DES_crypt().
948 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
949 involves renaming the source and generated shared-libs for
950 both. The engines will accept the corrected or legacy ids
951 ('ncipher' and '4758_cca' respectively) when binding. NB,
952 this only applies when building 'shared'.
953 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
955 *) Add attribute functions to EVP_PKEY structure. Modify
956 PKCS12_create() to recognize a CSP name attribute and
957 use it. Make -CSP option work again in pkcs12 utility.
960 *) Add new functionality to the bn blinding code:
961 - automatic re-creation of the BN_BLINDING parameters after
962 a fixed number of uses (currently 32)
963 - add new function for parameter creation
964 - introduce flags to control the update behaviour of the
965 BN_BLINDING parameters
966 - hide BN_BLINDING structure
967 Add a second BN_BLINDING slot to the RSA structure to improve
968 performance when a single RSA object is shared among several
972 *) Add support for DTLS.
973 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
975 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
976 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
979 *) Remove buggy and incompletet DH cert support from
980 ssl/ssl_rsa.c and ssl/s3_both.c
983 *) Use SHA-1 instead of MD5 as the default digest algorithm for
984 the apps/openssl applications.
987 *) Compile clean with "-Wall -Wmissing-prototypes
988 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
989 DEBUG_SAFESTACK must also be set.
992 *) Change ./Configure so that certain algorithms can be disabled by default.
993 The new counterpiece to "no-xxx" is "enable-xxx".
995 The patented RC5 and MDC2 algorithms will now be disabled unless
996 "enable-rc5" and "enable-mdc2", respectively, are specified.
998 (IDEA remains enabled despite being patented. This is because IDEA
999 is frequently required for interoperability, and there is no license
1000 fee for non-commercial use. As before, "no-idea" can be used to
1001 avoid this algorithm.)
1005 *) Add processing of proxy certificates (see RFC 3820). This work was
1006 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
1007 EGEE (Enabling Grids for E-science in Europe).
1010 *) RC4 performance overhaul on modern architectures/implementations, such
1011 as Intel P4, IA-64 and AMD64.
1014 *) New utility extract-section.pl. This can be used specify an alternative
1015 section number in a pod file instead of having to treat each file as
1016 a separate case in Makefile. This can be done by adding two lines to the
1019 =for comment openssl_section:XXX
1021 The blank line is mandatory.
1025 *) New arguments -certform, -keyform and -pass for s_client and s_server
1026 to allow alternative format key and certificate files and passphrase
1030 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
1031 update associated structures and add various utility functions.
1033 Add new policy related verify parameters, include policy checking in
1034 standard verify code. Enhance 'smime' application with extra parameters
1035 to support policy checking and print out.
1038 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
1039 Nehemiah processors. These extensions support AES encryption in hardware
1040 as well as RNG (though RNG support is currently disabled).
1041 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
1043 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
1046 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
1047 [Andy Polyakov and a number of other people]
1049 *) Improved PowerPC platform support. Most notably BIGNUM assembler
1050 implementation contributed by IBM.
1051 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
1053 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
1054 exponent rather than 'unsigned long'. There is a corresponding change to
1055 the new 'rsa_keygen' element of the RSA_METHOD structure.
1056 [Jelte Jansen, Geoff Thorpe]
1058 *) Functionality for creating the initial serial number file is now
1059 moved from CA.pl to the 'ca' utility with a new option -create_serial.
1061 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
1062 number file to 1, which is bound to cause problems. To avoid
1063 the problems while respecting compatibility between different 0.9.7
1064 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
1065 CA.pl for serial number initialization. With the new release 0.9.8,
1066 we can fix the problem directly in the 'ca' utility.)
1069 *) Reduced header interdepencies by declaring more opaque objects in
1070 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
1071 give fewer recursive includes, which could break lazy source code - so
1072 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
1073 developers should define this symbol when building and using openssl to
1074 ensure they track the recommended behaviour, interfaces, [etc], but
1075 backwards-compatible behaviour prevails when this isn't defined.
1078 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
1081 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
1082 This will generate a random key of the appropriate length based on the
1083 cipher context. The EVP_CIPHER can provide its own random key generation
1084 routine to support keys of a specific form. This is used in the des and
1085 3des routines to generate a key of the correct parity. Update S/MIME
1086 code to use new functions and hence generate correct parity DES keys.
1087 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
1088 valid (weak or incorrect parity).
1091 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
1092 as looking them up. This is useful when the verified structure may contain
1093 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
1094 present unless the new PKCS7_NO_CRL flag is asserted.
1097 *) Extend ASN1 oid configuration module. It now additionally accepts the
1100 shortName = some long name, 1.2.3.4
1103 *) Reimplemented the BN_CTX implementation. There is now no more static
1104 limitation on the number of variables it can handle nor the depth of the
1105 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
1106 information can now expand as required, and rather than having a single
1107 static array of bignums, BN_CTX now uses a linked-list of such arrays
1108 allowing it to expand on demand whilst maintaining the usefulness of
1109 BN_CTX's "bundling".
1112 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
1113 to allow all RSA operations to function using a single BN_CTX.
1116 *) Preliminary support for certificate policy evaluation and checking. This
1117 is initially intended to pass the tests outlined in "Conformance Testing
1118 of Relying Party Client Certificate Path Processing Logic" v1.07.
1121 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
1122 remained unused and not that useful. A variety of other little bignum
1123 tweaks and fixes have also been made continuing on from the audit (see
1127 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
1128 associated ASN1, EVP and SSL functions and old ASN1 macros.
1131 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
1132 and this should never fail. So the return value from the use of
1133 BN_set_word() (which can fail due to needless expansion) is now deprecated;
1134 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
1137 *) BN_CTX_get() should return zero-valued bignums, providing the same
1138 initialised value as BN_new().
1139 [Geoff Thorpe, suggested by Ulf Möller]
1141 *) Support for inhibitAnyPolicy certificate extension.
1144 *) An audit of the BIGNUM code is underway, for which debugging code is
1145 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
1146 is considered valid when processing BIGNUMs, and causes execution to
1147 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
1148 further steps are taken to deliberately pollute unused data in BIGNUM
1149 structures to try and expose faulty code further on. For now, openssl will
1150 (in its default mode of operation) continue to tolerate the inconsistent
1151 forms that it has tolerated in the past, but authors and packagers should
1152 consider trying openssl and their own applications when compiled with
1153 these debugging symbols defined. It will help highlight potential bugs in
1154 their own code, and will improve the test coverage for OpenSSL itself. At
1155 some point, these tighter rules will become openssl's default to improve
1156 maintainability, though the assert()s and other overheads will remain only
1157 in debugging configurations. See bn.h for more details.
1158 [Geoff Thorpe, Nils Larsch, Ulf Möller]
1160 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
1161 that can only be obtained through BN_CTX_new() (which implicitly
1162 initialises it). The presence of this function only made it possible
1163 to overwrite an existing structure (and cause memory leaks).
1166 *) Because of the callback-based approach for implementing LHASH as a
1167 template type, lh_insert() adds opaque objects to hash-tables and
1168 lh_doall() or lh_doall_arg() are typically used with a destructor callback
1169 to clean up those corresponding objects before destroying the hash table
1170 (and losing the object pointers). So some over-zealous constifications in
1171 LHASH have been relaxed so that lh_insert() does not take (nor store) the
1172 objects as "const" and the lh_doall[_arg] callback wrappers are not
1173 prototyped to have "const" restrictions on the object pointers they are
1174 given (and so aren't required to cast them away any more).
1177 *) The tmdiff.h API was so ugly and minimal that our own timing utility
1178 (speed) prefers to use its own implementation. The two implementations
1179 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
1180 its object type properly exposed (MS_TM) instead of casting to/from "char
1181 *". This may still change yet if someone realises MS_TM and "ms_time_***"
1182 aren't necessarily the greatest nomenclatures - but this is what was used
1183 internally to the implementation so I've used that for now.
1186 *) Ensure that deprecated functions do not get compiled when
1187 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
1188 the self-tests were still using deprecated key-generation functions so
1189 these have been updated also.
1192 *) Reorganise PKCS#7 code to separate the digest location functionality
1193 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
1194 New function PKCS7_set_digest() to set the digest type for PKCS#7
1195 digestedData type. Add additional code to correctly generate the
1196 digestedData type and add support for this type in PKCS7 initialization
1200 *) New function PKCS7_set0_type_other() this initializes a PKCS7
1201 structure of type "other".
1204 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
1205 sure the loop does correctly stop and breaking ("division by zero")
1206 modulus operations are not performed. The (pre-generated) prime
1207 table crypto/bn/bn_prime.h was already correct, but it could not be
1208 re-generated on some platforms because of the "division by zero"
1209 situation in the script.
1210 [Ralf S. Engelschall]
1212 *) Update support for ECC-based TLS ciphersuites according to
1213 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
1214 SHA-1 now is only used for "small" curves (where the
1215 representation of a field element takes up to 24 bytes); for
1216 larger curves, the field element resulting from ECDH is directly
1217 used as premaster secret.
1218 [Douglas Stebila (Sun Microsystems Laboratories)]
1220 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
1221 curve secp160r1 to the tests.
1222 [Douglas Stebila (Sun Microsystems Laboratories)]
1224 *) Add the possibility to load symbols globally with DSO.
1225 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
1227 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
1228 control of the error stack.
1231 *) Add support for STORE in ENGINE.
1234 *) Add the STORE type. The intention is to provide a common interface
1235 to certificate and key stores, be they simple file-based stores, or
1236 HSM-type store, or LDAP stores, or...
1237 NOTE: The code is currently UNTESTED and isn't really used anywhere.
1240 *) Add a generic structure called OPENSSL_ITEM. This can be used to
1241 pass a list of arguments to any function as well as provide a way
1242 for a function to pass data back to the caller.
1245 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
1246 works like BUF_strdup() but can be used to duplicate a portion of
1247 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
1251 *) Add the function sk_find_ex() which works like sk_find(), but will
1252 return an index to an element even if an exact match couldn't be
1253 found. The index is guaranteed to point at the element where the
1254 searched-for key would be inserted to preserve sorting order.
1257 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
1258 takes an extra flags argument for optional functionality. Currently,
1259 the following flags are defined:
1261 OBJ_BSEARCH_VALUE_ON_NOMATCH
1262 This one gets OBJ_bsearch_ex() to return a pointer to the first
1263 element where the comparing function returns a negative or zero
1266 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
1267 This one gets OBJ_bsearch_ex() to return a pointer to the first
1268 element where the comparing function returns zero. This is useful
1269 if there are more than one element where the comparing function
1273 *) Make it possible to create self-signed certificates with 'openssl ca'
1274 in such a way that the self-signed certificate becomes part of the
1275 CA database and uses the same mechanisms for serial number generation
1276 as all other certificate signing. The new flag '-selfsign' enables
1277 this functionality. Adapt CA.sh and CA.pl.in.
1280 *) Add functionality to check the public key of a certificate request
1281 against a given private. This is useful to check that a certificate
1282 request can be signed by that key (self-signing).
1285 *) Make it possible to have multiple active certificates with the same
1286 subject in the CA index file. This is done only if the keyword
1287 'unique_subject' is set to 'no' in the main CA section (default
1288 if 'CA_default') of the configuration file. The value is saved
1289 with the database itself in a separate index attribute file,
1290 named like the index file with '.attr' appended to the name.
1293 *) Generate muti valued AVAs using '+' notation in config files for
1297 *) Support for nameConstraints certificate extension.
1300 *) Support for policyConstraints certificate extension.
1303 *) Support for policyMappings certificate extension.
1306 *) Make sure the default DSA_METHOD implementation only uses its
1307 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
1308 and change its own handlers to be NULL so as to remove unnecessary
1309 indirection. This lets alternative implementations fallback to the
1310 default implementation more easily.
1313 *) Support for directoryName in GeneralName related extensions
1317 *) Make it possible to link applications using Makefile.shared.
1318 Make that possible even when linking against static libraries!
1321 *) Support for single pass processing for S/MIME signing. This now
1322 means that S/MIME signing can be done from a pipe, in addition
1323 cleartext signing (multipart/signed type) is effectively streaming
1324 and the signed data does not need to be all held in memory.
1326 This is done with a new flag PKCS7_STREAM. When this flag is set
1327 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
1328 is done after the data is output (and digests calculated) in
1329 SMIME_write_PKCS7().
1332 *) Add full support for -rpath/-R, both in shared libraries and
1333 applications, at least on the platforms where it's known how
1337 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
1338 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
1339 will now compute a table of multiples of the generator that
1340 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
1341 faster (notably in the case of a single point multiplication,
1342 scalar * generator).
1343 [Nils Larsch, Bodo Moeller]
1345 *) IPv6 support for certificate extensions. The various extensions
1346 which use the IP:a.b.c.d can now take IPv6 addresses using the
1347 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
1351 *) Added an ENGINE that implements RSA by performing private key
1352 exponentiations with the GMP library. The conversions to and from
1353 GMP's mpz_t format aren't optimised nor are any montgomery forms
1354 cached, and on x86 it appears OpenSSL's own performance has caught up.
1355 However there are likely to be other architectures where GMP could
1356 provide a boost. This ENGINE is not built in by default, but it can be
1357 specified at Configure time and should be accompanied by the necessary
1358 linker additions, eg;
1359 ./config -DOPENSSL_USE_GMP -lgmp
1362 *) "openssl engine" will not display ENGINE/DSO load failure errors when
1363 testing availability of engines with "-t" - the old behaviour is
1364 produced by increasing the feature's verbosity with "-tt".
1367 *) ECDSA routines: under certain error conditions uninitialized BN objects
1368 could be freed. Solution: make sure initialization is performed early
1369 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
1373 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
1374 and DH_METHOD (eg. by ENGINE implementations) to override the normal
1375 software implementations. For DSA and DH, parameter generation can
1376 also be overriden by providing the appropriate method callbacks.
1379 *) Change the "progress" mechanism used in key-generation and
1380 primality testing to functions that take a new BN_GENCB pointer in
1381 place of callback/argument pairs. The new API functions have "_ex"
1382 postfixes and the older functions are reimplemented as wrappers for
1383 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
1384 declarations of the old functions to help (graceful) attempts to
1385 migrate to the new functions. Also, the new key-generation API
1386 functions operate on a caller-supplied key-structure and return
1387 success/failure rather than returning a key or NULL - this is to
1388 help make "keygen" another member function of RSA_METHOD etc.
1390 Example for using the new callback interface:
1392 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
1396 BN_GENCB_set(&my_cb, my_callback, my_arg);
1398 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
1399 /* For the meaning of a, b in calls to my_callback(), see the
1400 * documentation of the function that calls the callback.
1401 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
1402 * my_callback should return 1 if it wants BN_is_prime_ex()
1403 * to continue, or 0 to stop.
1408 *) Change the ZLIB compression method to be stateful, and make it
1409 available to TLS with the number defined in
1410 draft-ietf-tls-compression-04.txt.
1413 *) Add the ASN.1 structures and functions for CertificatePair, which
1414 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
1416 CertificatePair ::= SEQUENCE {
1417 forward [0] Certificate OPTIONAL,
1418 reverse [1] Certificate OPTIONAL,
1419 -- at least one of the pair shall be present -- }
1421 Also implement the PEM functions to read and write certificate
1422 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
1424 This needed to be defined, mostly for the sake of the LDAP
1425 attribute crossCertificatePair, but may prove useful elsewhere as
1429 *) Make it possible to inhibit symlinking of shared libraries in
1430 Makefile.shared, for Cygwin's sake.
1433 *) Extend the BIGNUM API by creating a function
1434 void BN_set_negative(BIGNUM *a, int neg);
1435 and a macro that behave like
1436 int BN_is_negative(const BIGNUM *a);
1438 to avoid the need to access 'a->neg' directly in applications.
1441 *) Implement fast modular reduction for pseudo-Mersenne primes
1442 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
1443 EC_GROUP_new_curve_GFp() will now automatically use this
1445 [Nils Larsch <nla@trustcenter.de>]
1447 *) Add new lock type (CRYPTO_LOCK_BN).
1450 *) Change the ENGINE framework to automatically load engines
1451 dynamically from specific directories unless they could be
1452 found to already be built in or loaded. Move all the
1453 current engines except for the cryptodev one to a new
1455 The engines in engines/ are built as shared libraries if
1456 the "shared" options was given to ./Configure or ./config.
1457 Otherwise, they are inserted in libcrypto.a.
1458 /usr/local/ssl/engines is the default directory for dynamic
1459 engines, but that can be overriden at configure time through
1460 the usual use of --prefix and/or --openssldir, and at run
1461 time with the environment variable OPENSSL_ENGINES.
1462 [Geoff Thorpe and Richard Levitte]
1464 *) Add Makefile.shared, a helper makefile to build shared
1465 libraries. Addapt Makefile.org.
1468 *) Add version info to Win32 DLLs.
1469 [Peter 'Luna' Runestig" <peter@runestig.com>]
1471 *) Add new 'medium level' PKCS#12 API. Certificates and keys
1472 can be added using this API to created arbitrary PKCS#12
1473 files while avoiding the low level API.
1475 New options to PKCS12_create(), key or cert can be NULL and
1476 will then be omitted from the output file. The encryption
1477 algorithm NIDs can be set to -1 for no encryption, the mac
1478 iteration count can be set to 0 to omit the mac.
1480 Enhance pkcs12 utility by making the -nokeys and -nocerts
1481 options work when creating a PKCS#12 file. New option -nomac
1482 to omit the mac, NONE can be set for an encryption algorithm.
1483 New code is modified to use the enhanced PKCS12_create()
1484 instead of the low level API.
1487 *) Extend ASN1 encoder to support indefinite length constructed
1488 encoding. This can output sequences tags and octet strings in
1489 this form. Modify pk7_asn1.c to support indefinite length
1490 encoding. This is experimental and needs additional code to
1491 be useful, such as an ASN1 bio and some enhanced streaming
1494 Extend template encode functionality so that tagging is passed
1495 down to the template encoder.
1498 *) Let 'openssl req' fail if an argument to '-newkey' is not
1499 recognized instead of using RSA as a default.
1502 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
1503 As these are not official, they are not included in "ALL";
1504 the "ECCdraft" ciphersuite group alias can be used to select them.
1505 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
1507 *) Add ECDH engine support.
1508 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
1510 *) Add ECDH in new directory crypto/ecdh/.
1511 [Douglas Stebila (Sun Microsystems Laboratories)]
1513 *) Let BN_rand_range() abort with an error after 100 iterations
1514 without success (which indicates a broken PRNG).
1517 *) Change BN_mod_sqrt() so that it verifies that the input value
1518 is really the square of the return value. (Previously,
1519 BN_mod_sqrt would show GIGO behaviour.)
1522 *) Add named elliptic curves over binary fields from X9.62, SECG,
1523 and WAP/WTLS; add OIDs that were still missing.
1525 [Sheueling Chang Shantz and Douglas Stebila
1526 (Sun Microsystems Laboratories)]
1528 *) Extend the EC library for elliptic curves over binary fields
1529 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
1532 EC_GF2m_simple_method
1536 EC_GROUP_new_curve_GF2m
1537 EC_GROUP_set_curve_GF2m
1538 EC_GROUP_get_curve_GF2m
1539 EC_POINT_set_affine_coordinates_GF2m
1540 EC_POINT_get_affine_coordinates_GF2m
1541 EC_POINT_set_compressed_coordinates_GF2m
1543 Point compression for binary fields is disabled by default for
1544 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
1547 As binary polynomials are represented as BIGNUMs, various members
1548 of the EC_GROUP and EC_POINT data structures can be shared
1549 between the implementations for prime fields and binary fields;
1550 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
1551 are essentially identical to their ..._GFp counterparts.
1552 (For simplicity, the '..._GFp' prefix has been dropped from
1553 various internal method names.)
1555 An internal 'field_div' method (similar to 'field_mul' and
1556 'field_sqr') has been added; this is used only for binary fields.
1558 [Sheueling Chang Shantz and Douglas Stebila
1559 (Sun Microsystems Laboratories)]
1561 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
1562 through methods ('mul', 'precompute_mult').
1564 The generic implementations (now internally called 'ec_wNAF_mul'
1565 and 'ec_wNAF_precomputed_mult') remain the default if these
1566 methods are undefined.
1568 [Sheueling Chang Shantz and Douglas Stebila
1569 (Sun Microsystems Laboratories)]
1571 *) New function EC_GROUP_get_degree, which is defined through
1572 EC_METHOD. For curves over prime fields, this returns the bit
1573 length of the modulus.
1575 [Sheueling Chang Shantz and Douglas Stebila
1576 (Sun Microsystems Laboratories)]
1578 *) New functions EC_GROUP_dup, EC_POINT_dup.
1579 (These simply call ..._new and ..._copy).
1581 [Sheueling Chang Shantz and Douglas Stebila
1582 (Sun Microsystems Laboratories)]
1584 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
1585 Polynomials are represented as BIGNUMs (where the sign bit is not
1586 used) in the following functions [macros]:
1589 BN_GF2m_sub [= BN_GF2m_add]
1590 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
1591 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
1592 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
1594 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
1595 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
1596 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
1597 BN_GF2m_cmp [= BN_ucmp]
1599 (Note that only the 'mod' functions are actually for fields GF(2^m).
1600 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
1602 For some functions, an the irreducible polynomial defining a
1603 field can be given as an 'unsigned int[]' with strictly
1604 decreasing elements giving the indices of those bits that are set;
1605 i.e., p[] represents the polynomial
1606 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
1608 p[0] > p[1] > ... > p[k] = 0.
1609 This applies to the following functions:
1614 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
1615 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
1617 BN_GF2m_mod_sqrt_arr
1618 BN_GF2m_mod_solve_quad_arr
1622 Conversion can be performed by the following functions:
1627 bntest.c has additional tests for binary polynomial arithmetic.
1629 Two implementations for BN_GF2m_mod_div() are available.
1630 The default algorithm simply uses BN_GF2m_mod_inv() and
1631 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
1632 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
1633 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
1635 [Sheueling Chang Shantz and Douglas Stebila
1636 (Sun Microsystems Laboratories)]
1638 *) Add new error code 'ERR_R_DISABLED' that can be used when some
1639 functionality is disabled at compile-time.
1640 [Douglas Stebila <douglas.stebila@sun.com>]
1642 *) Change default behaviour of 'openssl asn1parse' so that more
1643 information is visible when viewing, e.g., a certificate:
1645 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
1646 mode the content of non-printable OCTET STRINGs is output in a
1647 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
1648 avoid the appearance of a printable string.
1649 [Nils Larsch <nla@trustcenter.de>]
1651 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
1653 EC_GROUP_set_asn1_flag()
1654 EC_GROUP_get_asn1_flag()
1655 EC_GROUP_set_point_conversion_form()
1656 EC_GROUP_get_point_conversion_form()
1657 These control ASN1 encoding details:
1658 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
1659 has been set to OPENSSL_EC_NAMED_CURVE.
1660 - Points are encoded in uncompressed form by default; options for
1661 asn1_for are as for point2oct, namely
1662 POINT_CONVERSION_COMPRESSED
1663 POINT_CONVERSION_UNCOMPRESSED
1664 POINT_CONVERSION_HYBRID
1666 Also add 'seed' and 'seed_len' members to EC_GROUP with access
1669 EC_GROUP_get0_seed()
1670 EC_GROUP_get_seed_len()
1671 This is used only for ASN1 purposes (so far).
1672 [Nils Larsch <nla@trustcenter.de>]
1674 *) Add 'field_type' member to EC_METHOD, which holds the NID
1675 of the appropriate field type OID. The new function
1676 EC_METHOD_get_field_type() returns this value.
1677 [Nils Larsch <nla@trustcenter.de>]
1682 EC_POINT_point2hex()
1683 EC_POINT_hex2point()
1684 providing useful interfaces to EC_POINT_point2oct() and
1685 EC_POINT_oct2point().
1686 [Nils Larsch <nla@trustcenter.de>]
1688 *) Change internals of the EC library so that the functions
1689 EC_GROUP_set_generator()
1690 EC_GROUP_get_generator()
1691 EC_GROUP_get_order()
1692 EC_GROUP_get_cofactor()
1693 are implemented directly in crypto/ec/ec_lib.c and not dispatched
1694 to methods, which would lead to unnecessary code duplication when
1695 adding different types of curves.
1696 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
1698 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
1699 arithmetic, and such that modified wNAFs are generated
1700 (which avoid length expansion in many cases).
1703 *) Add a function EC_GROUP_check_discriminant() (defined via
1704 EC_METHOD) that verifies that the curve discriminant is non-zero.
1706 Add a function EC_GROUP_check() that makes some sanity tests
1707 on a EC_GROUP, its generator and order. This includes
1708 EC_GROUP_check_discriminant().
1709 [Nils Larsch <nla@trustcenter.de>]
1711 *) Add ECDSA in new directory crypto/ecdsa/.
1713 Add applications 'openssl ecparam' and 'openssl ecdsa'
1714 (these are based on 'openssl dsaparam' and 'openssl dsa').
1716 ECDSA support is also included in various other files across the
1717 library. Most notably,
1718 - 'openssl req' now has a '-newkey ecdsa:file' option;
1719 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
1720 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
1721 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
1722 them suitable for ECDSA where domain parameters must be
1723 extracted before the specific public key;
1724 - ECDSA engine support has been added.
1725 [Nils Larsch <nla@trustcenter.de>]
1727 *) Include some named elliptic curves, and add OIDs from X9.62,
1728 SECG, and WAP/WTLS. Each curve can be obtained from the new
1730 EC_GROUP_new_by_curve_name(),
1731 and the list of available named curves can be obtained with
1732 EC_get_builtin_curves().
1733 Also add a 'curve_name' member to EC_GROUP objects, which can be
1735 EC_GROUP_set_curve_name()
1736 EC_GROUP_get_curve_name()
1737 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
1739 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1740 was actually never needed) and in BN_mul(). The removal in BN_mul()
1741 required a small change in bn_mul_part_recursive() and the addition
1742 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1743 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1744 bn_sub_words() and bn_add_words() except they take arrays with
1748 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
1750 *) Cleanse PEM buffers before freeing them since they may contain
1752 [Benjamin Bennett <ben@psc.edu>]
1754 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1755 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1756 authentication-only ciphersuites.
1759 *) Since AES128 and AES256 share a single mask bit in the logic of
1760 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1761 kludge to work properly if AES128 is available and AES256 isn't.
1764 *) Expand security boundary to match 1.1.1 module.
1767 *) Remove redundant features: hash file source, editing of test vectors
1768 modify fipsld to use external fips_premain.c signature.
1771 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
1772 run algorithm test programs.
1775 *) Make algorithm test programs more tolerant of whitespace.
1778 *) Have SSL/TLS server implementation tolerate "mismatched" record
1779 protocol version while receiving ClientHello even if the
1780 ClientHello is fragmented. (The server can't insist on the
1781 particular protocol version it has chosen before the ServerHello
1782 message has informed the client about his choice.)
1785 *) Load error codes if they are not already present instead of using a
1786 static variable. This allows them to be cleanly unloaded and reloaded.
1789 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
1791 *) Introduce limits to prevent malicious keys being able to
1792 cause a denial of service. (CVE-2006-2940)
1793 [Steve Henson, Bodo Moeller]
1795 *) Fix ASN.1 parsing of certain invalid structures that can result
1796 in a denial of service. (CVE-2006-2937) [Steve Henson]
1798 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1799 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1801 *) Fix SSL client code which could crash if connecting to a
1802 malicious SSLv2 server. (CVE-2006-4343)
1803 [Tavis Ormandy and Will Drewry, Google Security Team]
1805 *) Change ciphersuite string processing so that an explicit
1806 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
1807 will no longer include "AES128-SHA"), and any other similar
1808 ciphersuite (same bitmap) from *other* protocol versions (so that
1809 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
1810 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
1811 changes from 0.9.8b and 0.9.8d.
1814 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
1816 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1817 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1819 *) Change the Unix randomness entropy gathering to use poll() when
1820 possible instead of select(), since the latter has some
1821 undesirable limitations.
1822 [Darryl Miles via Richard Levitte and Bodo Moeller]
1824 *) Disable rogue ciphersuites:
1826 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1827 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1828 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1830 The latter two were purportedly from
1831 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1834 Also deactive the remaining ciphersuites from
1835 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1836 unofficial, and the ID has long expired.
1839 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1840 dual-core machines) and other potential thread-safety issues.
1843 Changes between 0.9.7i and 0.9.7j [04 May 2006]
1845 *) Adapt fipsld and the build system to link against the validated FIPS
1846 module in FIPS mode.
1849 *) Fixes for VC++ 2005 build under Windows.
1852 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
1853 from a Windows bash shell such as MSYS. It is autodetected from the
1854 "config" script when run from a VC++ environment. Modify standard VC++
1855 build to use fipscanister.o from the GNU make build.
1858 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
1860 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
1861 The value now differs depending on if you build for FIPS or not.
1862 BEWARE! A program linked with a shared FIPSed libcrypto can't be
1863 safely run with a non-FIPSed libcrypto, as it may crash because of
1864 the difference induced by this change.
1867 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
1869 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1870 (part of SSL_OP_ALL). This option used to disable the
1871 countermeasure against man-in-the-middle protocol-version
1872 rollback in the SSL 2.0 server implementation, which is a bad
1873 idea. (CVE-2005-2969)
1875 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1876 for Information Security, National Institute of Advanced Industrial
1877 Science and Technology [AIST], Japan)]
1879 *) Minimal support for X9.31 signatures and PSS padding modes. This is
1880 mainly for FIPS compliance and not fully integrated at this stage.
1883 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
1884 the exponentiation using a fixed-length exponent. (Otherwise,
1885 the information leaked through timing could expose the secret key
1886 after many signatures; cf. Bleichenbacher's attack on DSA with
1890 *) Make a new fixed-window mod_exp implementation the default for
1891 RSA, DSA, and DH private-key operations so that the sequence of
1892 squares and multiplies and the memory access pattern are
1893 independent of the particular secret key. This will mitigate
1894 cache-timing and potential related attacks.
1896 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
1897 and this is automatically used by BN_mod_exp_mont() if the new flag
1898 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
1899 will use this BN flag for private exponents unless the flag
1900 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
1901 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
1903 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
1905 *) Change the client implementation for SSLv23_method() and
1906 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
1907 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
1908 (Previously, the SSL 2.0 backwards compatible Client Hello
1909 message format would be used even with SSL_OP_NO_SSLv2.)
1912 *) Add support for smime-type MIME parameter in S/MIME messages which some
1916 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
1917 a threadsafe manner. Modify rsa code to use new function and add calls
1918 to dsa and dh code (which had race conditions before).
1921 *) Include the fixed error library code in the C error file definitions
1922 instead of fixing them up at runtime. This keeps the error code
1923 structures constant.
1926 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
1928 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
1931 *) Fixes for newer kerberos headers. NB: the casts are needed because
1932 the 'length' field is signed on one version and unsigned on another
1933 with no (?) obvious way to tell the difference, without these VC++
1934 complains. Also the "definition" of FAR (blank) is no longer included
1935 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
1936 some needed definitions.
1939 *) Undo Cygwin change.
1942 *) Added support for proxy certificates according to RFC 3820.
1943 Because they may be a security thread to unaware applications,
1944 they must be explicitely allowed in run-time. See
1945 docs/HOWTO/proxy_certificates.txt for further information.
1948 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
1950 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
1951 server and client random values. Previously
1952 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
1953 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
1955 This change has negligible security impact because:
1957 1. Server and client random values still have 24 bytes of pseudo random
1960 2. Server and client random values are sent in the clear in the initial
1963 3. The master secret is derived using the premaster secret (48 bytes in
1964 size for static RSA ciphersuites) as well as client server and random
1967 The OpenSSL team would like to thank the UK NISCC for bringing this issue
1970 [Stephen Henson, reported by UK NISCC]
1972 *) Use Windows randomness collection on Cygwin.
1975 *) Fix hang in EGD/PRNGD query when communication socket is closed
1976 prematurely by EGD/PRNGD.
1977 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
1979 *) Prompt for pass phrases when appropriate for PKCS12 input format.
1982 *) Back-port of selected performance improvements from development
1983 branch, as well as improved support for PowerPC platforms.
1986 *) Add lots of checks for memory allocation failure, error codes to indicate
1987 failure and freeing up memory if a failure occurs.
1988 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
1990 *) Add new -passin argument to dgst.
1993 *) Perform some character comparisons of different types in X509_NAME_cmp:
1994 this is needed for some certificates that reencode DNs into UTF8Strings
1995 (in violation of RFC3280) and can't or wont issue name rollover
1999 *) Make an explicit check during certificate validation to see that
2000 the CA setting in each certificate on the chain is correct. As a
2001 side effect always do the following basic checks on extensions,
2002 not just when there's an associated purpose to the check:
2004 - if there is an unhandled critical extension (unless the user
2005 has chosen to ignore this fault)
2006 - if the path length has been exceeded (if one is set at all)
2007 - that certain extensions fit the associated purpose (if one has
2011 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
2013 *) Avoid a race condition when CRLs are checked in a multi threaded
2014 environment. This would happen due to the reordering of the revoked
2015 entries during signature checking and serial number lookup. Now the
2016 encoding is cached and the serial number sort performed under a lock.
2017 Add new STACK function sk_is_sorted().
2020 *) Add Delta CRL to the extension code.
2023 *) Various fixes to s3_pkt.c so alerts are sent properly.
2024 [David Holmes <d.holmes@f5.com>]
2026 *) Reduce the chances of duplicate issuer name and serial numbers (in
2027 violation of RFC3280) using the OpenSSL certificate creation utilities.
2028 This is done by creating a random 64 bit value for the initial serial
2029 number when a serial number file is created or when a self signed
2030 certificate is created using 'openssl req -x509'. The initial serial
2031 number file is created using 'openssl x509 -next_serial' in CA.pl
2032 rather than being initialized to 1.
2035 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
2037 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
2038 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
2039 [Joe Orton, Steve Henson]
2041 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
2043 [Joe Orton, Steve Henson]
2045 *) Make it possible to have multiple active certificates with the same
2046 subject in the CA index file. This is done only if the keyword
2047 'unique_subject' is set to 'no' in the main CA section (default
2048 if 'CA_default') of the configuration file. The value is saved
2049 with the database itself in a separate index attribute file,
2050 named like the index file with '.attr' appended to the name.
2053 *) X509 verify fixes. Disable broken certificate workarounds when
2054 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
2055 keyUsage extension present. Don't accept CRLs with unhandled critical
2056 extensions: since verify currently doesn't process CRL extensions this
2057 rejects a CRL with *any* critical extensions. Add new verify error codes
2061 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
2062 A clarification of RFC2560 will require the use of OCTET STRINGs and
2063 some implementations cannot handle the current raw format. Since OpenSSL
2064 copies and compares OCSP nonces as opaque blobs without any attempt at
2065 parsing them this should not create any compatibility issues.
2068 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
2069 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
2070 this HMAC (and other) operations are several times slower than OpenSSL
2074 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
2075 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2077 *) Use the correct content when signing type "other".
2080 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
2082 *) Fix various bugs revealed by running the NISCC test suite:
2084 Stop out of bounds reads in the ASN1 code when presented with
2085 invalid tags (CVE-2003-0543 and CVE-2003-0544).
2087 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
2089 If verify callback ignores invalid public key errors don't try to check
2090 certificate signature with the NULL public key.
2094 *) New -ignore_err option in ocsp application to stop the server
2095 exiting on the first error in a request.
2098 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
2099 if the server requested one: as stated in TLS 1.0 and SSL 3.0
2103 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
2104 extra data after the compression methods not only for TLS 1.0
2105 but also for SSL 3.0 (as required by the specification).
2106 [Bodo Moeller; problem pointed out by Matthias Loepfe]
2108 *) Change X509_certificate_type() to mark the key as exported/exportable
2109 when it's 512 *bits* long, not 512 bytes.
2112 *) Change AES_cbc_encrypt() so it outputs exact multiple of
2113 blocks during encryption.
2116 *) Various fixes to base64 BIO and non blocking I/O. On write
2117 flushes were not handled properly if the BIO retried. On read
2118 data was not being buffered properly and had various logic bugs.
2119 This also affects blocking I/O when the data being decoded is a
2123 *) Various S/MIME bugfixes and compatibility changes:
2124 output correct application/pkcs7 MIME type if
2125 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
2126 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
2127 of files as .eml work). Correctly handle very long lines in MIME
2131 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
2133 *) Countermeasure against the Klima-Pokorny-Rosa extension of
2134 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
2135 a protocol version number mismatch like a decryption error
2136 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
2139 *) Turn on RSA blinding by default in the default implementation
2140 to avoid a timing attack. Applications that don't want it can call
2141 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
2142 They would be ill-advised to do so in most cases.
2143 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
2145 *) Change RSA blinding code so that it works when the PRNG is not
2146 seeded (in this case, the secret RSA exponent is abused as
2147 an unpredictable seed -- if it is not unpredictable, there
2148 is no point in blinding anyway). Make RSA blinding thread-safe
2149 by remembering the creator's thread ID in rsa->blinding and
2150 having all other threads use local one-time blinding factors
2151 (this requires more computation than sharing rsa->blinding, but
2152 avoids excessive locking; and if an RSA object is not shared
2153 between threads, blinding will still be very fast).
2156 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
2157 ENGINE as defaults for all supported algorithms irrespective of
2158 the 'flags' parameter. 'flags' is now honoured, so applications
2159 should make sure they are passing it correctly.
2162 *) Target "mingw" now allows native Windows code to be generated in
2163 the Cygwin environment as well as with the MinGW compiler.
2166 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
2168 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
2169 via timing by performing a MAC computation even if incorrrect
2170 block cipher padding has been found. This is a countermeasure
2171 against active attacks where the attacker has to distinguish
2172 between bad padding and a MAC verification error. (CVE-2003-0078)
2174 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
2175 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
2176 Martin Vuagnoux (EPFL, Ilion)]
2178 *) Make the no-err option work as intended. The intention with no-err
2179 is not to have the whole error stack handling routines removed from
2180 libcrypto, it's only intended to remove all the function name and
2181 reason texts, thereby removing some of the footprint that may not
2182 be interesting if those errors aren't displayed anyway.
2184 NOTE: it's still possible for any application or module to have it's
2185 own set of error texts inserted. The routines are there, just not
2186 used by default when no-err is given.
2189 *) Add support for FreeBSD on IA64.
2190 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
2192 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
2193 Kerberos function mit_des_cbc_cksum(). Before this change,
2194 the value returned by DES_cbc_cksum() was like the one from
2195 mit_des_cbc_cksum(), except the bytes were swapped.
2196 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
2198 *) Allow an application to disable the automatic SSL chain building.
2199 Before this a rather primitive chain build was always performed in
2200 ssl3_output_cert_chain(): an application had no way to send the
2201 correct chain if the automatic operation produced an incorrect result.
2203 Now the chain builder is disabled if either:
2205 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
2207 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
2209 The reasoning behind this is that an application would not want the
2210 auto chain building to take place if extra chain certificates are
2211 present and it might also want a means of sending no additional
2212 certificates (for example the chain has two certificates and the
2216 *) Add the possibility to build without the ENGINE framework.
2217 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2219 *) Under Win32 gmtime() can return NULL: check return value in
2220 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
2223 *) DSA routines: under certain error conditions uninitialized BN objects
2224 could be freed. Solution: make sure initialization is performed early
2225 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
2226 Nils Larsch <nla@trustcenter.de> via PR#459)
2229 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
2230 checked on reconnect on the client side, therefore session resumption
2231 could still fail with a "ssl session id is different" error. This
2232 behaviour is masked when SSL_OP_ALL is used due to
2233 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
2234 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2235 followup to PR #377.
2238 *) IA-32 assembler support enhancements: unified ELF targets, support
2239 for SCO/Caldera platforms, fix for Cygwin shared build.
2242 *) Add support for FreeBSD on sparc64. As a consequence, support for
2243 FreeBSD on non-x86 processors is separate from x86 processors on
2244 the config script, much like the NetBSD support.
2245 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
2247 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
2249 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
2252 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
2253 code (06) was taken as the first octet of the session ID and the last
2254 octet was ignored consequently. As a result SSLv2 client side session
2255 caching could not have worked due to the session ID mismatch between
2257 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2261 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
2262 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
2266 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
2267 seems that in spite of existing for more than a year, many application
2268 author have done nothing to provide the necessary callbacks, which
2269 means that this particular engine will not work properly anywhere.
2270 This is a very unfortunate situation which forces us, in the name
2271 of usability, to give the hw_ncipher.c a static lock, which is part
2273 NOTE: This is for the 0.9.7 series ONLY. This hack will never
2274 appear in 0.9.8 or later. We EXPECT application authors to have
2275 dealt properly with this when 0.9.8 is released (unless we actually
2276 make such changes in the libcrypto locking code that changes will
2277 have to be made anyway).
2280 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
2281 octets have been read, EOF or an error occurs. Without this change
2282 some truncated ASN1 structures will not produce an error.
2285 *) Disable Heimdal support, since it hasn't been fully implemented.
2286 Still give the possibility to force the use of Heimdal, but with
2287 warnings and a request that patches get sent to openssl-dev.
2290 *) Add the VC-CE target, introduce the WINCE sysname, and add
2291 INSTALL.WCE and appropriate conditionals to make it build.
2292 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2294 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
2295 cygssl-x.y.z.dll, where x, y and z are the major, minor and
2296 edit numbers of the version.
2297 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2299 *) Introduce safe string copy and catenation functions
2300 (BUF_strlcpy() and BUF_strlcat()).
2301 [Ben Laurie (CHATS) and Richard Levitte]
2303 *) Avoid using fixed-size buffers for one-line DNs.
2304 [Ben Laurie (CHATS)]
2306 *) Add BUF_MEM_grow_clean() to avoid information leakage when
2307 resizing buffers containing secrets, and use where appropriate.
2308 [Ben Laurie (CHATS)]
2310 *) Avoid using fixed size buffers for configuration file location.
2311 [Ben Laurie (CHATS)]
2313 *) Avoid filename truncation for various CA files.
2314 [Ben Laurie (CHATS)]
2316 *) Use sizeof in preference to magic numbers.
2317 [Ben Laurie (CHATS)]
2319 *) Avoid filename truncation in cert requests.
2320 [Ben Laurie (CHATS)]
2322 *) Add assertions to check for (supposedly impossible) buffer
2324 [Ben Laurie (CHATS)]
2326 *) Don't cache truncated DNS entries in the local cache (this could
2327 potentially lead to a spoofing attack).
2328 [Ben Laurie (CHATS)]
2330 *) Fix various buffers to be large enough for hex/decimal
2331 representations in a platform independent manner.
2332 [Ben Laurie (CHATS)]
2334 *) Add CRYPTO_realloc_clean() to avoid information leakage when
2335 resizing buffers containing secrets, and use where appropriate.
2336 [Ben Laurie (CHATS)]
2338 *) Add BIO_indent() to avoid much slightly worrying code to do
2340 [Ben Laurie (CHATS)]
2342 *) Convert sprintf()/BIO_puts() to BIO_printf().
2343 [Ben Laurie (CHATS)]
2345 *) buffer_gets() could terminate with the buffer only half
2347 [Ben Laurie (CHATS)]
2349 *) Add assertions to prevent user-supplied crypto functions from
2350 overflowing internal buffers by having large block sizes, etc.
2351 [Ben Laurie (CHATS)]
2353 *) New OPENSSL_assert() macro (similar to assert(), but enabled
2355 [Ben Laurie (CHATS)]
2357 *) Eliminate unused copy of key in RC4.
2358 [Ben Laurie (CHATS)]
2360 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
2361 [Ben Laurie (CHATS)]
2363 *) Fix off-by-one error in EGD path.
2364 [Ben Laurie (CHATS)]
2366 *) If RANDFILE path is too long, ignore instead of truncating.
2367 [Ben Laurie (CHATS)]
2369 *) Eliminate unused and incorrectly sized X.509 structure
2371 [Ben Laurie (CHATS)]
2373 *) Eliminate unused and dangerous function knumber().
2374 [Ben Laurie (CHATS)]
2376 *) Eliminate unused and dangerous structure, KSSL_ERR.
2377 [Ben Laurie (CHATS)]
2379 *) Protect against overlong session ID context length in an encoded
2380 session object. Since these are local, this does not appear to be
2382 [Ben Laurie (CHATS)]
2384 *) Change from security patch (see 0.9.6e below) that did not affect
2385 the 0.9.6 release series:
2387 Remote buffer overflow in SSL3 protocol - an attacker could
2388 supply an oversized master key in Kerberos-enabled versions.
2390 [Ben Laurie (CHATS)]
2392 *) Change the SSL kerb5 codes to match RFC 2712.
2395 *) Make -nameopt work fully for req and add -reqopt switch.
2396 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
2398 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
2399 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
2401 *) Make sure tests can be performed even if the corresponding algorithms
2402 have been removed entirely. This was also the last step to make
2403 OpenSSL compilable with DJGPP under all reasonable conditions.
2404 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
2406 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
2407 to allow version independent disabling of normally unselected ciphers,
2408 which may be activated as a side-effect of selecting a single cipher.
2410 (E.g., cipher list string "RSA" enables ciphersuites that are left
2411 out of "ALL" because they do not provide symmetric encryption.
2412 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
2413 [Lutz Jaenicke, Bodo Moeller]
2415 *) Add appropriate support for separate platform-dependent build
2416 directories. The recommended way to make a platform-dependent
2417 build directory is the following (tested on Linux), maybe with
2420 # Place yourself outside of the OpenSSL source tree. In
2421 # this example, the environment variable OPENSSL_SOURCE
2422 # is assumed to contain the absolute OpenSSL source directory.
2423 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
2424 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
2425 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
2426 mkdir -p `dirname $F`
2427 ln -s $OPENSSL_SOURCE/$F $F
2430 To be absolutely sure not to disturb the source tree, a "make clean"
2431 is a good thing. If it isn't successfull, don't worry about it,
2432 it probably means the source directory is very clean.
2435 *) Make sure any ENGINE control commands make local copies of string
2436 pointers passed to them whenever necessary. Otherwise it is possible
2437 the caller may have overwritten (or deallocated) the original string
2438 data when a later ENGINE operation tries to use the stored values.
2439 [Götz Babin-Ebell <babinebell@trustcenter.de>]
2441 *) Improve diagnostics in file reading and command-line digests.
2442 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
2444 *) Add AES modes CFB and OFB to the object database. Correct an
2445 error in AES-CFB decryption.
2448 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
2449 allows existing EVP_CIPHER_CTX structures to be reused after
2450 calling EVP_*Final(). This behaviour is used by encryption
2451 BIOs and some applications. This has the side effect that
2452 applications must explicitly clean up cipher contexts with
2453 EVP_CIPHER_CTX_cleanup() or they will leak memory.
2456 *) Check the values of dna and dnb in bn_mul_recursive before calling
2457 bn_mul_comba (a non zero value means the a or b arrays do not contain
2458 n2 elements) and fallback to bn_mul_normal if either is not zero.
2461 *) Fix escaping of non-ASCII characters when using the -subj option
2462 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
2465 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
2466 form for "surname", serialNumber has no short form.
2467 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
2468 therefore remove "mail" short name for "internet 7".
2469 The OID for unique identifiers in X509 certificates is
2470 x500UniqueIdentifier, not uniqueIdentifier.
2471 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
2474 *) Add an "init" command to the ENGINE config module and auto initialize
2475 ENGINEs. Without any "init" command the ENGINE will be initialized
2476 after all ctrl commands have been executed on it. If init=1 the
2477 ENGINE is initailized at that point (ctrls before that point are run
2478 on the uninitialized ENGINE and after on the initialized one). If
2479 init=0 then the ENGINE will not be iniatialized at all.
2482 *) Fix the 'app_verify_callback' interface so that the user-defined
2483 argument is actually passed to the callback: In the
2484 SSL_CTX_set_cert_verify_callback() prototype, the callback
2485 declaration has been changed from
2488 int (*cb)(X509_STORE_CTX *,void *);
2489 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
2490 i=s->ctx->app_verify_callback(&ctx)
2491 has been changed into
2492 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
2494 To update applications using SSL_CTX_set_cert_verify_callback(),
2495 a dummy argument can be added to their callback functions.
2496 [D. K. Smetters <smetters@parc.xerox.com>]
2498 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
2499 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
2501 *) Add and OPENSSL_LOAD_CONF define which will cause
2502 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
2503 This allows older applications to transparently support certain
2504 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
2505 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
2506 load the config file and OPENSSL_add_all_algorithms_conf() which will
2507 always load it have also been added.
2510 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
2511 Adjust NIDs and EVP layer.
2512 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2514 *) Config modules support in openssl utility.
2516 Most commands now load modules from the config file,
2517 though in a few (such as version) this isn't done
2518 because it couldn't be used for anything.
2520 In the case of ca and req the config file used is
2521 the same as the utility itself: that is the -config
2522 command line option can be used to specify an
2526 *) Move default behaviour from OPENSSL_config(). If appname is NULL
2527 use "openssl_conf" if filename is NULL use default openssl config file.
2530 *) Add an argument to OPENSSL_config() to allow the use of an alternative
2531 config section name. Add a new flag to tolerate a missing config file
2532 and move code to CONF_modules_load_file().
2535 *) Support for crypto accelerator cards from Accelerated Encryption
2536 Processing, www.aep.ie. (Use engine 'aep')
2537 The support was copied from 0.9.6c [engine] and adapted/corrected
2538 to work with the new engine framework.
2539 [AEP Inc. and Richard Levitte]
2541 *) Support for SureWare crypto accelerator cards from Baltimore
2542 Technologies. (Use engine 'sureware')
2543 The support was copied from 0.9.6c [engine] and adapted
2544 to work with the new engine framework.
2547 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
2548 make the newer ENGINE framework commands for the CHIL engine work.
2549 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
2551 *) Make it possible to produce shared libraries on ReliantUNIX.
2552 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
2554 *) Add the configuration target debug-linux-ppro.
2555 Make 'openssl rsa' use the general key loading routines
2556 implemented in apps.c, and make those routines able to
2557 handle the key format FORMAT_NETSCAPE and the variant
2559 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2561 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2562 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2564 *) Add -keyform to rsautl, and document -engine.
2565 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
2567 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
2568 BIO_R_NO_SUCH_FILE error code rather than the generic
2569 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
2572 *) Add new functions
2574 ERR_peek_last_error_line
2575 ERR_peek_last_error_line_data.
2576 These are similar to
2579 ERR_peek_error_line_data,
2580 but report on the latest error recorded rather than the first one
2581 still in the error queue.
2582 [Ben Laurie, Bodo Moeller]
2584 *) default_algorithms option in ENGINE config module. This allows things
2586 default_algorithms = ALL
2587 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
2590 *) Prelminary ENGINE config module.
2593 *) New experimental application configuration code.
2596 *) Change the AES code to follow the same name structure as all other
2597 symmetric ciphers, and behave the same way. Move everything to
2598 the directory crypto/aes, thereby obsoleting crypto/rijndael.
2599 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2601 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
2602 [Ben Laurie and Theo de Raadt]
2604 *) Add option to output public keys in req command.
2605 [Massimiliano Pala madwolf@openca.org]
2607 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
2608 (up to about 10% better than before for P-192 and P-224).
2611 *) New functions/macros
2613 SSL_CTX_set_msg_callback(ctx, cb)
2614 SSL_CTX_set_msg_callback_arg(ctx, arg)
2615 SSL_set_msg_callback(ssl, cb)
2616 SSL_set_msg_callback_arg(ssl, arg)
2618 to request calling a callback function
2620 void cb(int write_p, int version, int content_type,
2621 const void *buf, size_t len, SSL *ssl, void *arg)
2623 whenever a protocol message has been completely received
2624 (write_p == 0) or sent (write_p == 1). Here 'version' is the
2625 protocol version according to which the SSL library interprets
2626 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
2627 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
2628 the content type as defined in the SSL 3.0/TLS 1.0 protocol
2629 specification (change_cipher_spec(20), alert(21), handshake(22)).
2630 'buf' and 'len' point to the actual message, 'ssl' to the
2631 SSL object, and 'arg' is the application-defined value set by
2632 SSL[_CTX]_set_msg_callback_arg().
2634 'openssl s_client' and 'openssl s_server' have new '-msg' options
2635 to enable a callback that displays all protocol messages.
2638 *) Change the shared library support so shared libraries are built as
2639 soon as the corresponding static library is finished, and thereby get
2640 openssl and the test programs linked against the shared library.
2641 This still only happens when the keyword "shard" has been given to
2642 the configuration scripts.
2644 NOTE: shared library support is still an experimental thing, and
2645 backward binary compatibility is still not guaranteed.
2646 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
2648 *) Add support for Subject Information Access extension.
2649 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2651 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
2652 additional bytes when new memory had to be allocated, not just
2653 when reusing an existing buffer.
2656 *) New command line and configuration option 'utf8' for the req command.
2657 This allows field values to be specified as UTF8 strings.
2660 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
2661 runs for the former and machine-readable output for the latter.
2664 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
2665 of the e-mail address in the DN (i.e., it will go into a certificate
2666 extension only). The new configuration file option 'email_in_dn = no'
2667 has the same effect.
2668 [Massimiliano Pala madwolf@openca.org]
2670 *) Change all functions with names starting with des_ to be starting
2671 with DES_ instead. Add wrappers that are compatible with libdes,
2672 but are named _ossl_old_des_*. Finally, add macros that map the
2673 des_* symbols to the corresponding _ossl_old_des_* if libdes
2674 compatibility is desired. If OpenSSL 0.9.6c compatibility is
2675 desired, the des_* symbols will be mapped to DES_*, with one
2678 Since we provide two compatibility mappings, the user needs to
2679 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
2680 compatibility is desired. The default (i.e., when that macro
2681 isn't defined) is OpenSSL 0.9.6c compatibility.
2683 There are also macros that enable and disable the support of old
2684 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
2685 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
2686 are defined, the default will apply: to support the old des routines.
2688 In either case, one must include openssl/des.h to get the correct
2689 definitions. Do not try to just include openssl/des_old.h, that
2692 NOTE: This is a major break of an old API into a new one. Software
2693 authors are encouraged to switch to the DES_ style functions. Some
2694 time in the future, des_old.h and the libdes compatibility functions
2695 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
2696 default), and then completely removed.
2699 *) Test for certificates which contain unsupported critical extensions.
2700 If such a certificate is found during a verify operation it is
2701 rejected by default: this behaviour can be overridden by either
2702 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
2703 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
2704 X509_supported_extension() has also been added which returns 1 if a
2705 particular extension is supported.
2708 *) Modify the behaviour of EVP cipher functions in similar way to digests
2709 to retain compatibility with existing code.
2712 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
2713 compatibility with existing code. In particular the 'ctx' parameter does
2714 not have to be to be initialized before the call to EVP_DigestInit() and
2715 it is tidied up after a call to EVP_DigestFinal(). New function
2716 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
2717 EVP_MD_CTX_copy() changed to not require the destination to be
2718 initialized valid and new function EVP_MD_CTX_copy_ex() added which
2719 requires the destination to be valid.
2721 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
2722 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
2725 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
2726 so that complete 'Handshake' protocol structures are kept in memory
2727 instead of overwriting 'msg_type' and 'length' with 'body' data.
2730 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
2731 [Massimo Santin via Richard Levitte]
2733 *) Major restructuring to the underlying ENGINE code. This includes
2734 reduction of linker bloat, separation of pure "ENGINE" manipulation
2735 (initialisation, etc) from functionality dealing with implementations
2736 of specific crypto iterfaces. This change also introduces integrated
2737 support for symmetric ciphers and digest implementations - so ENGINEs
2738 can now accelerate these by providing EVP_CIPHER and EVP_MD
2739 implementations of their own. This is detailed in crypto/engine/README
2740 as it couldn't be adequately described here. However, there are a few
2741 API changes worth noting - some RSA, DSA, DH, and RAND functions that
2742 were changed in the original introduction of ENGINE code have now
2743 reverted back - the hooking from this code to ENGINE is now a good
2744 deal more passive and at run-time, operations deal directly with
2745 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
2746 dereferencing through an ENGINE pointer any more. Also, the ENGINE
2747 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
2748 they were not being used by the framework as there is no concept of a
2749 BIGNUM_METHOD and they could not be generalised to the new
2750 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
2751 ENGINE_cpy() has been removed as it cannot be consistently defined in
2755 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
2758 *) Change mkdef.pl to sort symbols that get the same entry number,
2759 and make sure the automatically generated functions ERR_load_*
2760 become part of libeay.num as well.
2763 *) New function SSL_renegotiate_pending(). This returns true once
2764 renegotiation has been requested (either SSL_renegotiate() call
2765 or HelloRequest/ClientHello receveived from the peer) and becomes
2766 false once a handshake has been completed.
2767 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
2768 sends a HelloRequest, but does not ensure that a handshake takes
2769 place. SSL_renegotiate_pending() is useful for checking if the
2770 client has followed the request.)
2773 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
2774 By default, clients may request session resumption even during
2775 renegotiation (if session ID contexts permit); with this option,
2776 session resumption is possible only in the first handshake.
2778 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
2779 more bits available for options that should not be part of
2780 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
2783 *) Add some demos for certificate and certificate request creation.
2786 *) Make maximum certificate chain size accepted from the peer application
2787 settable (SSL*_get/set_max_cert_list()), as proposed by
2788 "Douglas E. Engert" <deengert@anl.gov>.
2791 *) Add support for shared libraries for Unixware-7
2792 (Boyd Lynn Gerber <gerberb@zenez.com>).
2795 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
2796 be done prior to destruction. Use this to unload error strings from
2797 ENGINEs that load their own error strings. NB: This adds two new API
2798 functions to "get" and "set" this destroy handler in an ENGINE.
2801 *) Alter all existing ENGINE implementations (except "openssl" and
2802 "openbsd") to dynamically instantiate their own error strings. This
2803 makes them more flexible to be built both as statically-linked ENGINEs
2804 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
2805 Also, add stub code to each that makes building them as self-contained
2806 shared-libraries easier (see README.ENGINE).
2809 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
2810 implementations into applications that are completely implemented in
2811 self-contained shared-libraries. The "dynamic" ENGINE exposes control
2812 commands that can be used to configure what shared-library to load and
2813 to control aspects of the way it is handled. Also, made an update to
2814 the README.ENGINE file that brings its information up-to-date and
2815 provides some information and instructions on the "dynamic" ENGINE
2816 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
2819 *) Make it possible to unload ranges of ERR strings with a new
2820 "ERR_unload_strings" function.
2823 *) Add a copy() function to EVP_MD.
2826 *) Make EVP_MD routines take a context pointer instead of just the
2827 md_data void pointer.
2830 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
2831 that the digest can only process a single chunk of data
2832 (typically because it is provided by a piece of
2833 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
2834 is only going to provide a single chunk of data, and hence the
2835 framework needn't accumulate the data for oneshot drivers.
2838 *) As with "ERR", make it possible to replace the underlying "ex_data"
2839 functions. This change also alters the storage and management of global
2840 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
2841 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
2842 index counters. The API functions that use this state have been changed
2843 to take a "class_index" rather than pointers to the class's local STACK
2844 and counter, and there is now an API function to dynamically create new
2845 classes. This centralisation allows us to (a) plug a lot of the
2846 thread-safety problems that existed, and (b) makes it possible to clean
2847 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
2848 such data would previously have always leaked in application code and
2849 workarounds were in place to make the memory debugging turn a blind eye
2850 to it. Application code that doesn't use this new function will still
2851 leak as before, but their memory debugging output will announce it now
2852 rather than letting it slide.
2854 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
2855 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
2856 has a return value to indicate success or failure.
2859 *) Make it possible to replace the underlying "ERR" functions such that the
2860 global state (2 LHASH tables and 2 locks) is only used by the "default"
2861 implementation. This change also adds two functions to "get" and "set"
2862 the implementation prior to it being automatically set the first time
2863 any other ERR function takes place. Ie. an application can call "get",
2864 pass the return value to a module it has just loaded, and that module
2865 can call its own "set" function using that value. This means the
2866 module's "ERR" operations will use (and modify) the error state in the
2867 application and not in its own statically linked copy of OpenSSL code.
2870 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
2871 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
2872 the operation, and provides a more encapsulated way for external code
2873 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
2874 to use these functions rather than manually incrementing the counts.
2876 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
2879 *) Add EVP test program.
2882 *) Add symmetric cipher support to ENGINE. Expect the API to change!
2885 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
2886 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
2887 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
2888 These allow a CRL to be built without having to access X509_CRL fields
2889 directly. Modify 'ca' application to use new functions.
2892 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
2893 bug workarounds. Rollback attack detection is a security feature.
2894 The problem will only arise on OpenSSL servers when TLSv1 is not
2895 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
2896 Software authors not wanting to support TLSv1 will have special reasons
2897 for their choice and can explicitly enable this option.
2898 [Bodo Moeller, Lutz Jaenicke]
2900 *) Rationalise EVP so it can be extended: don't include a union of
2901 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
2902 (similar to those existing for EVP_CIPHER_CTX).
2907 EVP_MD_CTX_init(&md); /* new function call */
2908 EVP_DigestInit(&md, EVP_sha1());
2909 EVP_DigestUpdate(&md, in, len);
2910 EVP_DigestFinal(&md, out, NULL);
2911 EVP_MD_CTX_cleanup(&md); /* new function call */
2915 *) Make DES key schedule conform to the usual scheme, as well as
2916 correcting its structure. This means that calls to DES functions
2917 now have to pass a pointer to a des_key_schedule instead of a
2918 plain des_key_schedule (which was actually always a pointer
2921 des_key_schedule ks;
2923 des_set_key_checked(..., &ks);
2924 des_ncbc_encrypt(..., &ks, ...);
2926 (Note that a later change renames 'des_...' into 'DES_...'.)
2929 *) Initial reduction of linker bloat: the use of some functions, such as
2930 PEM causes large amounts of unused functions to be linked in due to
2931 poor organisation. For example pem_all.c contains every PEM function
2932 which has a knock on effect of linking in large amounts of (unused)
2933 ASN1 code. Grouping together similar functions and splitting unrelated
2934 functions prevents this.
2937 *) Cleanup of EVP macros.
2940 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
2941 correct _ecb suffix.
2944 *) Add initial OCSP responder support to ocsp application. The
2945 revocation information is handled using the text based index
2946 use by the ca application. The responder can either handle
2947 requests generated internally, supplied in files (for example
2948 via a CGI script) or using an internal minimal server.
2951 *) Add configuration choices to get zlib compression for TLS.
2954 *) Changes to Kerberos SSL for RFC 2712 compliance:
2955 1. Implemented real KerberosWrapper, instead of just using
2956 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
2957 2. Implemented optional authenticator field of KerberosWrapper.
2959 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
2960 and authenticator structs; see crypto/krb5/.
2962 Generalized Kerberos calls to support multiple Kerberos libraries.
2963 [Vern Staats <staatsvr@asc.hpc.mil>,
2964 Jeffrey Altman <jaltman@columbia.edu>
2965 via Richard Levitte]
2967 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
2968 already does with RSA. testdsa.h now has 'priv_key/pub_key'
2969 values for each of the key sizes rather than having just
2970 parameters (and 'speed' generating keys each time).
2973 *) Speed up EVP routines.
2976 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
2977 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
2978 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
2979 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
2981 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
2982 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
2983 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
2986 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
2988 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
2991 *) Added the OS2-EMX target.
2992 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
2994 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
2995 to support NCONF routines in extension code. New function CONF_set_nconf()
2996 to allow functions which take an NCONF to also handle the old LHASH
2997 structure: this means that the old CONF compatible routines can be
2998 retained (in particular wrt extensions) without having to duplicate the
2999 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
3002 *) Enhance the general user interface with mechanisms for inner control
3003 and with possibilities to have yes/no kind of prompts.
3006 *) Change all calls to low level digest routines in the library and
3007 applications to use EVP. Add missing calls to HMAC_cleanup() and
3008 don't assume HMAC_CTX can be copied using memcpy().
3009 [Verdon Walker <VWalker@novell.com>, Steve Henson]
3011 *) Add the possibility to control engines through control names but with
3012 arbitrary arguments instead of just a string.
3013 Change the key loaders to take a UI_METHOD instead of a callback
3014 function pointer. NOTE: this breaks binary compatibility with earlier
3015 versions of OpenSSL [engine].
3016 Adapt the nCipher code for these new conditions and add a card insertion
3020 *) Enhance the general user interface with mechanisms to better support
3021 dialog box interfaces, application-defined prompts, the possibility
3022 to use defaults (for example default passwords from somewhere else)
3023 and interrupts/cancellations.
3026 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
3027 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
3030 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
3031 tidy up some unnecessarily weird code in 'sk_new()').
3032 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
3034 *) Change the key loading routines for ENGINEs to use the same kind
3035 callback (pem_password_cb) as all other routines that need this
3039 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
3040 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
3041 than this minimum value is recommended.
3044 *) New random seeder for OpenVMS, using the system process statistics
3045 that are easily reachable.
3048 *) Windows apparently can't transparently handle global
3049 variables defined in DLLs. Initialisations such as:
3051 const ASN1_ITEM *it = &ASN1_INTEGER_it;
3053 wont compile. This is used by the any applications that need to
3054 declare their own ASN1 modules. This was fixed by adding the option
3055 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
3056 needed for static libraries under Win32.
3059 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
3060 setting of purpose and trust fields. New X509_STORE trust and
3061 purpose functions and tidy up setting in other SSL functions.
3064 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
3065 structure. These are inherited by X509_STORE_CTX when it is
3066 initialised. This allows various defaults to be set in the
3067 X509_STORE structure (such as flags for CRL checking and custom
3068 purpose or trust settings) for functions which only use X509_STORE_CTX
3069 internally such as S/MIME.
3071 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
3072 trust settings if they are not set in X509_STORE. This allows X509_STORE
3073 purposes and trust (in S/MIME for example) to override any set by default.
3075 Add command line options for CRL checking to smime, s_client and s_server
3079 *) Initial CRL based revocation checking. If the CRL checking flag(s)
3080 are set then the CRL is looked up in the X509_STORE structure and
3081 its validity and signature checked, then if the certificate is found
3082 in the CRL the verify fails with a revoked error.
3084 Various new CRL related callbacks added to X509_STORE_CTX structure.
3086 Command line options added to 'verify' application to support this.
3088 This needs some additional work, such as being able to handle multiple
3089 CRLs with different times, extension based lookup (rather than just
3090 by subject name) and ultimately more complete V2 CRL extension
3094 *) Add a general user interface API (crypto/ui/). This is designed
3095 to replace things like des_read_password and friends (backward
3096 compatibility functions using this new API are provided).
3097 The purpose is to remove prompting functions from the DES code
3098 section as well as provide for prompting through dialog boxes in
3099 a window system and the like.
3102 *) Add "ex_data" support to ENGINE so implementations can add state at a
3103 per-structure level rather than having to store it globally.
3106 *) Make it possible for ENGINE structures to be copied when retrieved by
3107 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
3108 This causes the "original" ENGINE structure to act like a template,
3109 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
3110 operational state can be localised to each ENGINE structure, despite the
3111 fact they all share the same "methods". New ENGINE structures returned in
3112 this case have no functional references and the return value is the single
3113 structural reference. This matches the single structural reference returned
3114 by ENGINE_by_id() normally, when it is incremented on the pre-existing
3118 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
3119 needs to match any other type at all we need to manually clear the
3123 *) Changes to the "openssl engine" utility to include;
3124 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
3125 about an ENGINE's available control commands.
3126 - executing control commands from command line arguments using the
3127 '-pre' and '-post' switches. '-post' is only used if '-t' is
3128 specified and the ENGINE is successfully initialised. The syntax for
3129 the individual commands are colon-separated, for example;
3130 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
3133 *) New dynamic control command support for ENGINEs. ENGINEs can now
3134 declare their own commands (numbers), names (strings), descriptions,
3135 and input types for run-time discovery by calling applications. A
3136 subset of these commands are implicitly classed as "executable"
3137 depending on their input type, and only these can be invoked through
3138 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
3139 can be based on user input, config files, etc). The distinction is
3140 that "executable" commands cannot return anything other than a boolean
3141 result and can only support numeric or string input, whereas some
3142 discoverable commands may only be for direct use through
3143 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
3144 pointers, or other custom uses. The "executable" commands are to
3145 support parameterisations of ENGINE behaviour that can be
3146 unambiguously defined by ENGINEs and used consistently across any
3147 OpenSSL-based application. Commands have been added to all the
3148 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
3149 control over shared-library paths without source code alterations.
3152 *) Changed all ENGINE implementations to dynamically allocate their
3153 ENGINEs rather than declaring them statically. Apart from this being
3154 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
3155 this also allows the implementations to compile without using the
3156 internal engine_int.h header.
3159 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
3160 'const' value. Any code that should be able to modify a RAND_METHOD
3161 should already have non-const pointers to it (ie. they should only
3162 modify their own ones).
3165 *) Made a variety of little tweaks to the ENGINE code.
3166 - "atalla" and "ubsec" string definitions were moved from header files
3167 to C code. "nuron" string definitions were placed in variables
3168 rather than hard-coded - allowing parameterisation of these values
3169 later on via ctrl() commands.
3170 - Removed unused "#if 0"'d code.
3171 - Fixed engine list iteration code so it uses ENGINE_free() to release
3172 structural references.
3173 - Constified the RAND_METHOD element of ENGINE structures.
3174 - Constified various get/set functions as appropriate and added
3175 missing functions (including a catch-all ENGINE_cpy that duplicates
3176 all ENGINE values onto a new ENGINE except reference counts/state).
3177 - Removed NULL parameter checks in get/set functions. Setting a method
3178 or function to NULL is a way of cancelling out a previously set
3179 value. Passing a NULL ENGINE parameter is just plain stupid anyway
3180 and doesn't justify the extra error symbols and code.
3181 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
3182 flags from engine_int.h to engine.h.
3183 - Changed prototypes for ENGINE handler functions (init(), finish(),
3184 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
3187 *) Implement binary inversion algorithm for BN_mod_inverse in addition
3188 to the algorithm using long division. The binary algorithm can be
3189 used only if the modulus is odd. On 32-bit systems, it is faster
3190 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
3191 roughly 5-15% for 256-bit moduli), so we use it only for moduli
3192 up to 450 bits. In 64-bit environments, the binary algorithm
3193 appears to be advantageous for much longer moduli; here we use it
3194 for moduli up to 2048 bits.
3197 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
3198 could not support the combine flag in choice fields.
3201 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
3202 extensions from a certificate request to the certificate.
3205 *) Allow multiple 'certopt' and 'nameopt' options to be separated
3206 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
3207 file: this allows the display of the certificate about to be
3208 signed to be customised, to allow certain fields to be included
3209 or excluded and extension details. The old system didn't display
3210 multicharacter strings properly, omitted fields not in the policy
3211 and couldn't display additional details such as extensions.
3214 *) Function EC_POINTs_mul for multiple scalar multiplication
3215 of an arbitrary number of elliptic curve points
3216 \sum scalars[i]*points[i],
3217 optionally including the generator defined for the EC_GROUP:
3218 scalar*generator + \sum scalars[i]*points[i].
3220 EC_POINT_mul is a simple wrapper function for the typical case
3221 that the point list has just one item (besides the optional
3225 *) First EC_METHODs for curves over GF(p):
3227 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
3228 operations and provides various method functions that can also
3229 operate with faster implementations of modular arithmetic.
3231 EC_GFp_mont_method() reuses most functions that are part of
3232 EC_GFp_simple_method, but uses Montgomery arithmetic.
3234 [Bodo Moeller; point addition and point doubling
3235 implementation directly derived from source code provided by
3236 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
3238 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
3239 crypto/ec/ec_lib.c):
3241 Curves are EC_GROUP objects (with an optional group generator)
3242 based on EC_METHODs that are built into the library.
3244 Points are EC_POINT objects based on EC_GROUP objects.
3246 Most of the framework would be able to handle curves over arbitrary
3247 finite fields, but as there are no obvious types for fields other
3248 than GF(p), some functions are limited to that for now.
3251 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
3252 that the file contains a complete HTTP response.
3255 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
3256 change the def and num file printf format specifier from "%-40sXXX"
3257 to "%-39s XXX". The latter will always guarantee a space after the
3258 field while the former will cause them to run together if the field
3259 is 40 of more characters long.
3262 *) Constify the cipher and digest 'method' functions and structures
3263 and modify related functions to take constant EVP_MD and EVP_CIPHER
3267 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
3268 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
3271 *) Modify EVP_Digest*() routines so they now return values. Although the
3272 internal software routines can never fail additional hardware versions
3276 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
3278 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
3279 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
3282 ERR_R_NESTED_ASN1_ERROR
3284 ERR_R_MISSING_ASN1_EOS
3285 were 4 .. 9, conflicting with
3286 ERR_LIB_RSA (= ERR_R_RSA_LIB)
3288 ERR_LIB_PEM (= ERR_R_PEM_LIB).
3289 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
3291 Add new error code 'ERR_R_INTERNAL_ERROR'.
3294 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
3298 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
3299 sets the subject name for a new request or supersedes the
3300 subject name in a given request. Formats that can be parsed are
3301 'CN=Some Name, OU=myOU, C=IT'
3303 'CN=Some Name/OU=myOU/C=IT'.
3305 Add options '-batch' and '-verbose' to 'openssl req'.
3306 [Massimiliano Pala <madwolf@hackmasters.net>]
3308 *) Introduce the possibility to access global variables through
3309 functions on platform were that's the best way to handle exporting
3310 global variables in shared libraries. To enable this functionality,
3311 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
3312 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
3313 is normally done by Configure or something similar).
3315 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
3316 in the source file (foo.c) like this:
3318 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
3319 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
3321 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
3322 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
3324 OPENSSL_DECLARE_GLOBAL(int,foo);
3325 #define foo OPENSSL_GLOBAL_REF(foo)
3326 OPENSSL_DECLARE_GLOBAL(double,bar);
3327 #define bar OPENSSL_GLOBAL_REF(bar)
3329 The #defines are very important, and therefore so is including the
3330 header file everywhere where the defined globals are used.
3332 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
3333 of ASN.1 items, but that structure is a bit different.
3335 The largest change is in util/mkdef.pl which has been enhanced with
3336 better and easier to understand logic to choose which symbols should
3337 go into the Windows .def files as well as a number of fixes and code
3338 cleanup (among others, algorithm keywords are now sorted
3339 lexicographically to avoid constant rewrites).
3342 *) In BN_div() keep a copy of the sign of 'num' before writing the
3343 result to 'rm' because if rm==num the value will be overwritten
3344 and produce the wrong result if 'num' is negative: this caused
3345 problems with BN_mod() and BN_nnmod().
3348 *) Function OCSP_request_verify(). This checks the signature on an
3349 OCSP request and verifies the signer certificate. The signer
3350 certificate is just checked for a generic purpose and OCSP request
3354 *) Add OCSP_check_validity() function to check the validity of OCSP
3355 responses. OCSP responses are prepared in real time and may only
3356 be a few seconds old. Simply checking that the current time lies
3357 between thisUpdate and nextUpdate max reject otherwise valid responses
3358 caused by either OCSP responder or client clock inaccuracy. Instead
3359 we allow thisUpdate and nextUpdate to fall within a certain period of
3360 the current time. The age of the response can also optionally be
3361 checked. Two new options -validity_period and -status_age added to
3365 *) If signature or public key algorithm is unrecognized print out its
3366 OID rather that just UNKNOWN.
3369 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
3370 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
3371 ID to be generated from the issuer certificate alone which can then be
3372 passed to OCSP_id_issuer_cmp().
3375 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
3376 ASN1 modules to export functions returning ASN1_ITEM pointers
3377 instead of the ASN1_ITEM structures themselves. This adds several
3378 new macros which allow the underlying ASN1 function/structure to
3379 be accessed transparently. As a result code should not use ASN1_ITEM
3380 references directly (such as &X509_it) but instead use the relevant
3381 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
3382 use of the new ASN1 code on platforms where exporting structures
3383 is problematical (for example in shared libraries) but exporting
3384 functions returning pointers to structures is not.
3387 *) Add support for overriding the generation of SSL/TLS session IDs.
3388 These callbacks can be registered either in an SSL_CTX or per SSL.
3389 The purpose of this is to allow applications to control, if they wish,
3390 the arbitrary values chosen for use as session IDs, particularly as it
3391 can be useful for session caching in multiple-server environments. A
3392 command-line switch for testing this (and any client code that wishes
3393 to use such a feature) has been added to "s_server".
3394 [Geoff Thorpe, Lutz Jaenicke]
3396 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
3397 of the form '#if defined(...) || defined(...) || ...' and
3398 '#if !defined(...) && !defined(...) && ...'. This also avoids
3399 the growing number of special cases it was previously handling.
3402 *) Make all configuration macros available for application by making
3403 sure they are available in opensslconf.h, by giving them names starting
3404 with "OPENSSL_" to avoid conflicts with other packages and by making
3405 sure e_os2.h will cover all platform-specific cases together with
3407 Additionally, it is now possible to define configuration/platform-
3408 specific names (called "system identities"). In the C code, these
3409 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
3410 macro with the name beginning with "OPENSSL_SYS_", which is determined
3411 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
3415 *) New option -set_serial to 'req' and 'x509' this allows the serial
3416 number to use to be specified on the command line. Previously self
3417 signed certificates were hard coded with serial number 0 and the
3418 CA options of 'x509' had to use a serial number in a file which was
3422 *) New options to 'ca' utility to support V2 CRL entry extensions.
3423 Currently CRL reason, invalidity date and hold instruction are
3424 supported. Add new CRL extensions to V3 code and some new objects.
3427 *) New function EVP_CIPHER_CTX_set_padding() this is used to
3428 disable standard block padding (aka PKCS#5 padding) in the EVP
3429 API, which was previously mandatory. This means that the data is
3430 not padded in any way and so the total length much be a multiple
3431 of the block size, otherwise an error occurs.
3434 *) Initial (incomplete) OCSP SSL support.
3437 *) New function OCSP_parse_url(). This splits up a URL into its host,
3438 port and path components: primarily to parse OCSP URLs. New -url
3439 option to ocsp utility.
3442 *) New nonce behavior. The return value of OCSP_check_nonce() now
3443 reflects the various checks performed. Applications can decide
3444 whether to tolerate certain situations such as an absent nonce
3445 in a response when one was present in a request: the ocsp application
3446 just prints out a warning. New function OCSP_add1_basic_nonce()
3447 this is to allow responders to include a nonce in a response even if
3448 the request is nonce-less.
3451 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
3452 skipped when using openssl x509 multiple times on a single input file,
3453 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
3456 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
3457 set string type: to handle setting ASN1_TIME structures. Fix ca
3458 utility to correctly initialize revocation date of CRLs.
3461 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
3462 the clients preferred ciphersuites and rather use its own preferences.
3463 Should help to work around M$ SGC (Server Gated Cryptography) bug in
3464 Internet Explorer by ensuring unchanged hash method during stepup.
3465 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
3468 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
3469 to aes and add a new 'exist' option to print out symbols that don't
3473 *) Additional options to ocsp utility to allow flags to be set and
3474 additional certificates supplied.
3477 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
3478 OCSP client a number of certificate to only verify the response
3482 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
3483 handle the new API. Currently only ECB, CBC modes supported. Add new
3486 Add TLS AES ciphersuites as described in RFC3268, "Advanced
3487 Encryption Standard (AES) Ciphersuites for Transport Layer
3488 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
3489 not enabled by default and were not part of the "ALL" ciphersuite
3490 alias because they were not yet official; they could be
3491 explicitly requested by specifying the "AESdraft" ciphersuite
3492 group alias. In the final release of OpenSSL 0.9.7, the group
3493 alias is called "AES" and is part of "ALL".)
3494 [Ben Laurie, Steve Henson, Bodo Moeller]
3496 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
3497 request to response.
3500 *) Functions for OCSP responders. OCSP_request_onereq_count(),
3501 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
3502 extract information from a certificate request. OCSP_response_create()
3503 creates a response and optionally adds a basic response structure.
3504 OCSP_basic_add1_status() adds a complete single response to a basic
3505 response and returns the OCSP_SINGLERESP structure just added (to allow
3506 extensions to be included for example). OCSP_basic_add1_cert() adds a
3507 certificate to a basic response and OCSP_basic_sign() signs a basic
3508 response with various flags. New helper functions ASN1_TIME_check()
3509 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
3510 (converts ASN1_TIME to GeneralizedTime).
3513 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
3514 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
3515 structure from a certificate. X509_pubkey_digest() digests the public_key
3516 contents: this is used in various key identifiers.
3519 *) Make sk_sort() tolerate a NULL argument.
3520 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
3522 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
3523 passed by the function are trusted implicitly. If any of them signed the
3524 response then it is assumed to be valid and is not verified.
3527 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
3528 to data. This was previously part of the PKCS7 ASN1 code. This
3529 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
3530 [Steve Henson, reported by Kenneth R. Robinette
3531 <support@securenetterm.com>]
3533 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
3534 routines: without these tracing memory leaks is very painful.
3535 Fix leaks in PKCS12 and PKCS7 routines.
3538 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
3539 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
3540 effectively meant GeneralizedTime would never be used. Now it
3541 is initialised to -1 but X509_time_adj() now has to check the value
3542 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
3543 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
3544 [Steve Henson, reported by Kenneth R. Robinette
3545 <support@securenetterm.com>]
3547 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
3548 result in a zero length in the ASN1_INTEGER structure which was
3549 not consistent with the structure when d2i_ASN1_INTEGER() was used
3550 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
3551 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
3552 where it did not print out a minus for negative ASN1_INTEGER.
3555 *) Add summary printout to ocsp utility. The various functions which
3556 convert status values to strings have been renamed to:
3557 OCSP_response_status_str(), OCSP_cert_status_str() and
3558 OCSP_crl_reason_str() and are no longer static. New options
3559 to verify nonce values and to disable verification. OCSP response
3560 printout format cleaned up.
3563 *) Add additional OCSP certificate checks. These are those specified
3564 in RFC2560. This consists of two separate checks: the CA of the
3565 certificate being checked must either be the OCSP signer certificate
3566 or the issuer of the OCSP signer certificate. In the latter case the
3567 OCSP signer certificate must contain the OCSP signing extended key
3568 usage. This check is performed by attempting to match the OCSP
3569 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
3570 in the OCSP_CERTID structures of the response.
3573 *) Initial OCSP certificate verification added to OCSP_basic_verify()
3574 and related routines. This uses the standard OpenSSL certificate
3575 verify routines to perform initial checks (just CA validity) and
3576 to obtain the certificate chain. Then additional checks will be
3577 performed on the chain. Currently the root CA is checked to see
3578 if it is explicitly trusted for OCSP signing. This is used to set
3579 a root CA as a global signing root: that is any certificate that
3580 chains to that CA is an acceptable OCSP signing certificate.
3583 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
3584 extensions from a separate configuration file.
3585 As when reading extensions from the main configuration file,
3586 the '-extensions ...' option may be used for specifying the
3588 [Massimiliano Pala <madwolf@comune.modena.it>]
3590 *) New OCSP utility. Allows OCSP requests to be generated or
3591 read. The request can be sent to a responder and the output
3592 parsed, outputed or printed in text form. Not complete yet:
3593 still needs to check the OCSP response validity.
3596 *) New subcommands for 'openssl ca':
3597 'openssl ca -status <serial>' prints the status of the cert with
3598 the given serial number (according to the index file).
3599 'openssl ca -updatedb' updates the expiry status of certificates
3601 [Massimiliano Pala <madwolf@comune.modena.it>]
3603 *) New '-newreq-nodes' command option to CA.pl. This is like
3604 '-newreq', but calls 'openssl req' with the '-nodes' option
3605 so that the resulting key is not encrypted.
3606 [Damien Miller <djm@mindrot.org>]
3608 *) New configuration for the GNU Hurd.
3609 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
3611 *) Initial code to implement OCSP basic response verify. This
3612 is currently incomplete. Currently just finds the signer's
3613 certificate and verifies the signature on the response.
3616 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
3617 value of OPENSSLDIR. This is available via the new '-d' option
3618 to 'openssl version', and is also included in 'openssl version -a'.
3621 *) Allowing defining memory allocation callbacks that will be given
3622 file name and line number information in additional arguments
3623 (a const char* and an int). The basic functionality remains, as
3624 well as the original possibility to just replace malloc(),
3625 realloc() and free() by functions that do not know about these
3626 additional arguments. To register and find out the current
3627 settings for extended allocation functions, the following
3628 functions are provided:
3630 CRYPTO_set_mem_ex_functions
3631 CRYPTO_set_locked_mem_ex_functions
3632 CRYPTO_get_mem_ex_functions
3633 CRYPTO_get_locked_mem_ex_functions
3635 These work the same way as CRYPTO_set_mem_functions and friends.
3636 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
3637 extended allocation function is enabled.
3638 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
3639 a conventional allocation function is enabled.
3640 [Richard Levitte, Bodo Moeller]
3642 *) Finish off removing the remaining LHASH function pointer casts.
3643 There should no longer be any prototype-casting required when using
3644 the LHASH abstraction, and any casts that remain are "bugs". See
3645 the callback types and macros at the head of lhash.h for details
3646 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
3649 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
3650 If /dev/[u]random devices are not available or do not return enough
3651 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
3653 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
3654 /etc/entropy will be queried once each in this sequence, quering stops
3655 when enough entropy was collected without querying more sockets.
3658 *) Change the Unix RAND_poll() variant to be able to poll several
3659 random devices, as specified by DEVRANDOM, until a sufficient amount
3660 of data has been collected. We spend at most 10 ms on each file
3661 (select timeout) and read in non-blocking mode. DEVRANDOM now
3662 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
3663 (previously it was just the string "/dev/urandom"), so on typical
3664 platforms the 10 ms delay will never occur.
3665 Also separate out the Unix variant to its own file, rand_unix.c.
3666 For VMS, there's a currently-empty rand_vms.c.
3669 *) Move OCSP client related routines to ocsp_cl.c. These
3670 provide utility functions which an application needing
3671 to issue a request to an OCSP responder and analyse the
3672 response will typically need: as opposed to those which an
3673 OCSP responder itself would need which will be added later.
3675 OCSP_request_sign() signs an OCSP request with an API similar
3676 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
3677 response. OCSP_response_get1_basic() extracts basic response
3678 from response. OCSP_resp_find_status(): finds and extracts status
3679 information from an OCSP_CERTID structure (which will be created
3680 when the request structure is built). These are built from lower
3681 level functions which work on OCSP_SINGLERESP structures but
3682 wont normally be used unless the application wishes to examine
3683 extensions in the OCSP response for example.
3685 Replace nonce routines with a pair of functions.
3686 OCSP_request_add1_nonce() adds a nonce value and optionally
3687 generates a random value. OCSP_check_nonce() checks the
3688 validity of the nonce in an OCSP response.
3691 *) Change function OCSP_request_add() to OCSP_request_add0_id().
3692 This doesn't copy the supplied OCSP_CERTID and avoids the
3693 need to free up the newly created id. Change return type
3694 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
3695 This can then be used to add extensions to the request.
3696 Deleted OCSP_request_new(), since most of its functionality
3697 is now in OCSP_REQUEST_new() (and the case insensitive name
3698 clash) apart from the ability to set the request name which
3699 will be added elsewhere.
3702 *) Update OCSP API. Remove obsolete extensions argument from
3703 various functions. Extensions are now handled using the new
3704 OCSP extension code. New simple OCSP HTTP function which
3705 can be used to send requests and parse the response.
3708 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
3709 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
3710 uses the special reorder version of SET OF to sort the attributes
3711 and reorder them to match the encoded order. This resolves a long
3712 standing problem: a verify on a PKCS7 structure just after signing
3713 it used to fail because the attribute order did not match the
3714 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
3715 it uses the received order. This is necessary to tolerate some broken
3716 software that does not order SET OF. This is handled by encoding
3717 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
3718 to produce the required SET OF.
3721 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
3722 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
3723 files to get correct declarations of the ASN.1 item variables.
3726 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
3727 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
3728 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
3729 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
3730 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
3731 ASN1_ITEM and no wrapper functions.
3734 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
3735 replace the old function pointer based I/O routines. Change most of
3736 the *_d2i_bio() and *_d2i_fp() functions to use these.
3739 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
3740 lines, recognice more "algorithms" that can be deselected, and make
3741 it complain about algorithm deselection that isn't recognised.
3744 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
3745 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
3746 to use new functions. Add NO_ASN1_OLD which can be set to remove
3747 some old style ASN1 functions: this can be used to determine if old
3748 code will still work when these eventually go away.
3751 *) New extension functions for OCSP structures, these follow the
3752 same conventions as certificates and CRLs.
3755 *) New function X509V3_add1_i2d(). This automatically encodes and
3756 adds an extension. Its behaviour can be customised with various
3757 flags to append, replace or delete. Various wrappers added for
3758 certifcates and CRLs.
3761 *) Fix to avoid calling the underlying ASN1 print routine when
3762 an extension cannot be parsed. Correct a typo in the
3763 OCSP_SERVICELOC extension. Tidy up print OCSP format.
3766 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
3767 entries for variables.
3770 *) Add functionality to apps/openssl.c for detecting locking
3771 problems: As the program is single-threaded, all we have
3772 to do is register a locking callback using an array for
3773 storing which locks are currently held by the program.
3776 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
3777 SSL_get_ex_data_X509_STORE_idx(), which is used in
3778 ssl_verify_cert_chain() and thus can be called at any time
3779 during TLS/SSL handshakes so that thread-safety is essential.
3780 Unfortunately, the ex_data design is not at all suited
3781 for multi-threaded use, so it probably should be abolished.
3784 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
3785 [Broadcom, tweaked and integrated by Geoff Thorpe]
3787 *) Move common extension printing code to new function
3788 X509V3_print_extensions(). Reorganise OCSP print routines and
3789 implement some needed OCSP ASN1 functions. Add OCSP extensions.
3792 *) New function X509_signature_print() to remove duplication in some
3796 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
3797 set (this was treated exactly the same as SET OF previously). This
3798 is used to reorder the STACK representing the structure to match the
3799 encoding. This will be used to get round a problem where a PKCS7
3800 structure which was signed could not be verified because the STACK
3801 order did not reflect the encoded order.
3804 *) Reimplement the OCSP ASN1 module using the new code.
3807 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
3808 for its ASN1 operations. The old style function pointers still exist
3809 for now but they will eventually go away.
3812 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
3813 completely replaces the old ASN1 functionality with a table driven
3814 encoder and decoder which interprets an ASN1_ITEM structure describing
3815 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
3816 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
3817 has also been converted to the new form.
3820 *) Change BN_mod_exp_recp so that negative moduli are tolerated
3821 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
3822 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
3823 for negative moduli.
3826 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
3827 of not touching the result's sign bit.
3830 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
3834 *) Changed the LHASH code to use prototypes for callbacks, and created
3835 macros to declare and implement thin (optionally static) functions
3836 that provide type-safety and avoid function pointer casting for the
3837 type-specific callbacks.
3840 *) Added Kerberos Cipher Suites to be used with TLS, as written in
3842 [Veers Staats <staatsvr@asc.hpc.mil>,
3843 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
3845 *) Reformat the FAQ so the different questions and answers can be divided
3846 in sections depending on the subject.
3849 *) Have the zlib compression code load ZLIB.DLL dynamically under
3853 *) New function BN_mod_sqrt for computing square roots modulo a prime
3854 (using the probabilistic Tonelli-Shanks algorithm unless
3855 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
3856 be handled deterministically).
3857 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3859 *) Make BN_mod_inverse faster by explicitly handling small quotients
3860 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
3861 512 bits], about 30% for larger ones [1024 or 2048 bits].)
3864 *) New function BN_kronecker.
3867 *) Fix BN_gcd so that it works on negative inputs; the result is
3868 positive unless both parameters are zero.
3869 Previously something reasonably close to an infinite loop was
3870 possible because numbers could be growing instead of shrinking
3871 in the implementation of Euclid's algorithm.
3874 *) Fix BN_is_word() and BN_is_one() macros to take into account the
3875 sign of the number in question.
3877 Fix BN_is_word(a,w) to work correctly for w == 0.
3879 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
3880 because its test if the absolute value of 'a' equals 'w'.
3881 Note that BN_abs_is_word does *not* handle w == 0 reliably;
3882 it exists mostly for use in the implementations of BN_is_zero(),
3883 BN_is_one(), and BN_is_word().
3886 *) New function BN_swap.
3889 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
3890 the exponentiation functions are more likely to produce reasonable
3891 results on negative inputs.
3894 *) Change BN_mod_mul so that the result is always non-negative.
3895 Previously, it could be negative if one of the factors was negative;
3896 I don't think anyone really wanted that behaviour.
3899 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
3900 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
3901 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
3902 and add new functions:
3911 BN_mod_lshift1_quick
3915 These functions always generate non-negative results.
3917 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
3918 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
3920 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
3921 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
3922 be reduced modulo m.
3923 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3926 The following entry accidentily appeared in the CHANGES file
3927 distributed with OpenSSL 0.9.7. The modifications described in
3928 it do *not* apply to OpenSSL 0.9.7.
3930 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3931 was actually never needed) and in BN_mul(). The removal in BN_mul()
3932 required a small change in bn_mul_part_recursive() and the addition
3933 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3934 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3935 bn_sub_words() and bn_add_words() except they take arrays with
3940 *) In 'openssl passwd', verify passwords read from the terminal
3941 unless the '-salt' option is used (which usually means that
3942 verification would just waste user's time since the resulting
3943 hash is going to be compared with some given password hash)
3944 or the new '-noverify' option is used.
3946 This is an incompatible change, but it does not affect
3947 non-interactive use of 'openssl passwd' (passwords on the command
3948 line, '-stdin' option, '-in ...' option) and thus should not
3952 *) Remove all references to RSAref, since there's no more need for it.
3955 *) Make DSO load along a path given through an environment variable
3956 (SHLIB_PATH) with shl_load().
3959 *) Constify the ENGINE code as a result of BIGNUM constification.
3960 Also constify the RSA code and most things related to it. In a
3961 few places, most notable in the depth of the ASN.1 code, ugly
3962 casts back to non-const were required (to be solved at a later
3966 *) Make it so the openssl application has all engines loaded by default.
3969 *) Constify the BIGNUM routines a little more.
3972 *) Add the following functions:
3974 ENGINE_load_cswift()
3976 ENGINE_load_atalla()
3978 ENGINE_load_builtin_engines()
3980 That way, an application can itself choose if external engines that
3981 are built-in in OpenSSL shall ever be used or not. The benefit is
3982 that applications won't have to be linked with libdl or other dso
3983 libraries unless it's really needed.
3985 Changed 'openssl engine' to load all engines on demand.
3986 Changed the engine header files to avoid the duplication of some
3987 declarations (they differed!).
3990 *) 'openssl engine' can now list capabilities.
3993 *) Better error reporting in 'openssl engine'.
3996 *) Never call load_dh_param(NULL) in s_server.
3999 *) Add engine application. It can currently list engines by name and
4000 identity, and test if they are actually available.
4003 *) Improve RPM specification file by forcing symbolic linking and making
4004 sure the installed documentation is also owned by root.root.
4005 [Damien Miller <djm@mindrot.org>]
4007 *) Give the OpenSSL applications more possibilities to make use of
4008 keys (public as well as private) handled by engines.
4011 *) Add OCSP code that comes from CertCo.
4014 *) Add VMS support for the Rijndael code.
4017 *) Added untested support for Nuron crypto accelerator.
4020 *) Add support for external cryptographic devices. This code was
4021 previously distributed separately as the "engine" branch.
4022 [Geoff Thorpe, Richard Levitte]
4024 *) Rework the filename-translation in the DSO code. It is now possible to
4025 have far greater control over how a "name" is turned into a filename
4026 depending on the operating environment and any oddities about the
4027 different shared library filenames on each system.
4030 *) Support threads on FreeBSD-elf in Configure.
4033 *) Fix for SHA1 assembly problem with MASM: it produces
4034 warnings about corrupt line number information when assembling
4035 with debugging information. This is caused by the overlapping
4037 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
4040 NCONF_get_number() has no error checking at all. As a replacement,
4041 NCONF_get_number_e() is defined (_e for "error checking") and is
4042 promoted strongly. The old NCONF_get_number is kept around for
4043 binary backward compatibility.
4044 Make it possible for methods to load from something other than a BIO,
4045 by providing a function pointer that is given a name instead of a BIO.
4046 For example, this could be used to load configuration data from an
4050 *) Fix for non blocking accept BIOs. Added new I/O special reason
4051 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
4052 with non blocking I/O was not possible because no retry code was
4053 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
4057 *) Added the beginnings of Rijndael support.
4060 *) Fix for bug in DirectoryString mask setting. Add support for
4061 X509_NAME_print_ex() in 'req' and X509_print_ex() function
4062 to allow certificate printing to more controllable, additional
4063 'certopt' option to 'x509' to allow new printing options to be
4067 *) Clean old EAY MD5 hack from e_os.h.
4070 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
4072 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4073 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4074 [Joe Orton, Steve Henson]
4076 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
4078 *) Fix additional bug revealed by the NISCC test suite:
4080 Stop bug triggering large recursion when presented with
4081 certain ASN.1 tags (CVE-2003-0851)
4084 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
4086 *) Fix various bugs revealed by running the NISCC test suite:
4088 Stop out of bounds reads in the ASN1 code when presented with
4089 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4091 If verify callback ignores invalid public key errors don't try to check
4092 certificate signature with the NULL public key.
4096 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4097 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4101 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4102 extra data after the compression methods not only for TLS 1.0
4103 but also for SSL 3.0 (as required by the specification).
4104 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4106 *) Change X509_certificate_type() to mark the key as exported/exportable
4107 when it's 512 *bits* long, not 512 bytes.
4110 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
4112 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4113 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4114 a protocol version number mismatch like a decryption error
4115 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4118 *) Turn on RSA blinding by default in the default implementation
4119 to avoid a timing attack. Applications that don't want it can call
4120 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4121 They would be ill-advised to do so in most cases.
4122 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4124 *) Change RSA blinding code so that it works when the PRNG is not
4125 seeded (in this case, the secret RSA exponent is abused as
4126 an unpredictable seed -- if it is not unpredictable, there
4127 is no point in blinding anyway). Make RSA blinding thread-safe
4128 by remembering the creator's thread ID in rsa->blinding and
4129 having all other threads use local one-time blinding factors
4130 (this requires more computation than sharing rsa->blinding, but
4131 avoids excessive locking; and if an RSA object is not shared
4132 between threads, blinding will still be very fast).
4135 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
4137 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4138 via timing by performing a MAC computation even if incorrrect
4139 block cipher padding has been found. This is a countermeasure
4140 against active attacks where the attacker has to distinguish
4141 between bad padding and a MAC verification error. (CVE-2003-0078)
4143 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4144 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4145 Martin Vuagnoux (EPFL, Ilion)]
4147 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
4149 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
4150 memory from it's contents. This is done with a counter that will
4151 place alternating values in each byte. This can be used to solve
4152 two issues: 1) the removal of calls to memset() by highly optimizing
4153 compilers, and 2) cleansing with other values than 0, since those can
4154 be read through on certain media, for example a swap space on disk.
4157 *) Bugfix: client side session caching did not work with external caching,
4158 because the session->cipher setting was not restored when reloading
4159 from the external cache. This problem was masked, when
4160 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
4161 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
4164 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
4165 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
4166 [Zeev Lieber <zeev-l@yahoo.com>]
4168 *) Undo an undocumented change introduced in 0.9.6e which caused
4169 repeated calls to OpenSSL_add_all_ciphers() and
4170 OpenSSL_add_all_digests() to be ignored, even after calling
4174 *) Change the default configuration reader to deal with last line not
4175 being properly terminated.
4178 *) Change X509_NAME_cmp() so it applies the special rules on handling
4179 DN values that are of type PrintableString, as well as RDNs of type
4180 emailAddress where the value has the type ia5String.
4181 [stefank@valicert.com via Richard Levitte]
4183 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
4184 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
4185 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
4186 the bitwise-OR of the two for use by the majority of applications
4187 wanting this behaviour, and update the docs. The documented
4188 behaviour and actual behaviour were inconsistent and had been
4189 changing anyway, so this is more a bug-fix than a behavioural
4191 [Geoff Thorpe, diagnosed by Nadav Har'El]
4193 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
4194 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
4197 *) Fix initialization code race conditions in
4198 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
4199 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
4200 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
4201 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
4202 ssl2_get_cipher_by_char(),
4203 ssl3_get_cipher_by_char().
4204 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
4206 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
4207 the cached sessions are flushed, as the remove_cb() might use ex_data
4208 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
4209 (see [openssl.org #212]).
4210 [Geoff Thorpe, Lutz Jaenicke]
4212 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
4213 length, instead of the encoding length to d2i_ASN1_OBJECT.
4216 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
4218 *) [In 0.9.6g-engine release:]
4219 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
4220 [Lynn Gazis <lgazis@rainbow.com>]
4222 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
4224 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
4225 and get fix the header length calculation.
4226 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
4227 Alon Kantor <alonk@checkpoint.com> (and others),
4230 *) Use proper error handling instead of 'assertions' in buffer
4231 overflow checks added in 0.9.6e. This prevents DoS (the
4232 assertions could call abort()).
4233 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
4235 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
4237 *) Add various sanity checks to asn1_get_length() to reject
4238 the ASN1 length bytes if they exceed sizeof(long), will appear
4239 negative or the content length exceeds the length of the
4241 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4243 *) Fix cipher selection routines: ciphers without encryption had no flags
4244 for the cipher strength set and where therefore not handled correctly
4245 by the selection routines (PR #130).
4248 *) Fix EVP_dsa_sha macro.
4252 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
4253 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
4254 that was added in OpenSSL 0.9.6d.
4256 As the countermeasure turned out to be incompatible with some
4257 broken SSL implementations, the new option is part of SSL_OP_ALL.
4258 SSL_OP_ALL is usually employed when compatibility with weird SSL
4259 implementations is desired (e.g. '-bugs' option to 's_client' and
4260 's_server'), so the new option is automatically set in many
4264 *) Changes in security patch:
4266 Changes marked "(CHATS)" were sponsored by the Defense Advanced
4267 Research Projects Agency (DARPA) and Air Force Research Laboratory,
4268 Air Force Materiel Command, USAF, under agreement number
4271 *) Add various sanity checks to asn1_get_length() to reject
4272 the ASN1 length bytes if they exceed sizeof(long), will appear
4273 negative or the content length exceeds the length of the
4274 supplied buffer. (CVE-2002-0659)
4275 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4277 *) Assertions for various potential buffer overflows, not known to
4279 [Ben Laurie (CHATS)]
4281 *) Various temporary buffers to hold ASCII versions of integers were
4282 too small for 64 bit platforms. (CVE-2002-0655)
4283 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
4285 *) Remote buffer overflow in SSL3 protocol - an attacker could
4286 supply an oversized session ID to a client. (CVE-2002-0656)
4287 [Ben Laurie (CHATS)]
4289 *) Remote buffer overflow in SSL2 protocol - an attacker could
4290 supply an oversized client master key. (CVE-2002-0656)
4291 [Ben Laurie (CHATS)]
4293 Changes between 0.9.6c and 0.9.6d [9 May 2002]
4295 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
4296 encoded as NULL) with id-dsa-with-sha1.
4297 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
4299 *) Check various X509_...() return values in apps/req.c.
4300 [Nils Larsch <nla@trustcenter.de>]
4302 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
4303 an end-of-file condition would erronously be flagged, when the CRLF
4304 was just at the end of a processed block. The bug was discovered when
4305 processing data through a buffering memory BIO handing the data to a
4306 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
4307 <ptsekov@syntrex.com> and Nedelcho Stanev.
4310 *) Implement a countermeasure against a vulnerability recently found
4311 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
4312 before application data chunks to avoid the use of known IVs
4313 with data potentially chosen by the attacker.
4316 *) Fix length checks in ssl3_get_client_hello().
4319 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
4320 to prevent ssl3_read_internal() from incorrectly assuming that
4321 ssl3_read_bytes() found application data while handshake
4322 processing was enabled when in fact s->s3->in_read_app_data was
4323 merely automatically cleared during the initial handshake.
4324 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
4326 *) Fix object definitions for Private and Enterprise: they were not
4327 recognized in their shortname (=lowercase) representation. Extend
4328 obj_dat.pl to issue an error when using undefined keywords instead
4329 of silently ignoring the problem (Svenning Sorensen
4330 <sss@sss.dnsalias.net>).
4333 *) Fix DH_generate_parameters() so that it works for 'non-standard'
4334 generators, i.e. generators other than 2 and 5. (Previously, the
4335 code did not properly initialise the 'add' and 'rem' values to
4336 BN_generate_prime().)
4338 In the new general case, we do not insist that 'generator' is
4339 actually a primitive root: This requirement is rather pointless;
4340 a generator of the order-q subgroup is just as good, if not
4344 *) Map new X509 verification errors to alerts. Discovered and submitted by
4345 Tom Wu <tom@arcot.com>.
4348 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
4349 returning non-zero before the data has been completely received
4350 when using non-blocking I/O.
4351 [Bodo Moeller; problem pointed out by John Hughes]
4353 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
4354 [Ben Laurie, Lutz Jaenicke]
4356 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
4357 Yoram Zahavi <YoramZ@gilian.com>).
4360 *) Add information about CygWin 1.3 and on, and preserve proper
4361 configuration for the versions before that.
4362 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4364 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
4365 check whether we deal with a copy of a session and do not delete from
4366 the cache in this case. Problem reported by "Izhar Shoshani Levi"
4367 <izhar@checkpoint.com>.
4370 *) Do not store session data into the internal session cache, if it
4371 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
4372 flag is set). Proposed by Aslam <aslam@funk.com>.
4375 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
4379 *) [In 0.9.6d-engine release:]
4380 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4381 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4383 *) Add the configuration target linux-s390x.
4384 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
4386 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
4387 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
4388 variable as an indication that a ClientHello message has been
4389 received. As the flag value will be lost between multiple
4390 invocations of ssl3_accept when using non-blocking I/O, the
4391 function may not be aware that a handshake has actually taken
4392 place, thus preventing a new session from being added to the
4395 To avoid this problem, we now set s->new_session to 2 instead of
4396 using a local variable.
4397 [Lutz Jaenicke, Bodo Moeller]
4399 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
4400 if the SSL_R_LENGTH_MISMATCH error is detected.
4401 [Geoff Thorpe, Bodo Moeller]
4403 *) New 'shared_ldflag' column in Configure platform table.
4406 *) Fix EVP_CIPHER_mode macro.
4407 ["Dan S. Camper" <dan@bti.net>]
4409 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
4410 type, we must throw them away by setting rr->length to 0.
4411 [D P Chang <dpc@qualys.com>]
4413 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
4415 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
4416 <Dominikus.Scherkl@biodata.com>. (The previous implementation
4417 worked incorrectly for those cases where range = 10..._2 and
4418 3*range is two bits longer than range.)
4421 *) Only add signing time to PKCS7 structures if it is not already
4425 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
4426 OBJ_ld_ce should be OBJ_id_ce.
4427 Also some ip-pda OIDs in crypto/objects/objects.txt were
4428 incorrect (cf. RFC 3039).
4429 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
4431 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
4432 returns early because it has nothing to do.
4433 [Andy Schneider <andy.schneider@bjss.co.uk>]
4435 *) [In 0.9.6c-engine release:]
4436 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
4437 [Andy Schneider <andy.schneider@bjss.co.uk>]
4439 *) [In 0.9.6c-engine release:]
4440 Add support for Cryptographic Appliance's keyserver technology.
4441 (Use engine 'keyclient')
4442 [Cryptographic Appliances and Geoff Thorpe]
4444 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
4445 is called via tools/c89.sh because arguments have to be
4446 rearranged (all '-L' options must appear before the first object
4448 [Richard Shapiro <rshapiro@abinitio.com>]
4450 *) [In 0.9.6c-engine release:]
4451 Add support for Broadcom crypto accelerator cards, backported
4453 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
4455 *) [In 0.9.6c-engine release:]
4456 Add support for SureWare crypto accelerator cards from
4457 Baltimore Technologies. (Use engine 'sureware')
4458 [Baltimore Technologies and Mark Cox]
4460 *) [In 0.9.6c-engine release:]
4461 Add support for crypto accelerator cards from Accelerated
4462 Encryption Processing, www.aep.ie. (Use engine 'aep')
4463 [AEP Inc. and Mark Cox]
4465 *) Add a configuration entry for gcc on UnixWare.
4466 [Gary Benson <gbenson@redhat.com>]
4468 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
4469 messages are stored in a single piece (fixed-length part and
4470 variable-length part combined) and fix various bugs found on the way.
4473 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
4474 instead. BIO_gethostbyname() does not know what timeouts are
4475 appropriate, so entries would stay in cache even when they have
4477 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
4479 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
4480 faced with a pathologically small ClientHello fragment that does
4481 not contain client_version: Instead of aborting with an error,
4482 simply choose the highest available protocol version (i.e.,
4483 TLS 1.0 unless it is disabled). In practice, ClientHello
4484 messages are never sent like this, but this change gives us
4485 strictly correct behaviour at least for TLS.
4488 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
4489 never resets s->method to s->ctx->method when called from within
4490 one of the SSL handshake functions.
4491 [Bodo Moeller; problem pointed out by Niko Baric]
4493 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
4494 (sent using the client's version number) if client_version is
4495 smaller than the protocol version in use. Also change
4496 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
4497 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
4498 the client will at least see that alert.
4501 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
4505 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
4506 client receives HelloRequest while in a handshake.
4507 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
4509 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
4510 should end in 'break', not 'goto end' which circuments various
4511 cleanups done in state SSL_ST_OK. But session related stuff
4512 must be disabled for SSL_ST_OK in the case that we just sent a
4515 Also avoid some overhead by not calling ssl_init_wbio_buffer()
4516 before just sending a HelloRequest.
4517 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
4519 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
4520 reveal whether illegal block cipher padding was found or a MAC
4521 verification error occured. (Neither SSLerr() codes nor alerts
4522 are directly visible to potential attackers, but the information
4523 may leak via logfiles.)
4525 Similar changes are not required for the SSL 2.0 implementation
4526 because the number of padding bytes is sent in clear for SSL 2.0,
4527 and the extra bytes are just ignored. However ssl/s2_pkt.c
4528 failed to verify that the purported number of padding bytes is in
4532 *) Add OpenUNIX-8 support including shared libraries
4533 (Boyd Lynn Gerber <gerberb@zenez.com>).
4536 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
4537 'wristwatch attack' using huge encoding parameters (cf.
4538 James H. Manger's CRYPTO 2001 paper). Note that the
4539 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
4540 encoding parameters and hence was not vulnerable.
4543 *) BN_sqr() bug fix.
4544 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
4546 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
4547 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
4548 followed by modular reduction.
4549 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
4551 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
4552 equivalent based on BN_pseudo_rand() instead of BN_rand().
4555 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
4556 This function was broken, as the check for a new client hello message
4557 to handle SGC did not allow these large messages.
4558 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
4561 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
4564 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
4565 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
4568 *) Rework the configuration and shared library support for Tru64 Unix.
4569 The configuration part makes use of modern compiler features and
4570 still retains old compiler behavior for those that run older versions
4571 of the OS. The shared library support part includes a variant that
4572 uses the RPATH feature, and is available through the special
4573 configuration target "alpha-cc-rpath", which will never be selected
4575 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
4577 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
4578 with the same message size as in ssl3_get_certificate_request().
4579 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
4580 messages might inadvertently be reject as too long.
4581 [Petr Lampa <lampa@fee.vutbr.cz>]
4583 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
4586 *) Modified SSL library such that the verify_callback that has been set
4587 specificly for an SSL object with SSL_set_verify() is actually being
4588 used. Before the change, a verify_callback set with this function was
4589 ignored and the verify_callback() set in the SSL_CTX at the time of
4590 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
4591 to allow the necessary settings.
4594 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
4595 explicitly to NULL, as at least on Solaris 8 this seems not always to be
4596 done automatically (in contradiction to the requirements of the C
4597 standard). This made problems when used from OpenSSH.
4600 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
4601 dh->length and always used
4603 BN_rand_range(priv_key, dh->p).
4605 BN_rand_range() is not necessary for Diffie-Hellman, and this
4606 specific range makes Diffie-Hellman unnecessarily inefficient if
4607 dh->length (recommended exponent length) is much smaller than the
4608 length of dh->p. We could use BN_rand_range() if the order of
4609 the subgroup was stored in the DH structure, but we only have
4614 BN_rand(priv_key, l, ...)
4616 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
4622 RSA_eay_public_encrypt
4623 RSA_eay_private_decrypt
4624 RSA_eay_private_encrypt (signing)
4625 RSA_eay_public_decrypt (signature verification)
4627 (default implementations for RSA_public_encrypt,
4628 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
4629 always reject numbers >= n.
4632 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
4633 to synchronize access to 'locking_thread'. This is necessary on
4634 systems where access to 'locking_thread' (an 'unsigned long'
4635 variable) is not atomic.
4638 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
4639 *before* setting the 'crypto_lock_rand' flag. The previous code had
4640 a race condition if 0 is a valid thread ID.
4641 [Travis Vitek <vitek@roguewave.com>]
4643 *) Add support for shared libraries under Irix.
4644 [Albert Chin-A-Young <china@thewrittenword.com>]
4646 *) Add configuration option to build on Linux on both big-endian and
4648 [Ralf Baechle <ralf@uni-koblenz.de>]
4650 *) Add the possibility to create shared libraries on HP-UX.
4653 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
4655 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
4656 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
4657 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
4658 PRNG state recovery was possible based on the output of
4659 one PRNG request appropriately sized to gain knowledge on
4660 'md' followed by enough consecutive 1-byte PRNG requests
4661 to traverse all of 'state'.
4663 1. When updating 'md_local' (the current thread's copy of 'md')
4664 during PRNG output generation, hash all of the previous
4665 'md_local' value, not just the half used for PRNG output.
4667 2. Make the number of bytes from 'state' included into the hash
4668 independent from the number of PRNG bytes requested.
4670 The first measure alone would be sufficient to avoid
4671 Markku-Juhani's attack. (Actually it had never occurred
4672 to me that the half of 'md_local' used for chaining was the
4673 half from which PRNG output bytes were taken -- I had always
4674 assumed that the secret half would be used.) The second
4675 measure makes sure that additional data from 'state' is never
4676 mixed into 'md_local' in small portions; this heuristically
4677 further strengthens the PRNG.
4680 *) Fix crypto/bn/asm/mips3.s.
4683 *) When only the key is given to "enc", the IV is undefined. Print out
4684 an error message in this case.
4687 *) Handle special case when X509_NAME is empty in X509 printing routines.
4690 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
4691 positive and less than q.
4694 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
4695 used: it isn't thread safe and the add_lock_callback should handle
4697 [Paul Rose <Paul.Rose@bridge.com>]
4699 *) Verify that incoming data obeys the block size in
4700 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
4704 [Ulf Möller, Bodo Möller]
4706 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
4707 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
4708 when fixing the server behaviour for backwards-compatible 'client
4709 hello' messages. (Note that the attack is impractical against
4710 SSL 3.0 and TLS 1.0 anyway because length and version checking
4711 means that the probability of guessing a valid ciphertext is
4712 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
4715 Before 0.9.5, the countermeasure (hide the error by generating a
4716 random 'decryption result') did not work properly because
4717 ERR_clear_error() was missing, meaning that SSL_get_error() would
4718 detect the supposedly ignored error.
4720 Both problems are now fixed.
4723 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
4724 (previously it was 1024).
4727 *) Fix for compatibility mode trust settings: ignore trust settings
4728 unless some valid trust or reject settings are present.
4731 *) Fix for blowfish EVP: its a variable length cipher.
4734 *) Fix various bugs related to DSA S/MIME verification. Handle missing
4735 parameters in DSA public key structures and return an error in the
4736 DSA routines if parameters are absent.
4739 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
4740 in the current directory if neither $RANDFILE nor $HOME was set.
4741 RAND_file_name() in 0.9.6a returned NULL in this case. This has
4742 caused some confusion to Windows users who haven't defined $HOME.
4743 Thus RAND_file_name() is changed again: e_os.h can define a
4744 DEFAULT_HOME, which will be used if $HOME is not set.
4745 For Windows, we use "C:"; on other platforms, we still require
4746 environment variables.
4748 *) Move 'if (!initialized) RAND_poll()' into regions protected by
4749 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
4750 having multiple threads call RAND_poll() concurrently.
4753 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
4754 combination of a flag and a thread ID variable.
4755 Otherwise while one thread is in ssleay_rand_bytes (which sets the
4756 flag), *other* threads can enter ssleay_add_bytes without obeying
4757 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
4758 that they do not hold after the first thread unsets add_do_not_lock).
4761 *) Change bctest again: '-x' expressions are not available in all
4765 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
4767 *) Fix a couple of memory leaks in PKCS7_dataDecode()
4768 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
4770 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
4771 the default extension for executables, if any. Also, make the perl
4772 scripts that use symlink() to test if it really exists and use "cp"
4773 if it doesn't. All this made OpenSSL compilable and installable in
4777 *) Fix for asn1_GetSequence() for indefinite length constructed data.
4778 If SEQUENCE is length is indefinite just set c->slen to the total
4779 amount of data available.
4780 [Steve Henson, reported by shige@FreeBSD.org]
4781 [This change does not apply to 0.9.7.]
4783 *) Change bctest to avoid here-documents inside command substitution
4784 (workaround for FreeBSD /bin/sh bug).
4785 For compatibility with Ultrix, avoid shell functions (introduced
4786 in the bctest version that searches along $PATH).
4789 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
4790 with des_encrypt() defined on some operating systems, like Solaris
4794 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
4795 On the Importance of Eliminating Errors in Cryptographic
4796 Computations, J. Cryptology 14 (2001) 2, 101-119,
4797 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
4800 *) MIPS assembler BIGNUM division bug fix.
4803 *) Disabled incorrect Alpha assembler code.
4806 *) Fix PKCS#7 decode routines so they correctly update the length
4807 after reading an EOC for the EXPLICIT tag.
4809 [This change does not apply to 0.9.7.]
4811 *) Fix bug in PKCS#12 key generation routines. This was triggered
4812 if a 3DES key was generated with a 0 initial byte. Include
4813 PKCS12_BROKEN_KEYGEN compilation option to retain the old
4814 (but broken) behaviour.
4817 *) Enhance bctest to search for a working bc along $PATH and print
4819 [Tim Rice <tim@multitalents.net> via Richard Levitte]
4821 *) Fix memory leaks in err.c: free err_data string if necessary;
4822 don't write to the wrong index in ERR_set_error_data.
4825 *) Implement ssl23_peek (analogous to ssl23_read), which previously
4829 *) Replace rdtsc with _emit statements for VC++ version 5.
4830 [Jeremy Cooper <jeremy@baymoo.org>]
4832 *) Make it possible to reuse SSLv2 sessions.
4835 *) In copy_email() check for >= 0 as a return value for
4836 X509_NAME_get_index_by_NID() since 0 is a valid index.
4837 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
4839 *) Avoid coredump with unsupported or invalid public keys by checking if
4840 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
4841 PKCS7_verify() fails with non detached data.
4844 *) Don't use getenv in library functions when run as setuid/setgid.
4845 New function OPENSSL_issetugid().
4848 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
4849 due to incorrect handling of multi-threading:
4851 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
4853 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
4855 3. Count how many times MemCheck_off() has been called so that
4856 nested use can be treated correctly. This also avoids
4857 inband-signalling in the previous code (which relied on the
4858 assumption that thread ID 0 is impossible).
4861 *) Add "-rand" option also to s_client and s_server.
4864 *) Fix CPU detection on Irix 6.x.
4865 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
4866 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4868 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
4871 [This change does not apply to 0.9.7.]
4873 *) Use the cached encoding of an X509_NAME structure rather than
4874 copying it. This is apparently the reason for the libsafe "errors"
4875 but the code is actually correct.
4878 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
4879 Bleichenbacher's DSA attack.
4880 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
4881 to be set and top=0 forces the highest bit to be set; top=-1 is new
4882 and leaves the highest bit random.
4883 [Ulf Moeller, Bodo Moeller]
4885 *) In the NCONF_...-based implementations for CONF_... queries
4886 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
4887 a temporary CONF structure with the data component set to NULL
4888 (which gives segmentation faults in lh_retrieve).
4889 Instead, use NULL for the CONF pointer in CONF_get_string and
4890 CONF_get_number (which may use environment variables) and directly
4891 return NULL from CONF_get_section.
4894 *) Fix potential buffer overrun for EBCDIC.
4897 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
4898 keyUsage if basicConstraints absent for a CA.
4901 *) Make SMIME_write_PKCS7() write mail header values with a format that
4902 is more generally accepted (no spaces before the semicolon), since
4903 some programs can't parse those values properly otherwise. Also make
4904 sure BIO's that break lines after each write do not create invalid
4908 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
4909 macros previously used would not encode an empty SEQUENCE OF
4910 and break the signature.
4912 [This change does not apply to 0.9.7.]
4914 *) Zero the premaster secret after deriving the master secret in
4918 *) Add some EVP_add_digest_alias registrations (as found in
4919 OpenSSL_add_all_digests()) to SSL_library_init()
4920 aka OpenSSL_add_ssl_algorithms(). This provides improved
4921 compatibility with peers using X.509 certificates
4922 with unconventional AlgorithmIdentifier OIDs.
4925 *) Fix for Irix with NO_ASM.
4926 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4928 *) ./config script fixes.
4929 [Ulf Moeller, Richard Levitte]
4931 *) Fix 'openssl passwd -1'.
4934 *) Change PKCS12_key_gen_asc() so it can cope with non null
4935 terminated strings whose length is passed in the passlen
4936 parameter, for example from PEM callbacks. This was done
4937 by adding an extra length parameter to asc2uni().
4938 [Steve Henson, reported by <oddissey@samsung.co.kr>]
4940 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
4941 call failed, free the DSA structure.
4944 *) Fix to uni2asc() to cope with zero length Unicode strings.
4945 These are present in some PKCS#12 files.
4948 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
4949 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
4950 when writing a 32767 byte record.
4951 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
4953 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
4954 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
4956 (RSA objects have a reference count access to which is protected
4957 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
4958 so they are meant to be shared between threads.)
4959 [Bodo Moeller, Geoff Thorpe; original patch submitted by
4960 "Reddie, Steven" <Steven.Reddie@ca.com>]
4962 *) Fix a deadlock in CRYPTO_mem_leaks().
4965 *) Use better test patterns in bntest.
4968 *) rand_win.c fix for Borland C.
4971 *) BN_rshift bugfix for n == 0.
4974 *) Add a 'bctest' script that checks for some known 'bc' bugs
4975 so that 'make test' does not abort just because 'bc' is broken.
4978 *) Store verify_result within SSL_SESSION also for client side to
4979 avoid potential security hole. (Re-used sessions on the client side
4980 always resulted in verify_result==X509_V_OK, not using the original
4981 result of the server certificate verification.)
4984 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
4985 SSL3_RT_APPLICATION_DATA, return 0.
4986 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
4990 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
4991 releases, have been re-implemented by renaming the previous
4992 implementations of ssl2_read and ssl3_read to ssl2_read_internal
4993 and ssl3_read_internal, respectively, and adding 'peek' parameters
4994 to them. The new ssl[23]_{read,peek} functions are calls to
4995 ssl[23]_read_internal with the 'peek' flag set appropriately.
4996 A 'peek' parameter has also been added to ssl3_read_bytes, which
4997 does the actual work for ssl3_read_internal.
5000 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
5001 the method-specific "init()" handler. Also clean up ex_data after
5002 calling the method-specific "finish()" handler. Previously, this was
5003 happening the other way round.
5006 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
5007 The previous value, 12, was not always sufficient for BN_mod_exp().
5010 *) Make sure that shared libraries get the internal name engine with
5011 the full version number and not just 0. This should mark the
5012 shared libraries as not backward compatible. Of course, this should
5013 be changed again when we can guarantee backward binary compatibility.
5016 *) Fix typo in get_cert_by_subject() in by_dir.c
5017 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
5019 *) Rework the system to generate shared libraries:
5021 - Make note of the expected extension for the shared libraries and
5022 if there is a need for symbolic links from for example libcrypto.so.0
5023 to libcrypto.so.0.9.7. There is extended info in Configure for
5026 - Make as few rebuilds of the shared libraries as possible.
5028 - Still avoid linking the OpenSSL programs with the shared libraries.
5030 - When installing, install the shared libraries separately from the
5034 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
5036 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
5037 and not in SSL_clear because the latter is also used by the
5038 accept/connect functions; previously, the settings made by
5039 SSL_set_read_ahead would be lost during the handshake.
5040 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
5042 *) Correct util/mkdef.pl to be selective about disabled algorithms.
5043 Previously, it would create entries for disableed algorithms no
5047 *) Added several new manual pages for SSL_* function.
5050 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
5052 *) In ssl23_get_client_hello, generate an error message when faced
5053 with an initial SSL 3.0/TLS record that is too small to contain the
5054 first two bytes of the ClientHello message, i.e. client_version.
5055 (Note that this is a pathologic case that probably has never happened
5056 in real life.) The previous approach was to use the version number
5057 from the record header as a substitute; but our protocol choice
5058 should not depend on that one because it is not authenticated
5059 by the Finished messages.
5062 *) More robust randomness gathering functions for Windows.
5063 [Jeffrey Altman <jaltman@columbia.edu>]
5065 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
5066 not set then we don't setup the error code for issuer check errors
5067 to avoid possibly overwriting other errors which the callback does
5068 handle. If an application does set the flag then we assume it knows
5069 what it is doing and can handle the new informational codes
5073 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
5074 a general "ANY" type, as such it should be able to decode anything
5075 including tagged types. However it didn't check the class so it would
5076 wrongly interpret tagged types in the same way as their universal
5077 counterpart and unknown types were just rejected. Changed so that the
5078 tagged and unknown types are handled in the same way as a SEQUENCE:
5079 that is the encoding is stored intact. There is also a new type
5080 "V_ASN1_OTHER" which is used when the class is not universal, in this
5081 case we have no idea what the actual type is so we just lump them all
5085 *) On VMS, stdout may very well lead to a file that is written to
5086 in a record-oriented fashion. That means that every write() will
5087 write a separate record, which will be read separately by the
5088 programs trying to read from it. This can be very confusing.
5090 The solution is to put a BIO filter in the way that will buffer
5091 text until a linefeed is reached, and then write everything a
5092 line at a time, so every record written will be an actual line,
5093 not chunks of lines and not (usually doesn't happen, but I've
5094 seen it once) several lines in one record. BIO_f_linebuffer() is
5097 Currently, it's a VMS-only method, because that's where it has
5098 been tested well enough.
5101 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
5102 it can return incorrect results.
5103 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
5104 but it was in 0.9.6-beta[12].)
5107 *) Disable the check for content being present when verifying detached
5108 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
5109 include zero length content when signing messages.
5112 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
5113 BIO_ctrl (for BIO pairs).
5116 *) Add DSO method for VMS.
5119 *) Bug fix: Montgomery multiplication could produce results with the
5123 *) Add RPM specification openssl.spec and modify it to build three
5124 packages. The default package contains applications, application
5125 documentation and run-time libraries. The devel package contains
5126 include files, static libraries and function documentation. The
5127 doc package contains the contents of the doc directory. The original
5128 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
5131 *) Add a large number of documentation files for many SSL routines.
5132 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5134 *) Add a configuration entry for Sony News 4.
5135 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
5137 *) Don't set the two most significant bits to one when generating a
5138 random number < q in the DSA library.
5141 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
5142 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
5143 the underlying transport is blocking) if a handshake took place.
5144 (The default behaviour is needed by applications such as s_client
5145 and s_server that use select() to determine when to use SSL_read;
5146 but for applications that know in advance when to expect data, it
5147 just makes things more complicated.)
5150 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
5154 *) Add a few more EBCDIC conditionals that make `req' and `x509'
5155 work better on such systems.
5156 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5158 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
5159 Update PKCS12_parse() so it copies the friendlyName and the
5160 keyid to the certificates aux info.
5163 *) Fix bug in PKCS7_verify() which caused an infinite loop
5164 if there was more than one signature.
5165 [Sven Uszpelkat <su@celocom.de>]
5167 *) Major change in util/mkdef.pl to include extra information
5168 about each symbol, as well as presentig variables as well
5169 as functions. This change means that there's n more need
5170 to rebuild the .num files when some algorithms are excluded.
5173 *) Allow the verify time to be set by an application,
5174 rather than always using the current time.
5177 *) Phase 2 verify code reorganisation. The certificate
5178 verify code now looks up an issuer certificate by a
5179 number of criteria: subject name, authority key id
5180 and key usage. It also verifies self signed certificates
5181 by the same criteria. The main comparison function is
5182 X509_check_issued() which performs these checks.
5184 Lot of changes were necessary in order to support this
5185 without completely rewriting the lookup code.
5187 Authority and subject key identifier are now cached.
5189 The LHASH 'certs' is X509_STORE has now been replaced
5190 by a STACK_OF(X509_OBJECT). This is mainly because an
5191 LHASH can't store or retrieve multiple objects with
5192 the same hash value.
5194 As a result various functions (which were all internal
5195 use only) have changed to handle the new X509_STORE
5196 structure. This will break anything that messed round
5197 with X509_STORE internally.
5199 The functions X509_STORE_add_cert() now checks for an
5200 exact match, rather than just subject name.
5202 The X509_STORE API doesn't directly support the retrieval
5203 of multiple certificates matching a given criteria, however
5204 this can be worked round by performing a lookup first
5205 (which will fill the cache with candidate certificates)
5206 and then examining the cache for matches. This is probably
5207 the best we can do without throwing out X509_LOOKUP
5208 entirely (maybe later...).
5210 The X509_VERIFY_CTX structure has been enhanced considerably.
5212 All certificate lookup operations now go via a get_issuer()
5213 callback. Although this currently uses an X509_STORE it
5214 can be replaced by custom lookups. This is a simple way
5215 to bypass the X509_STORE hackery necessary to make this
5216 work and makes it possible to use more efficient techniques
5217 in future. A very simple version which uses a simple
5218 STACK for its trusted certificate store is also provided
5219 using X509_STORE_CTX_trusted_stack().
5221 The verify_cb() and verify() callbacks now have equivalents
5222 in the X509_STORE_CTX structure.
5224 X509_STORE_CTX also has a 'flags' field which can be used
5225 to customise the verify behaviour.
5228 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
5229 excludes S/MIME capabilities.
5232 *) When a certificate request is read in keep a copy of the
5233 original encoding of the signed data and use it when outputing
5234 again. Signatures then use the original encoding rather than
5235 a decoded, encoded version which may cause problems if the
5236 request is improperly encoded.
5239 *) For consistency with other BIO_puts implementations, call
5240 buffer_write(b, ...) directly in buffer_puts instead of calling
5243 In BIO_puts, increment b->num_write as in BIO_write.
5244 [Peter.Sylvester@EdelWeb.fr]
5246 *) Fix BN_mul_word for the case where the word is 0. (We have to use
5247 BN_zero, we may not return a BIGNUM with an array consisting of
5251 *) Avoid calling abort() from within the library when problems are
5252 detected, except if preprocessor symbols have been defined
5253 (such as REF_CHECK, BN_DEBUG etc.).
5256 *) New openssl application 'rsautl'. This utility can be
5257 used for low level RSA operations. DER public key
5258 BIO/fp routines also added.
5261 *) New Configure entry and patches for compiling on QNX 4.
5262 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
5264 *) A demo state-machine implementation was sponsored by
5265 Nuron (http://www.nuron.com/) and is now available in
5266 demos/state_machine.
5269 *) New options added to the 'dgst' utility for signature
5270 generation and verification.
5273 *) Unrecognized PKCS#7 content types are now handled via a
5274 catch all ASN1_TYPE structure. This allows unsupported
5275 types to be stored as a "blob" and an application can
5276 encode and decode it manually.
5279 *) Fix various signed/unsigned issues to make a_strex.c
5281 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
5283 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
5284 length if passed a buffer. ASN1_INTEGER_to_BN failed
5285 if passed a NULL BN and its argument was negative.
5286 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
5288 *) Modification to PKCS#7 encoding routines to output definite
5289 length encoding. Since currently the whole structures are in
5290 memory there's not real point in using indefinite length
5291 constructed encoding. However if OpenSSL is compiled with
5292 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
5295 *) Added BIO_vprintf() and BIO_vsnprintf().
5298 *) Added more prefixes to parse for in the the strings written
5299 through a logging bio, to cover all the levels that are available
5300 through syslog. The prefixes are now:
5302 PANIC, EMERG, EMR => LOG_EMERG
5303 ALERT, ALR => LOG_ALERT
5304 CRIT, CRI => LOG_CRIT
5305 ERROR, ERR => LOG_ERR
5306 WARNING, WARN, WAR => LOG_WARNING
5307 NOTICE, NOTE, NOT => LOG_NOTICE
5308 INFO, INF => LOG_INFO
5309 DEBUG, DBG => LOG_DEBUG
5311 and as before, if none of those prefixes are present at the
5312 beginning of the string, LOG_ERR is chosen.
5314 On Win32, the LOG_* levels are mapped according to this:
5316 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
5317 LOG_WARNING => EVENTLOG_WARNING_TYPE
5318 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
5322 *) Made it possible to reconfigure with just the configuration
5323 argument "reconf" or "reconfigure". The command line arguments
5324 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
5325 and are retrieved from there when reconfiguring.
5329 [Assar Westerlund <assar@sics.se>, Richard Levitte]
5331 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
5334 *) The obj_dat.pl script was messing up the sorting of object
5335 names. The reason was that it compared the quoted version
5336 of strings as a result "OCSP" > "OCSP Signing" because
5337 " > SPACE. Changed script to store unquoted versions of
5338 names and add quotes on output. It was also omitting some
5339 names from the lookup table if they were given a default
5340 value (that is if SN is missing it is given the same
5341 value as LN and vice versa), these are now added on the
5342 grounds that if an object has a name we should be able to
5343 look it up. Finally added warning output when duplicate
5344 short or long names are found.
5347 *) Changes needed for Tandem NSK.
5348 [Scott Uroff <scott@xypro.com>]
5350 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
5351 RSA_padding_check_SSLv23(), special padding was never detected
5352 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
5353 version rollback attacks was not effective.
5355 In s23_clnt.c, don't use special rollback-attack detection padding
5356 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
5357 client; similarly, in s23_srvr.c, don't do the rollback check if
5358 SSL 2.0 is the only protocol enabled in the server.
5361 *) Make it possible to get hexdumps of unprintable data with 'openssl
5362 asn1parse'. By implication, the functions ASN1_parse_dump() and
5363 BIO_dump_indent() are added.
5366 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
5367 these print out strings and name structures based on various
5368 flags including RFC2253 support and proper handling of
5369 multibyte characters. Added options to the 'x509' utility
5370 to allow the various flags to be set.
5373 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
5374 Also change the functions X509_cmp_current_time() and
5375 X509_gmtime_adj() work with an ASN1_TIME structure,
5376 this will enable certificates using GeneralizedTime in validity
5377 dates to be checked.
5380 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
5381 negative public key encodings) on by default,
5382 NO_NEG_PUBKEY_BUG can be set to disable it.
5385 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
5386 content octets. An i2c_ASN1_OBJECT is unnecessary because
5387 the encoding can be trivially obtained from the structure.
5390 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
5391 not read locks (CRYPTO_r_[un]lock).
5394 *) A first attempt at creating official support for shared
5395 libraries through configuration. I've kept it so the
5396 default is static libraries only, and the OpenSSL programs
5397 are always statically linked for now, but there are
5398 preparations for dynamic linking in place.
5399 This has been tested on Linux and Tru64.
5402 *) Randomness polling function for Win9x, as described in:
5403 Peter Gutmann, Software Generation of Practically Strong
5407 *) Fix so PRNG is seeded in req if using an already existing
5411 *) New options to smime application. -inform and -outform
5412 allow alternative formats for the S/MIME message including
5413 PEM and DER. The -content option allows the content to be
5414 specified separately. This should allow things like Netscape
5415 form signing output easier to verify.
5418 *) Fix the ASN1 encoding of tags using the 'long form'.
5421 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
5422 STRING types. These convert content octets to and from the
5423 underlying type. The actual tag and length octets are
5424 already assumed to have been read in and checked. These
5425 are needed because all other string types have virtually
5426 identical handling apart from the tag. By having versions
5427 of the ASN1 functions that just operate on content octets
5428 IMPLICIT tagging can be handled properly. It also allows
5429 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
5430 and ASN1_INTEGER are identical apart from the tag.
5433 *) Change the handling of OID objects as follows:
5435 - New object identifiers are inserted in objects.txt, following
5436 the syntax given in objects.README.
5437 - objects.pl is used to process obj_mac.num and create a new
5439 - obj_dat.pl is used to create a new obj_dat.h, using the data in
5442 This is currently kind of a hack, and the perl code in objects.pl
5443 isn't very elegant, but it works as I intended. The simplest way
5444 to check that it worked correctly is to look in obj_dat.h and
5445 check the array nid_objs and make sure the objects haven't moved
5446 around (this is important!). Additions are OK, as well as
5447 consistent name changes.
5450 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
5453 *) Addition of the command line parameter '-rand file' to 'openssl req'.
5454 The given file adds to whatever has already been seeded into the
5455 random pool through the RANDFILE configuration file option or
5456 environment variable, or the default random state file.
5459 *) mkstack.pl now sorts each macro group into lexical order.
5460 Previously the output order depended on the order the files
5461 appeared in the directory, resulting in needless rewriting
5465 *) Patches to make OpenSSL compile under Win32 again. Mostly
5466 work arounds for the VC++ problem that it treats func() as
5467 func(void). Also stripped out the parts of mkdef.pl that
5468 added extra typesafe functions: these no longer exist.
5471 *) Reorganisation of the stack code. The macros are now all
5472 collected in safestack.h . Each macro is defined in terms of
5473 a "stack macro" of the form SKM_<name>(type, a, b). The
5474 DEBUG_SAFESTACK is now handled in terms of function casts,
5475 this has the advantage of retaining type safety without the
5476 use of additional functions. If DEBUG_SAFESTACK is not defined
5477 then the non typesafe macros are used instead. Also modified the
5478 mkstack.pl script to handle the new form. Needs testing to see
5479 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
5480 the default if no major problems. Similar behaviour for ASN1_SET_OF
5481 and PKCS12_STACK_OF.
5484 *) When some versions of IIS use the 'NET' form of private key the
5485 key derivation algorithm is different. Normally MD5(password) is
5486 used as a 128 bit RC4 key. In the modified case
5487 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
5488 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
5489 as the old Netscape_RSA functions except they have an additional
5490 'sgckey' parameter which uses the modified algorithm. Also added
5491 an -sgckey command line option to the rsa utility. Thanks to
5492 Adrian Peck <bertie@ncipher.com> for posting details of the modified
5493 algorithm to openssl-dev.
5496 *) The evp_local.h macros were using 'c.##kname' which resulted in
5497 invalid expansion on some systems (SCO 5.0.5 for example).
5498 Corrected to 'c.kname'.
5499 [Phillip Porch <root@theporch.com>]
5501 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
5502 a STACK of email addresses from a certificate or request, these look
5503 in the subject name and the subject alternative name extensions and
5504 omit any duplicate addresses.
5507 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
5508 This makes DSA verification about 2 % faster.
5511 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
5512 (meaning that now 2^5 values will be precomputed, which is only 4 KB
5513 plus overhead for 1024 bit moduli).
5514 This makes exponentiations about 0.5 % faster for 1024 bit
5515 exponents (as measured by "openssl speed rsa2048").
5518 *) Rename memory handling macros to avoid conflicts with other
5520 Malloc => OPENSSL_malloc
5521 Malloc_locked => OPENSSL_malloc_locked
5522 Realloc => OPENSSL_realloc
5523 Free => OPENSSL_free
5526 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
5527 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
5530 *) CygWin32 support.
5531 [John Jarvie <jjarvie@newsguy.com>]
5533 *) The type-safe stack code has been rejigged. It is now only compiled
5534 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
5535 by default all type-specific stack functions are "#define"d back to
5536 standard stack functions. This results in more streamlined output
5537 but retains the type-safety checking possibilities of the original
5541 *) The STACK code has been cleaned up, and certain type declarations
5542 that didn't make a lot of sense have been brought in line. This has
5543 also involved a cleanup of sorts in safestack.h to more correctly
5544 map type-safe stack functions onto their plain stack counterparts.
5545 This work has also resulted in a variety of "const"ifications of
5546 lots of the code, especially "_cmp" operations which should normally
5547 be prototyped with "const" parameters anyway.
5550 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
5551 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
5552 (The PRNG state consists of two parts, the large pool 'state' and 'md',
5553 where all of 'md' is used each time the PRNG is used, but 'state'
5554 is used only indexed by a cyclic counter. As entropy may not be
5555 well distributed from the beginning, 'md' is important as a
5556 chaining variable. However, the output function chains only half
5557 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
5558 all of 'md', and seeding with STATE_SIZE dummy bytes will result
5559 in all of 'state' being rewritten, with the new values depending
5560 on virtually all of 'md'. This overcomes the 80 bit limitation.)
5563 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
5564 the handshake is continued after ssl_verify_cert_chain();
5565 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
5566 can lead to 'unexplainable' connection aborts later.
5567 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
5569 *) Major EVP API cipher revision.
5570 Add hooks for extra EVP features. This allows various cipher
5571 parameters to be set in the EVP interface. Support added for variable
5572 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
5573 setting of RC2 and RC5 parameters.
5575 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
5578 Remove lots of duplicated code from the EVP library. For example *every*
5579 cipher init() function handles the 'iv' in the same way according to the
5580 cipher mode. They also all do nothing if the 'key' parameter is NULL and
5581 for CFB and OFB modes they zero ctx->num.
5583 New functionality allows removal of S/MIME code RC2 hack.
5585 Most of the routines have the same form and so can be declared in terms
5588 By shifting this to the top level EVP_CipherInit() it can be removed from
5589 all individual ciphers. If the cipher wants to handle IVs or keys
5590 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
5593 Change lots of functions like EVP_EncryptUpdate() to now return a
5594 value: although software versions of the algorithms cannot fail
5595 any installed hardware versions can.
5598 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
5599 this option is set, tolerate broken clients that send the negotiated
5600 protocol version number instead of the requested protocol version
5604 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
5605 i.e. non-zero for export ciphersuites, zero otherwise.
5606 Previous versions had this flag inverted, inconsistent with
5607 rsa_tmp_cb (..._TMP_RSA_CB).
5608 [Bodo Moeller; problem reported by Amit Chopra]
5610 *) Add missing DSA library text string. Work around for some IIS
5611 key files with invalid SEQUENCE encoding.
5614 *) Add a document (doc/standards.txt) that list all kinds of standards
5615 and so on that are implemented in OpenSSL.
5618 *) Enhance c_rehash script. Old version would mishandle certificates
5619 with the same subject name hash and wouldn't handle CRLs at all.
5620 Added -fingerprint option to crl utility, to support new c_rehash
5624 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
5627 *) Fix for SSL server purpose checking. Server checking was
5628 rejecting certificates which had extended key usage present
5629 but no ssl client purpose.
5630 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
5632 *) Make PKCS#12 code work with no password. The PKCS#12 spec
5633 is a little unclear about how a blank password is handled.
5634 Since the password in encoded as a BMPString with terminating
5635 double NULL a zero length password would end up as just the
5636 double NULL. However no password at all is different and is
5637 handled differently in the PKCS#12 key generation code. NS
5638 treats a blank password as zero length. MSIE treats it as no
5639 password on export: but it will try both on import. We now do
5640 the same: PKCS12_parse() tries zero length and no password if
5641 the password is set to "" or NULL (NULL is now a valid password:
5642 it wasn't before) as does the pkcs12 application.
5645 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
5646 perror when PEM_read_bio_X509_REQ fails, the error message must
5647 be obtained from the error queue.
5650 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
5651 it in ERR_remove_state if appropriate, and change ERR_get_state
5652 accordingly to avoid race conditions (this is necessary because
5653 thread_hash is no longer constant once set).
5656 *) Bugfix for linux-elf makefile.one.
5659 *) RSA_get_default_method() will now cause a default
5660 RSA_METHOD to be chosen if one doesn't exist already.
5661 Previously this was only set during a call to RSA_new()
5662 or RSA_new_method(NULL) meaning it was possible for
5663 RSA_get_default_method() to return NULL.
5666 *) Added native name translation to the existing DSO code
5667 that will convert (if the flag to do so is set) filenames
5668 that are sufficiently small and have no path information
5669 into a canonical native form. Eg. "blah" converted to
5670 "libblah.so" or "blah.dll" etc.
5673 *) New function ERR_error_string_n(e, buf, len) which is like
5674 ERR_error_string(e, buf), but writes at most 'len' bytes
5675 including the 0 terminator. For ERR_error_string_n, 'buf'
5677 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
5679 *) CONF library reworked to become more general. A new CONF
5680 configuration file reader "class" is implemented as well as a
5681 new functions (NCONF_*, for "New CONF") to handle it. The now
5682 old CONF_* functions are still there, but are reimplemented to
5683 work in terms of the new functions. Also, a set of functions
5684 to handle the internal storage of the configuration data is
5685 provided to make it easier to write new configuration file
5686 reader "classes" (I can definitely see something reading a
5687 configuration file in XML format, for example), called _CONF_*,
5688 or "the configuration storage API"...
5690 The new configuration file reading functions are:
5692 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
5693 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
5695 NCONF_default, NCONF_WIN32
5697 NCONF_dump_fp, NCONF_dump_bio
5699 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
5700 NCONF_new creates a new CONF object. This works in the same way
5701 as other interfaces in OpenSSL, like the BIO interface.
5702 NCONF_dump_* dump the internal storage of the configuration file,
5703 which is useful for debugging. All other functions take the same
5704 arguments as the old CONF_* functions wth the exception of the
5705 first that must be a `CONF *' instead of a `LHASH *'.
5707 To make it easer to use the new classes with the old CONF_* functions,
5708 the function CONF_set_default_method is provided.
5711 *) Add '-tls1' option to 'openssl ciphers', which was already
5712 mentioned in the documentation but had not been implemented.
5713 (This option is not yet really useful because even the additional
5714 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
5717 *) Initial DSO code added into libcrypto for letting OpenSSL (and
5718 OpenSSL-based applications) load shared libraries and bind to
5719 them in a portable way.
5720 [Geoff Thorpe, with contributions from Richard Levitte]
5722 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
5724 *) Make sure _lrotl and _lrotr are only used with MSVC.
5726 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
5727 (the default implementation of RAND_status).
5729 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
5730 to '-clrext' (= clear extensions), as intended and documented.
5731 [Bodo Moeller; inconsistency pointed out by Michael Attili
5734 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
5735 was larger than the MD block size.
5736 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
5738 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
5739 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
5740 using the passed key: if the passed key was a private key the result
5741 of X509_print(), for example, would be to print out all the private key
5745 *) des_quad_cksum() byte order bug fix.
5746 [Ulf Möller, using the problem description in krb4-0.9.7, where
5747 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
5749 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
5751 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
5753 *) For easily testing in shell scripts whether some command
5754 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
5755 returns with exit code 0 iff no command of the given name is available.
5756 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
5757 the output goes to stdout and nothing is printed to stderr.
5758 Additional arguments are always ignored.
5760 Since for each cipher there is a command of the same name,
5761 the 'no-cipher' compilation switches can be tested this way.
5763 ('openssl no-XXX' is not able to detect pseudo-commands such
5764 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
5767 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
5770 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
5771 is set; it will be thrown away anyway because each handshake creates
5773 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
5774 to parameters -- in previous versions (since OpenSSL 0.9.3) the
5775 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
5776 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
5779 *) New s_client option -ign_eof: EOF at stdin is ignored, and
5780 'Q' and 'R' lose their special meanings (quit/renegotiate).
5781 This is part of what -quiet does; unlike -quiet, -ign_eof
5782 does not suppress any output.
5785 *) Add compatibility options to the purpose and trust code. The
5786 purpose X509_PURPOSE_ANY is "any purpose" which automatically
5787 accepts a certificate or CA, this was the previous behaviour,
5788 with all the associated security issues.
5790 X509_TRUST_COMPAT is the old trust behaviour: only and
5791 automatically trust self signed roots in certificate store. A
5792 new trust setting X509_TRUST_DEFAULT is used to specify that
5793 a purpose has no associated trust setting and it should instead
5794 use the value in the default purpose.
5797 *) Fix the PKCS#8 DSA private key code so it decodes keys again
5798 and fix a memory leak.
5801 *) In util/mkerr.pl (which implements 'make errors'), preserve
5802 reason strings from the previous version of the .c file, as
5803 the default to have only downcase letters (and digits) in
5804 automatically generated reasons codes is not always appropriate.
5807 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
5808 using strerror. Previously, ERR_reason_error_string() returned
5809 library names as reason strings for SYSerr; but SYSerr is a special
5810 case where small numbers are errno values, not library numbers.
5813 *) Add '-dsaparam' option to 'openssl dhparam' application. This
5814 converts DSA parameters into DH parameters. (When creating parameters,
5815 DSA_generate_parameters is used.)
5818 *) Include 'length' (recommended exponent length) in C code generated
5819 by 'openssl dhparam -C'.
5822 *) The second argument to set_label in perlasm was already being used
5823 so couldn't be used as a "file scope" flag. Moved to third argument
5827 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
5828 instead of RAND_bytes for encryption IVs and salts.
5831 *) Include RAND_status() into RAND_METHOD instead of implementing
5832 it only for md_rand.c Otherwise replacing the PRNG by calling
5833 RAND_set_rand_method would be impossible.
5836 *) Don't let DSA_generate_key() enter an infinite loop if the random
5837 number generation fails.
5840 *) New 'rand' application for creating pseudo-random output.
5843 *) Added configuration support for Linux/IA64
5844 [Rolf Haberrecker <rolf@suse.de>]
5846 *) Assembler module support for Mingw32.
5849 *) Shared library support for HPUX (in shlib/).
5850 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
5852 *) Shared library support for Solaris gcc.
5853 [Lutz Behnke <behnke@trustcenter.de>]
5855 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
5857 *) PKCS7_encrypt() was adding text MIME headers twice because they
5858 were added manually and by SMIME_crlf_copy().
5861 *) In bntest.c don't call BN_rand with zero bits argument.
5862 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
5864 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
5865 case was implemented. This caused BN_div_recp() to fail occasionally.
5868 *) Add an optional second argument to the set_label() in the perl
5869 assembly language builder. If this argument exists and is set
5870 to 1 it signals that the assembler should use a symbol whose
5871 scope is the entire file, not just the current function. This
5872 is needed with MASM which uses the format label:: for this scope.
5873 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
5875 *) Change the ASN1 types so they are typedefs by default. Before
5876 almost all types were #define'd to ASN1_STRING which was causing
5877 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
5881 *) Change names of new functions to the new get1/get0 naming
5882 convention: After 'get1', the caller owns a reference count
5883 and has to call ..._free; 'get0' returns a pointer to some
5884 data structure without incrementing reference counters.
5885 (Some of the existing 'get' functions increment a reference
5886 counter, some don't.)
5887 Similarly, 'set1' and 'add1' functions increase reference
5888 counters or duplicate objects.
5891 *) Allow for the possibility of temp RSA key generation failure:
5892 the code used to assume it always worked and crashed on failure.
5895 *) Fix potential buffer overrun problem in BIO_printf().
5896 [Ulf Möller, using public domain code by Patrick Powell; problem
5897 pointed out by David Sacerdote <das33@cornell.edu>]
5899 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
5900 RAND_egd() and RAND_status(). In the command line application,
5901 the EGD socket can be specified like a seed file using RANDFILE
5905 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
5906 Some CAs (e.g. Verisign) distribute certificates in this form.
5909 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
5910 list to exclude them. This means that no special compilation option
5911 is needed to use anonymous DH: it just needs to be included in the
5915 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
5916 EVP_MD_type. The old functionality is available in a new macro called
5917 EVP_MD_md(). Change code that uses it and update docs.
5920 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
5921 where the 'void *' argument is replaced by a function pointer argument.
5922 Previously 'void *' was abused to point to functions, which works on
5923 many platforms, but is not correct. As these functions are usually
5924 called by macros defined in OpenSSL header files, most source code
5925 should work without changes.
5928 *) <openssl/opensslconf.h> (which is created by Configure) now contains
5929 sections with information on -D... compiler switches used for
5930 compiling the library so that applications can see them. To enable
5931 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
5932 must be defined. E.g.,
5933 #define OPENSSL_ALGORITHM_DEFINES
5934 #include <openssl/opensslconf.h>
5935 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
5936 [Richard Levitte, Ulf and Bodo Möller]
5938 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
5942 *) Change the 'other' type in certificate aux info to a STACK_OF
5943 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
5944 the required ASN1 format: arbitrary types determined by an OID.
5947 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
5948 argument to 'req'. This is not because the function is newer or
5949 better than others it just uses the work 'NEW' in the certificate
5950 request header lines. Some software needs this.
5953 *) Reorganise password command line arguments: now passwords can be
5954 obtained from various sources. Delete the PEM_cb function and make
5955 it the default behaviour: i.e. if the callback is NULL and the
5956 usrdata argument is not NULL interpret it as a null terminated pass
5957 phrase. If usrdata and the callback are NULL then the pass phrase
5958 is prompted for as usual.
5961 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
5962 the support is automatically enabled. The resulting binaries will
5963 autodetect the card and use it if present.
5964 [Ben Laurie and Compaq Inc.]
5966 *) Work around for Netscape hang bug. This sends certificate request
5967 and server done in one record. Since this is perfectly legal in the
5968 SSL/TLS protocol it isn't a "bug" option and is on by default. See
5969 the bugs/SSLv3 entry for more info.
5972 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
5975 *) Add -rand argument to smime and pkcs12 applications and read/write
5979 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
5982 *) Add command line password options to the remaining applications.
5985 *) Bug fix for BN_div_recp() for numerators with an even number of
5989 *) More tests in bntest.c, and changed test_bn output.
5992 *) ./config recognizes MacOS X now.
5995 *) Bug fix for BN_div() when the first words of num and divsor are
5996 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
5999 *) Add support for various broken PKCS#8 formats, and command line
6000 options to produce them.
6003 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
6004 get temporary BIGNUMs from a BN_CTX.
6007 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
6011 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
6012 include a #define from the old name to the new. The original intent
6013 was that statically linked binaries could for example just call
6014 SSLeay_add_all_ciphers() to just add ciphers to the table and not
6015 link with digests. This never worked becayse SSLeay_add_all_digests()
6016 and SSLeay_add_all_ciphers() were in the same source file so calling
6017 one would link with the other. They are now in separate source files.
6020 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
6023 *) Use a less unusual form of the Miller-Rabin primality test (it used
6024 a binary algorithm for exponentiation integrated into the Miller-Rabin
6025 loop, our standard modexp algorithms are faster).
6028 *) Support for the EBCDIC character set completed.
6029 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
6031 *) Source code cleanups: use const where appropriate, eliminate casts,
6032 use void * instead of char * in lhash.
6035 *) Bugfix: ssl3_send_server_key_exchange was not restartable
6036 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
6037 this the server could overwrite ephemeral keys that the client
6041 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
6042 using 50 iterations of the Rabin-Miller test.
6044 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
6045 iterations of the Rabin-Miller test as required by the appendix
6046 to FIPS PUB 186[-1]) instead of DSA_is_prime.
6047 As BN_is_prime_fasttest includes trial division, DSA parameter
6048 generation becomes much faster.
6050 This implies a change for the callback functions in DSA_is_prime
6051 and DSA_generate_parameters: The callback function is called once
6052 for each positive witness in the Rabin-Miller test, not just
6053 occasionally in the inner loop; and the parameters to the
6054 callback function now provide an iteration count for the outer
6055 loop rather than for the current invocation of the inner loop.
6056 DSA_generate_parameters additionally can call the callback
6057 function with an 'iteration count' of -1, meaning that a
6058 candidate has passed the trial division test (when q is generated
6059 from an application-provided seed, trial division is skipped).
6062 *) New function BN_is_prime_fasttest that optionally does trial
6063 division before starting the Rabin-Miller test and has
6064 an additional BN_CTX * argument (whereas BN_is_prime always
6065 has to allocate at least one BN_CTX).
6066 'callback(1, -1, cb_arg)' is called when a number has passed the
6067 trial division stage.
6070 *) Fix for bug in CRL encoding. The validity dates weren't being handled
6074 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
6077 *) New function BN_pseudo_rand().
6080 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
6081 bignum version of BN_from_montgomery() with the working code from
6082 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
6086 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
6087 made it impossible to use the same SSL_SESSION data structure in
6088 SSL2 clients in multiple threads.
6091 *) The return value of RAND_load_file() no longer counts bytes obtained
6092 by stat(). RAND_load_file(..., -1) is new and uses the complete file
6093 to seed the PRNG (previously an explicit byte count was required).
6094 [Ulf Möller, Bodo Möller]
6096 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
6097 used (char *) instead of (void *) and had casts all over the place.
6100 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
6103 *) Retain source code compatibility for BN_prime_checks macro:
6104 BN_is_prime(..., BN_prime_checks, ...) now uses
6105 BN_prime_checks_for_size to determine the appropriate number of
6106 Rabin-Miller iterations.
6109 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
6110 DH_CHECK_P_NOT_SAFE_PRIME.
6111 (Check if this is true? OpenPGP calls them "strong".)
6114 *) Merge the functionality of "dh" and "gendh" programs into a new program
6115 "dhparam". The old programs are retained for now but will handle DH keys
6116 (instead of parameters) in future.
6119 *) Make the ciphers, s_server and s_client programs check the return values
6120 when a new cipher list is set.
6123 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
6124 ciphers. Before when the 56bit ciphers were enabled the sorting was
6127 The syntax for the cipher sorting has been extended to support sorting by
6128 cipher-strength (using the strength_bits hard coded in the tables).
6129 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
6131 Fix a bug in the cipher-command parser: when supplying a cipher command
6132 string with an "undefined" symbol (neither command nor alphanumeric
6133 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
6134 an error is flagged.
6136 Due to the strength-sorting extension, the code of the
6137 ssl_create_cipher_list() function was completely rearranged. I hope that
6138 the readability was also increased :-)
6139 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6141 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
6142 for the first serial number and places 2 in the serial number file. This
6143 avoids problems when the root CA is created with serial number zero and
6144 the first user certificate has the same issuer name and serial number
6148 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
6149 the new code. Add documentation for this stuff.
6152 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
6153 X509_*() to X509at_*() on the grounds that they don't handle X509
6154 structures and behave in an analagous way to the X509v3 functions:
6155 they shouldn't be called directly but wrapper functions should be used
6158 So we also now have some wrapper functions that call the X509at functions
6159 when passed certificate requests. (TO DO: similar things can be done with
6160 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
6161 things. Some of these need some d2i or i2d and print functionality
6162 because they handle more complex structures.)
6165 *) Add missing #ifndefs that caused missing symbols when building libssl
6166 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
6167 NO_RSA in ssl/s2*.c.
6168 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
6170 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
6171 has a return value which indicates the quality of the random data
6172 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
6173 error queue. New function RAND_pseudo_bytes() generates output that is
6174 guaranteed to be unique but not unpredictable. RAND_add is like
6175 RAND_seed, but takes an extra argument for an entropy estimate
6176 (RAND_seed always assumes full entropy).
6179 *) Do more iterations of Rabin-Miller probable prime test (specifically,
6180 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
6181 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
6182 in crypto/bn/bn_prime.c for the complete table). This guarantees a
6183 false-positive rate of at most 2^-80 for random input.
6186 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
6189 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
6190 in the 0.9.5 release), this returns the chain
6191 from an X509_CTX structure with a dup of the stack and all
6192 the X509 reference counts upped: so the stack will exist
6193 after X509_CTX_cleanup() has been called. Modify pkcs12.c
6196 Also make SSL_SESSION_print() print out the verify return
6200 *) Add manpage for the pkcs12 command. Also change the default
6201 behaviour so MAC iteration counts are used unless the new
6202 -nomaciter option is used. This improves file security and
6203 only older versions of MSIE (4.0 for example) need it.
6206 *) Honor the no-xxx Configure options when creating .DEF files.
6209 *) Add PKCS#10 attributes to field table: challengePassword,
6210 unstructuredName and unstructuredAddress. These are taken from
6211 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
6212 international characters are used.
6214 More changes to X509_ATTRIBUTE code: allow the setting of types
6215 based on strings. Remove the 'loc' parameter when adding
6216 attributes because these will be a SET OF encoding which is sorted
6220 *) Initial changes to the 'req' utility to allow request generation
6221 automation. This will allow an application to just generate a template
6222 file containing all the field values and have req construct the
6225 Initial support for X509_ATTRIBUTE handling. Stacks of these are
6226 used all over the place including certificate requests and PKCS#7
6227 structures. They are currently handled manually where necessary with
6228 some primitive wrappers for PKCS#7. The new functions behave in a
6229 manner analogous to the X509 extension functions: they allow
6230 attributes to be looked up by NID and added.
6232 Later something similar to the X509V3 code would be desirable to
6233 automatically handle the encoding, decoding and printing of the
6234 more complex types. The string types like challengePassword can
6235 be handled by the string table functions.
6237 Also modified the multi byte string table handling. Now there is
6238 a 'global mask' which masks out certain types. The table itself
6239 can use the flag STABLE_NO_MASK to ignore the mask setting: this
6240 is useful when for example there is only one permissible type
6241 (as in countryName) and using the mask might result in no valid
6245 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
6246 SSL_get_peer_finished to allow applications to obtain the latest
6247 Finished messages sent to the peer or expected from the peer,
6248 respectively. (SSL_get_peer_finished is usually the Finished message
6249 actually received from the peer, otherwise the protocol will be aborted.)
6251 As the Finished message are message digests of the complete handshake
6252 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
6253 be used for external authentication procedures when the authentication
6254 provided by SSL/TLS is not desired or is not enough.
6257 *) Enhanced support for Alpha Linux is added. Now ./config checks if
6258 the host supports BWX extension and if Compaq C is present on the
6259 $PATH. Just exploiting of the BWX extension results in 20-30%
6260 performance kick for some algorithms, e.g. DES and RC4 to mention
6261 a couple. Compaq C in turn generates ~20% faster code for MD5 and
6265 *) Add support for MS "fast SGC". This is arguably a violation of the
6266 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
6267 weak crypto and after checking the certificate is SGC a second one
6268 with strong crypto. MS SGC stops the first handshake after receiving
6269 the server certificate message and sends a second client hello. Since
6270 a server will typically do all the time consuming operations before
6271 expecting any further messages from the client (server key exchange
6272 is the most expensive) there is little difference between the two.
6274 To get OpenSSL to support MS SGC we have to permit a second client
6275 hello message after we have sent server done. In addition we have to
6276 reset the MAC if we do get this second client hello.
6279 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
6280 if a DER encoded private key is RSA or DSA traditional format. Changed
6281 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
6282 format DER encoded private key. Newer code should use PKCS#8 format which
6283 has the key type encoded in the ASN1 structure. Added DER private key
6284 support to pkcs8 application.
6287 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
6288 ciphersuites has been selected (as required by the SSL 3/TLS 1
6289 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
6290 is set, we interpret this as a request to violate the specification
6291 (the worst that can happen is a handshake failure, and 'correct'
6292 behaviour would result in a handshake failure anyway).
6295 *) In SSL_CTX_add_session, take into account that there might be multiple
6296 SSL_SESSION structures with the same session ID (e.g. when two threads
6297 concurrently obtain them from an external cache).
6298 The internal cache can handle only one SSL_SESSION with a given ID,
6299 so if there's a conflict, we now throw out the old one to achieve
6303 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
6304 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
6305 some routines that use cipher OIDs: some ciphers do not have OIDs
6306 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
6310 *) Simplify the trust setting structure and code. Now we just have
6311 two sequences of OIDs for trusted and rejected settings. These will
6312 typically have values the same as the extended key usage extension
6313 and any application specific purposes.
6315 The trust checking code now has a default behaviour: it will just
6316 check for an object with the same NID as the passed id. Functions can
6317 be provided to override either the default behaviour or the behaviour
6318 for a given id. SSL client, server and email already have functions
6319 in place for compatibility: they check the NID and also return "trusted"
6320 if the certificate is self signed.
6323 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
6324 traditional format into an EVP_PKEY structure.
6327 *) Add a password callback function PEM_cb() which either prompts for
6328 a password if usr_data is NULL or otherwise assumes it is a null
6329 terminated password. Allow passwords to be passed on command line
6330 environment or config files in a few more utilities.
6333 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
6334 keys. Add some short names for PKCS#8 PBE algorithms and allow them
6335 to be specified on the command line for the pkcs8 and pkcs12 utilities.
6336 Update documentation.
6339 *) Support for ASN1 "NULL" type. This could be handled before by using
6340 ASN1_TYPE but there wasn't any function that would try to read a NULL
6341 and produce an error if it couldn't. For compatibility we also have
6342 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
6343 don't allocate anything because they don't need to.
6346 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
6348 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
6350 *) Rebuild of the memory allocation routines used by OpenSSL code and
6351 possibly others as well. The purpose is to make an interface that
6352 provide hooks so anyone can build a separate set of allocation and
6353 deallocation routines to be used by OpenSSL, for example memory
6354 pool implementations, or something else, which was previously hard
6355 since Malloc(), Realloc() and Free() were defined as macros having
6356 the values malloc, realloc and free, respectively (except for Win32
6357 compilations). The same is provided for memory debugging code.
6358 OpenSSL already comes with functionality to find memory leaks, but
6359 this gives people a chance to debug other memory problems.
6361 With these changes, a new set of functions and macros have appeared:
6363 CRYPTO_set_mem_debug_functions() [F]
6364 CRYPTO_get_mem_debug_functions() [F]
6365 CRYPTO_dbg_set_options() [F]
6366 CRYPTO_dbg_get_options() [F]
6367 CRYPTO_malloc_debug_init() [M]
6369 The memory debug functions are NULL by default, unless the library
6370 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
6371 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
6372 gives the standard debugging functions that come with OpenSSL) or
6373 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
6374 provided by the library user) must be used. When the standard
6375 debugging functions are used, CRYPTO_dbg_set_options can be used to
6376 request additional information:
6377 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
6378 the CRYPTO_MDEBUG_xxx macro when compiling the library.
6380 Also, things like CRYPTO_set_mem_functions will always give the
6381 expected result (the new set of functions is used for allocation
6382 and deallocation) at all times, regardless of platform and compiler
6385 To finish it up, some functions that were never use in any other
6386 way than through macros have a new API and new semantic:
6389 CRYPTO_dbg_realloc()
6392 All macros of value have retained their old syntax.
6393 [Richard Levitte and Bodo Moeller]
6395 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
6396 ordering of SMIMECapabilities wasn't in "strength order" and there
6397 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
6401 *) Some ASN1 types with illegal zero length encoding (INTEGER,
6402 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
6403 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
6405 *) Merge in my S/MIME library for OpenSSL. This provides a simple
6406 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
6407 functionality to handle multipart/signed properly) and a utility
6408 called 'smime' to call all this stuff. This is based on code I
6409 originally wrote for Celo who have kindly allowed it to be
6410 included in OpenSSL.
6413 *) Add variants des_set_key_checked and des_set_key_unchecked of
6414 des_set_key (aka des_key_sched). Global variable des_check_key
6415 decides which of these is called by des_set_key; this way
6416 des_check_key behaves as it always did, but applications and
6417 the library itself, which was buggy for des_check_key == 1,
6418 have a cleaner way to pick the version they need.
6421 *) New function PKCS12_newpass() which changes the password of a
6425 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
6426 dynamic mix. In both cases the ids can be used as an index into the
6427 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
6428 functions so they accept a list of the field values and the
6429 application doesn't need to directly manipulate the X509_TRUST
6433 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
6437 *) Modify the way the V3 extension code looks up extensions. This now
6438 works in a similar way to the object code: we have some "standard"
6439 extensions in a static table which is searched with OBJ_bsearch()
6440 and the application can add dynamic ones if needed. The file
6441 crypto/x509v3/ext_dat.h now has the info: this file needs to be
6442 updated whenever a new extension is added to the core code and kept
6443 in ext_nid order. There is a simple program 'tabtest.c' which checks
6444 this. New extensions are not added too often so this file can readily
6445 be maintained manually.
6447 There are two big advantages in doing things this way. The extensions
6448 can be looked up immediately and no longer need to be "added" using
6449 X509V3_add_standard_extensions(): this function now does nothing.
6450 [Side note: I get *lots* of email saying the extension code doesn't
6451 work because people forget to call this function]
6452 Also no dynamic allocation is done unless new extensions are added:
6453 so if we don't add custom extensions there is no need to call
6454 X509V3_EXT_cleanup().
6457 *) Modify enc utility's salting as follows: make salting the default. Add a
6458 magic header, so unsalted files fail gracefully instead of just decrypting
6459 to garbage. This is because not salting is a big security hole, so people
6460 should be discouraged from doing it.
6463 *) Fixes and enhancements to the 'x509' utility. It allowed a message
6464 digest to be passed on the command line but it only used this
6465 parameter when signing a certificate. Modified so all relevant
6466 operations are affected by the digest parameter including the
6467 -fingerprint and -x509toreq options. Also -x509toreq choked if a
6468 DSA key was used because it didn't fix the digest.
6471 *) Initial certificate chain verify code. Currently tests the untrusted
6472 certificates for consistency with the verify purpose (which is set
6473 when the X509_STORE_CTX structure is set up) and checks the pathlength.
6475 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
6476 this is because it will reject chains with invalid extensions whereas
6477 every previous version of OpenSSL and SSLeay made no checks at all.
6479 Trust code: checks the root CA for the relevant trust settings. Trust
6480 settings have an initial value consistent with the verify purpose: e.g.
6481 if the verify purpose is for SSL client use it expects the CA to be
6482 trusted for SSL client use. However the default value can be changed to
6483 permit custom trust settings: one example of this would be to only trust
6484 certificates from a specific "secure" set of CAs.
6486 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
6487 which should be used for version portability: especially since the
6488 verify structure is likely to change more often now.
6490 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
6491 to set them. If not set then assume SSL clients will verify SSL servers
6494 Two new options to the verify program: -untrusted allows a set of
6495 untrusted certificates to be passed in and -purpose which sets the
6496 intended purpose of the certificate. If a purpose is set then the
6497 new chain verify code is used to check extension consistency.
6500 *) Support for the authority information access extension.
6503 *) Modify RSA and DSA PEM read routines to transparently handle
6504 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
6505 public keys in a format compatible with certificate
6506 SubjectPublicKeyInfo structures. Unfortunately there were already
6507 functions called *_PublicKey_* which used various odd formats so
6508 these are retained for compatibility: however the DSA variants were
6509 never in a public release so they have been deleted. Changed dsa/rsa
6510 utilities to handle the new format: note no releases ever handled public
6511 keys so we should be OK.
6513 The primary motivation for this change is to avoid the same fiasco
6514 that dogs private keys: there are several incompatible private key
6515 formats some of which are standard and some OpenSSL specific and
6516 require various evil hacks to allow partial transparent handling and
6517 even then it doesn't work with DER formats. Given the option anything
6518 other than PKCS#8 should be dumped: but the other formats have to
6519 stay in the name of compatibility.
6521 With public keys and the benefit of hindsight one standard format
6522 is used which works with EVP_PKEY, RSA or DSA structures: though
6523 it clearly returns an error if you try to read the wrong kind of key.
6525 Added a -pubkey option to the 'x509' utility to output the public key.
6526 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
6527 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
6528 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
6529 that do the same as the EVP_PKEY_assign_*() except they up the
6530 reference count of the added key (they don't "swallow" the
6534 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
6535 CRLs would fail if the file contained no certificates or no CRLs:
6536 added a new function to read in both types and return the number
6537 read: this means that if none are read it will be an error. The
6538 DER versions of the certificate and CRL reader would always fail
6539 because it isn't possible to mix certificates and CRLs in DER format
6540 without choking one or the other routine. Changed this to just read
6541 a certificate: this is the best we can do. Also modified the code
6542 in apps/verify.c to take notice of return codes: it was previously
6543 attempting to read in certificates from NULL pointers and ignoring
6544 any errors: this is one reason why the cert and CRL reader seemed
6545 to work. It doesn't check return codes from the default certificate
6546 routines: these may well fail if the certificates aren't installed.
6549 *) Code to support otherName option in GeneralName.
6552 *) First update to verify code. Change the verify utility
6553 so it warns if it is passed a self signed certificate:
6554 for consistency with the normal behaviour. X509_verify
6555 has been modified to it will now verify a self signed
6556 certificate if *exactly* the same certificate appears
6557 in the store: it was previously impossible to trust a
6558 single self signed certificate. This means that:
6559 openssl verify ss.pem
6560 now gives a warning about a self signed certificate but
6561 openssl verify -CAfile ss.pem ss.pem
6565 *) For servers, store verify_result in SSL_SESSION data structure
6566 (and add it to external session representation).
6567 This is needed when client certificate verifications fails,
6568 but an application-provided verification callback (set by
6569 SSL_CTX_set_cert_verify_callback) allows accepting the session
6570 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
6571 but returns 1): When the session is reused, we have to set
6572 ssl->verify_result to the appropriate error code to avoid
6574 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
6576 *) Fix a bug in the new PKCS#7 code: it didn't consider the
6577 case in PKCS7_dataInit() where the signed PKCS7 structure
6578 didn't contain any existing data because it was being created.
6579 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
6581 *) Add a salt to the key derivation routines in enc.c. This
6582 forms the first 8 bytes of the encrypted file. Also add a
6583 -S option to allow a salt to be input on the command line.
6586 *) New function X509_cmp(). Oddly enough there wasn't a function
6587 to compare two certificates. We do this by working out the SHA1
6588 hash and comparing that. X509_cmp() will be needed by the trust
6592 *) SSL_get1_session() is like SSL_get_session(), but increments
6593 the reference count in the SSL_SESSION returned.
6594 [Geoff Thorpe <geoff@eu.c2.net>]
6596 *) Fix for 'req': it was adding a null to request attributes.
6597 Also change the X509_LOOKUP and X509_INFO code to handle
6598 certificate auxiliary information.
6601 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
6605 *) Add the possibility to add extra information to the memory leak
6606 detecting output, to form tracebacks, showing from where each
6607 allocation was originated: CRYPTO_push_info("constant string") adds
6608 the string plus current file name and line number to a per-thread
6609 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
6610 is like calling CYRPTO_pop_info() until the stack is empty.
6611 Also updated memory leak detection code to be multi-thread-safe.
6614 *) Add options -text and -noout to pkcs7 utility and delete the
6615 encryption options which never did anything. Update docs.
6618 *) Add options to some of the utilities to allow the pass phrase
6619 to be included on either the command line (not recommended on
6620 OSes like Unix) or read from the environment. Update the
6621 manpages and fix a few bugs.
6624 *) Add a few manpages for some of the openssl commands.
6627 *) Fix the -revoke option in ca. It was freeing up memory twice,
6628 leaking and not finding already revoked certificates.
6631 *) Extensive changes to support certificate auxiliary information.
6632 This involves the use of X509_CERT_AUX structure and X509_AUX
6633 functions. An X509_AUX function such as PEM_read_X509_AUX()
6634 can still read in a certificate file in the usual way but it
6635 will also read in any additional "auxiliary information". By
6636 doing things this way a fair degree of compatibility can be
6637 retained: existing certificates can have this information added
6638 using the new 'x509' options.
6640 Current auxiliary information includes an "alias" and some trust
6641 settings. The trust settings will ultimately be used in enhanced
6642 certificate chain verification routines: currently a certificate
6643 can only be trusted if it is self signed and then it is trusted
6647 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
6648 The problem was that one of the replacement routines had not been working
6649 since SSLeay releases. For now the offending routine has been replaced
6650 with non-optimised assembler. Even so, this now gives around 95%
6651 performance improvement for 1024 bit RSA signs.
6654 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
6655 handling. Most clients have the effective key size in bits equal to
6656 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
6657 A few however don't do this and instead use the size of the decrypted key
6658 to determine the RC2 key length and the AlgorithmIdentifier to determine
6659 the effective key length. In this case the effective key length can still
6660 be 40 bits but the key length can be 168 bits for example. This is fixed
6661 by manually forcing an RC2 key into the EVP_PKEY structure because the
6662 EVP code can't currently handle unusual RC2 key sizes: it always assumes
6663 the key length and effective key length are equal.
6666 *) Add a bunch of functions that should simplify the creation of
6667 X509_NAME structures. Now you should be able to do:
6668 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
6669 and have it automatically work out the correct field type and fill in
6670 the structures. The more adventurous can try:
6671 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
6672 and it will (hopefully) work out the correct multibyte encoding.
6675 *) Change the 'req' utility to use the new field handling and multibyte
6676 copy routines. Before the DN field creation was handled in an ad hoc
6677 way in req, ca, and x509 which was rather broken and didn't support
6678 BMPStrings or UTF8Strings. Since some software doesn't implement
6679 BMPStrings or UTF8Strings yet, they can be enabled using the config file
6680 using the dirstring_type option. See the new comment in the default
6681 openssl.cnf for more info.
6684 *) Make crypto/rand/md_rand.c more robust:
6685 - Assure unique random numbers after fork().
6686 - Make sure that concurrent threads access the global counter and
6687 md serializably so that we never lose entropy in them
6688 or use exactly the same state in multiple threads.
6689 Access to the large state is not always serializable because
6690 the additional locking could be a performance killer, and
6691 md should be large enough anyway.
6694 *) New file apps/app_rand.c with commonly needed functionality
6695 for handling the random seed file.
6697 Use the random seed file in some applications that previously did not:
6699 dsaparam -genkey (which also ignored its '-rand' option),
6702 x509 (when signing).
6703 Except on systems with /dev/urandom, it is crucial to have a random
6704 seed file at least for key creation, DSA signing, and for DH exchanges;
6705 for RSA signatures we could do without one.
6707 gendh and gendsa (unlike genrsa) used to read only the first byte
6708 of each file listed in the '-rand' option. The function as previously
6709 found in genrsa is now in app_rand.c and is used by all programs
6710 that support '-rand'.
6713 *) In RAND_write_file, use mode 0600 for creating files;
6714 don't just chmod when it may be too late.
6717 *) Report an error from X509_STORE_load_locations
6718 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
6721 *) New function ASN1_mbstring_copy() this copies a string in either
6722 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
6723 into an ASN1_STRING type. A mask of permissible types is passed
6724 and it chooses the "minimal" type to use or an error if not type
6728 *) Add function equivalents to the various macros in asn1.h. The old
6729 macros are retained with an M_ prefix. Code inside the library can
6730 use the M_ macros. External code (including the openssl utility)
6731 should *NOT* in order to be "shared library friendly".
6734 *) Add various functions that can check a certificate's extensions
6735 to see if it usable for various purposes such as SSL client,
6736 server or S/MIME and CAs of these types. This is currently
6737 VERY EXPERIMENTAL but will ultimately be used for certificate chain
6738 verification. Also added a -purpose flag to x509 utility to
6739 print out all the purposes.
6742 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
6746 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
6747 for, obtain and decode and extension and obtain its critical flag.
6748 This allows all the necessary extension code to be handled in a
6749 single function call.
6752 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
6753 platforms. See crypto/rc4/rc4_enc.c for further details.
6756 *) New -noout option to asn1parse. This causes no output to be produced
6757 its main use is when combined with -strparse and -out to extract data
6758 from a file (which may not be in ASN.1 format).
6761 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
6762 when producing the local key id.
6763 [Richard Levitte <levitte@stacken.kth.se>]
6765 *) New option -dhparam in s_server. This allows a DH parameter file to be
6766 stated explicitly. If it is not stated then it tries the first server
6767 certificate file. The previous behaviour hard coded the filename
6771 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
6772 a public key to be input or output. For example:
6773 openssl rsa -in key.pem -pubout -out pubkey.pem
6774 Also added necessary DSA public key functions to handle this.
6777 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
6778 in the message. This was handled by allowing
6779 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
6780 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
6782 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
6783 to the end of the strings whereas this didn't. This would cause problems
6784 if strings read with d2i_ASN1_bytes() were later modified.
6785 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
6787 *) Fix for base64 decode bug. When a base64 bio reads only one line of
6788 data and it contains EOF it will end up returning an error. This is
6789 caused by input 46 bytes long. The cause is due to the way base64
6790 BIOs find the start of base64 encoded data. They do this by trying a
6791 trial decode on each line until they find one that works. When they
6792 do a flag is set and it starts again knowing it can pass all the
6793 data directly through the decoder. Unfortunately it doesn't reset
6794 the context it uses. This means that if EOF is reached an attempt
6795 is made to pass two EOFs through the context and this causes the
6796 resulting error. This can also cause other problems as well. As is
6797 usual with these problems it takes *ages* to find and the fix is
6798 trivial: move one line.
6799 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
6801 *) Ugly workaround to get s_client and s_server working under Windows. The
6802 old code wouldn't work because it needed to select() on sockets and the
6803 tty (for keypresses and to see if data could be written). Win32 only
6804 supports select() on sockets so we select() with a 1s timeout on the
6805 sockets and then see if any characters are waiting to be read, if none
6806 are present then we retry, we also assume we can always write data to
6807 the tty. This isn't nice because the code then blocks until we've
6808 received a complete line of data and it is effectively polling the
6809 keyboard at 1s intervals: however it's quite a bit better than not
6810 working at all :-) A dedicated Windows application might handle this
6811 with an event loop for example.
6814 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
6815 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
6816 will be called when RSA_sign() and RSA_verify() are used. This is useful
6817 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
6818 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
6819 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
6820 This necessitated the support of an extra signature type NID_md5_sha1
6821 for SSL signatures and modifications to the SSL library to use it instead
6822 of calling RSA_public_decrypt() and RSA_private_encrypt().
6825 *) Add new -verify -CAfile and -CApath options to the crl program, these
6826 will lookup a CRL issuers certificate and verify the signature in a
6827 similar way to the verify program. Tidy up the crl program so it
6828 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
6829 less strict. It will now permit CRL extensions even if it is not
6830 a V2 CRL: this will allow it to tolerate some broken CRLs.
6833 *) Initialize all non-automatic variables each time one of the openssl
6834 sub-programs is started (this is necessary as they may be started
6835 multiple times from the "OpenSSL>" prompt).
6836 [Lennart Bang, Bodo Moeller]
6838 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
6839 removing all other RSA functionality (this is what NO_RSA does). This
6840 is so (for example) those in the US can disable those operations covered
6841 by the RSA patent while allowing storage and parsing of RSA keys and RSA
6845 *) Non-copying interface to BIO pairs.
6846 (still largely untested)
6849 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
6850 ASCII string. This was handled independently in various places before.
6853 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
6854 UTF8 strings a character at a time.
6857 *) Use client_version from client hello to select the protocol
6858 (s23_srvr.c) and for RSA client key exchange verification
6859 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
6862 *) Add various utility functions to handle SPKACs, these were previously
6863 handled by poking round in the structure internals. Added new function
6864 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
6865 print, verify and generate SPKACs. Based on an original idea from
6866 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
6869 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
6872 *) Allow the config file extension section to be overwritten on the
6873 command line. Based on an original idea from Massimiliano Pala
6874 <madwolf@comune.modena.it>. The new option is called -extensions
6875 and can be applied to ca, req and x509. Also -reqexts to override
6876 the request extensions in req and -crlexts to override the crl extensions
6880 *) Add new feature to the SPKAC handling in ca. Now you can include
6881 the same field multiple times by preceding it by "XXXX." for example:
6884 this is the same syntax as used in the req config file.
6887 *) Allow certificate extensions to be added to certificate requests. These
6888 are specified in a 'req_extensions' option of the req section of the
6889 config file. They can be printed out with the -text option to req but
6890 are otherwise ignored at present.
6893 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
6894 data read consists of only the final block it would not decrypted because
6895 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
6896 A misplaced 'break' also meant the decrypted final block might not be
6897 copied until the next read.
6900 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
6901 a few extra parameters to the DH structure: these will be useful if
6902 for example we want the value of 'q' or implement X9.42 DH.
6905 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
6906 provides hooks that allow the default DSA functions or functions on a
6907 "per key" basis to be replaced. This allows hardware acceleration and
6908 hardware key storage to be handled without major modification to the
6909 library. Also added low level modexp hooks and CRYPTO_EX structure and
6910 associated functions.
6913 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
6914 as "read only": it can't be written to and the buffer it points to will
6915 not be freed. Reading from a read only BIO is much more efficient than
6916 a normal memory BIO. This was added because there are several times when
6917 an area of memory needs to be read from a BIO. The previous method was
6918 to create a memory BIO and write the data to it, this results in two
6919 copies of the data and an O(n^2) reading algorithm. There is a new
6920 function BIO_new_mem_buf() which creates a read only memory BIO from
6921 an area of memory. Also modified the PKCS#7 routines to use read only
6925 *) Bugfix: ssl23_get_client_hello did not work properly when called in
6926 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
6927 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
6928 but a retry condition occured while trying to read the rest.
6931 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
6932 NID_pkcs7_encrypted by default: this was wrong since this should almost
6933 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
6934 the encrypted data type: this is a more sensible place to put it and it
6935 allows the PKCS#12 code to be tidied up that duplicated this
6939 *) Changed obj_dat.pl script so it takes its input and output files on
6940 the command line. This should avoid shell escape redirection problems
6944 *) Initial support for certificate extension requests, these are included
6945 in things like Xenroll certificate requests. Included functions to allow
6946 extensions to be obtained and added.
6949 *) -crlf option to s_client and s_server for sending newlines as
6950 CRLF (as required by many protocols).
6953 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
6955 *) Install libRSAglue.a when OpenSSL is built with RSAref.
6956 [Ralf S. Engelschall]
6958 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
6959 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
6961 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
6965 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
6966 DH parameters/keys (q is lost during that conversion, but the resulting
6967 DH parameters contain its length).
6969 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
6970 much faster than DH_generate_parameters (which creates parameters
6971 where p = 2*q + 1), and also the smaller q makes DH computations
6972 much more efficient (160-bit exponentiation instead of 1024-bit
6973 exponentiation); so this provides a convenient way to support DHE
6974 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
6975 utter importance to use
6976 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
6978 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
6979 when such DH parameters are used, because otherwise small subgroup
6980 attacks may become possible!
6983 *) Avoid memory leak in i2d_DHparams.
6986 *) Allow the -k option to be used more than once in the enc program:
6987 this allows the same encrypted message to be read by multiple recipients.
6990 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
6991 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
6992 it will always use the numerical form of the OID, even if it has a short
6996 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
6997 method only got called if p,q,dmp1,dmq1,iqmp components were present,
6998 otherwise bn_mod_exp was called. In the case of hardware keys for example
6999 no private key components need be present and it might store extra data
7000 in the RSA structure, which cannot be accessed from bn_mod_exp.
7001 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
7002 private key operations.
7005 *) Added support for SPARC Linux.
7008 *) pem_password_cb function type incompatibly changed from
7009 typedef int pem_password_cb(char *buf, int size, int rwflag);
7011 ....(char *buf, int size, int rwflag, void *userdata);
7012 so that applications can pass data to their callbacks:
7013 The PEM[_ASN1]_{read,write}... functions and macros now take an
7014 additional void * argument, which is just handed through whenever
7015 the password callback is called.
7016 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
7018 New function SSL_CTX_set_default_passwd_cb_userdata.
7020 Compatibility note: As many C implementations push function arguments
7021 onto the stack in reverse order, the new library version is likely to
7022 interoperate with programs that have been compiled with the old
7023 pem_password_cb definition (PEM_whatever takes some data that
7024 happens to be on the stack as its last argument, and the callback
7025 just ignores this garbage); but there is no guarantee whatsoever that
7028 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
7029 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
7030 problems not only on Windows, but also on some Unix platforms.
7031 To avoid problematic command lines, these definitions are now in an
7032 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
7033 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
7036 *) MIPS III/IV assembler module is reimplemented.
7039 *) More DES library cleanups: remove references to srand/rand and
7040 delete an unused file.
7043 *) Add support for the the free Netwide assembler (NASM) under Win32,
7044 since not many people have MASM (ml) and it can be hard to obtain.
7045 This is currently experimental but it seems to work OK and pass all
7046 the tests. Check out INSTALL.W32 for info.
7049 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
7050 without temporary keys kept an extra copy of the server key,
7051 and connections with temporary keys did not free everything in case
7055 *) New function RSA_check_key and new openssl rsa option -check
7056 for verifying the consistency of RSA keys.
7057 [Ulf Moeller, Bodo Moeller]
7059 *) Various changes to make Win32 compile work:
7060 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
7061 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
7062 comparison" warnings.
7063 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
7066 *) Add a debugging option to PKCS#5 v2 key generation function: when
7067 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
7068 derived keys are printed to stderr.
7071 *) Copy the flags in ASN1_STRING_dup().
7072 [Roman E. Pavlov <pre@mo.msk.ru>]
7074 *) The x509 application mishandled signing requests containing DSA
7075 keys when the signing key was also DSA and the parameters didn't match.
7077 It was supposed to omit the parameters when they matched the signing key:
7078 the verifying software was then supposed to automatically use the CA's
7079 parameters if they were absent from the end user certificate.
7081 Omitting parameters is no longer recommended. The test was also
7082 the wrong way round! This was probably due to unusual behaviour in
7083 EVP_cmp_parameters() which returns 1 if the parameters match.
7084 This meant that parameters were omitted when they *didn't* match and
7085 the certificate was useless. Certificates signed with 'ca' didn't have
7087 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
7089 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
7090 The interface is as follows:
7091 Applications can use
7092 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
7093 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
7094 "off" is now the default.
7095 The library internally uses
7096 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
7097 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
7098 to disable memory-checking temporarily.
7100 Some inconsistent states that previously were possible (and were
7101 even the default) are now avoided.
7103 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
7104 with each memory chunk allocated; this is occasionally more helpful
7105 than just having a counter.
7107 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
7109 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
7113 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
7114 which largely parallels "options", but is for changing API behaviour,
7115 whereas "options" are about protocol behaviour.
7116 Initial "mode" flags are:
7118 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
7119 a single record has been written.
7120 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
7121 retries use the same buffer location.
7122 (But all of the contents must be
7126 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
7129 *) Fix problems with no-hmac etc.
7130 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
7132 *) New functions RSA_get_default_method(), RSA_set_method() and
7133 RSA_get_method(). These allows replacement of RSA_METHODs without having
7134 to mess around with the internals of an RSA structure.
7137 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
7138 Also really enable memory leak checks in openssl.c and in some
7140 [Chad C. Mulligan, Bodo Moeller]
7142 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
7143 up the length of negative integers. This has now been simplified to just
7144 store the length when it is first determined and use it later, rather
7145 than trying to keep track of where data is copied and updating it to
7147 [Steve Henson, reported by Brien Wheeler
7148 <bwheeler@authentica-security.com>]
7150 *) Add a new function PKCS7_signatureVerify. This allows the verification
7151 of a PKCS#7 signature but with the signing certificate passed to the
7152 function itself. This contrasts with PKCS7_dataVerify which assumes the
7153 certificate is present in the PKCS#7 structure. This isn't always the
7154 case: certificates can be omitted from a PKCS#7 structure and be
7155 distributed by "out of band" means (such as a certificate database).
7158 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
7159 function prototypes in pem.h, also change util/mkdef.pl to add the
7160 necessary function names.
7163 *) mk1mf.pl (used by Windows builds) did not properly read the
7164 options set by Configure in the top level Makefile, and Configure
7165 was not even able to write more than one option correctly.
7166 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
7169 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
7170 file to be loaded from a BIO or FILE pointer. The BIO version will
7171 for example allow memory BIOs to contain config info.
7174 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
7175 Whoever hopes to achieve shared-library compatibility across versions
7176 must use this, not the compile-time macro.
7177 (Exercise 0.9.4: Which is the minimum library version required by
7179 Note: All this applies only to multi-threaded programs, others don't
7183 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
7184 through a BIO pair triggered the default case, i.e.
7185 SSLerr(...,SSL_R_UNKNOWN_STATE).
7188 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
7189 can use the SSL library even if none of the specific BIOs is
7193 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
7194 for the encoded length.
7195 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
7197 *) Add initial documentation of the X509V3 functions.
7200 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
7201 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
7202 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
7203 secure PKCS#8 private key format with a high iteration count.
7206 *) Fix determination of Perl interpreter: A perl or perl5
7207 _directory_ in $PATH was also accepted as the interpreter.
7208 [Ralf S. Engelschall]
7210 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
7211 wrong with it but it was very old and did things like calling
7212 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
7216 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
7217 to use the new extension code.
7220 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
7221 with macros. This should make it easier to change their form, add extra
7222 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
7226 *) Add to configuration table a new entry that can specify an alternative
7227 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
7228 according to Mark Crispin <MRC@Panda.COM>.
7232 *) DES CBC did not update the IV. Weird.
7235 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
7236 Changing the behaviour of the former might break existing programs --
7237 where IV updating is needed, des_ncbc_encrypt can be used.
7240 *) When bntest is run from "make test" it drives bc to check its
7241 calculations, as well as internally checking them. If an internal check
7242 fails, it needs to cause bc to give a non-zero result or make test carries
7243 on without noticing the failure. Fixed.
7246 *) DES library cleanups.
7249 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
7250 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
7251 ciphers. NOTE: although the key derivation function has been verified
7252 against some published test vectors it has not been extensively tested
7253 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
7257 *) Instead of "mkdir -p", which is not fully portable, use new
7258 Perl script "util/mkdir-p.pl".
7261 *) Rewrite the way password based encryption (PBE) is handled. It used to
7262 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
7263 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
7264 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
7265 the 'parameter' field of the AlgorithmIdentifier is passed to the
7266 underlying key generation function so it must do its own ASN1 parsing.
7267 This has also changed the EVP_PBE_CipherInit() function which now has a
7268 'parameter' argument instead of literal salt and iteration count values
7269 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
7272 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
7273 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
7274 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
7275 KEY" because this clashed with PKCS#8 unencrypted string. Since this
7276 value was just used as a "magic string" and not used directly its
7277 value doesn't matter.
7280 *) Introduce some semblance of const correctness to BN. Shame C doesn't
7284 *) "linux-sparc64" configuration (ultrapenguin).
7285 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
7286 "linux-sparc" configuration.
7287 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
7289 *) config now generates no-xxx options for missing ciphers.
7292 *) Support the EBCDIC character set (work in progress).
7293 File ebcdic.c not yet included because it has a different license.
7294 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7296 *) Support BS2000/OSD-POSIX.
7297 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7299 *) Make callbacks for key generation use void * instead of char *.
7302 *) Make S/MIME samples compile (not yet tested).
7305 *) Additional typesafe stacks.
7308 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
7312 Changes between 0.9.3 and 0.9.3a [29 May 1999]
7314 *) New configuration variant "sco5-gcc".
7316 *) Updated some demos.
7317 [Sean O Riordain, Wade Scholine]
7319 *) Add missing BIO_free at exit of pkcs12 application.
7322 *) Fix memory leak in conf.c.
7325 *) Updates for Win32 to assembler version of MD5.
7328 *) Set #! path to perl in apps/der_chop to where we found it
7329 instead of using a fixed path.
7332 *) SHA library changes for irix64-mips4-cc.
7335 *) Improvements for VMS support.
7339 Changes between 0.9.2b and 0.9.3 [24 May 1999]
7341 *) Bignum library bug fix. IRIX 6 passes "make test" now!
7342 This also avoids the problems with SC4.2 and unpatched SC5.
7343 [Andy Polyakov <appro@fy.chalmers.se>]
7345 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
7346 These are required because of the typesafe stack would otherwise break
7347 existing code. If old code used a structure member which used to be STACK
7348 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
7349 sk_num or sk_value it would produce an error because the num, data members
7350 are not present in STACK_OF. Now it just produces a warning. sk_set
7351 replaces the old method of assigning a value to sk_value
7352 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
7353 that does this will no longer work (and should use sk_set instead) but
7354 this could be regarded as a "questionable" behaviour anyway.
7357 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
7358 correctly handle encrypted S/MIME data.
7361 *) Change type of various DES function arguments from des_cblock
7362 (which means, in function argument declarations, pointer to char)
7363 to des_cblock * (meaning pointer to array with 8 char elements),
7364 which allows the compiler to do more typechecking; it was like
7365 that back in SSLeay, but with lots of ugly casts.
7367 Introduce new type const_des_cblock.
7370 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
7371 problems: find RecipientInfo structure that matches recipient certificate
7372 and initialise the ASN1 structures properly based on passed cipher.
7375 *) Belatedly make the BN tests actually check the results.
7378 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
7379 to and from BNs: it was completely broken. New compilation option
7380 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
7381 key elements as negative integers.
7384 *) Reorganize and speed up MD5.
7385 [Andy Polyakov <appro@fy.chalmers.se>]
7388 [Richard Levitte <richard@levitte.org>]
7390 *) New option -out to asn1parse to allow the parsed structure to be
7391 output to a file. This is most useful when combined with the -strparse
7392 option to examine the output of things like OCTET STRINGS.
7395 *) Make SSL library a little more fool-proof by not requiring any longer
7396 that SSL_set_{accept,connect}_state be called before
7397 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
7398 in many applications because usually everything *appeared* to work as
7399 intended anyway -- now it really works as intended).
7402 *) Move openssl.cnf out of lib/.
7405 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
7406 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
7407 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
7408 [Ralf S. Engelschall]
7410 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
7411 handle PKCS#7 enveloped data properly.
7412 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
7414 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
7415 copying pointers. The cert_st handling is changed by this in
7416 various ways (and thus what used to be known as ctx->default_cert
7417 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
7418 any longer when s->cert does not give us what we need).
7419 ssl_cert_instantiate becomes obsolete by this change.
7420 As soon as we've got the new code right (possibly it already is?),
7421 we have solved a couple of bugs of the earlier code where s->cert
7422 was used as if it could not have been shared with other SSL structures.
7424 Note that using the SSL API in certain dirty ways now will result
7425 in different behaviour than observed with earlier library versions:
7426 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
7427 does not influence s as it used to.
7429 In order to clean up things more thoroughly, inside SSL_SESSION
7430 we don't use CERT any longer, but a new structure SESS_CERT
7431 that holds per-session data (if available); currently, this is
7432 the peer's certificate chain and, for clients, the server's certificate
7433 and temporary key. CERT holds only those values that can have
7434 meaningful defaults in an SSL_CTX.
7437 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
7438 from the internal representation. Various PKCS#7 fixes: remove some
7439 evil casts and set the enc_dig_alg field properly based on the signing
7443 *) Allow PKCS#12 password to be set from the command line or the
7444 environment. Let 'ca' get its config file name from the environment
7445 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
7449 *) Allow certificate policies extension to use an IA5STRING for the
7450 organization field. This is contrary to the PKIX definition but
7451 VeriSign uses it and IE5 only recognises this form. Document 'x509'
7455 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
7456 without disallowing inline assembler and the like for non-pedantic builds.
7459 *) Support Borland C++ builder.
7460 [Janez Jere <jj@void.si>, modified by Ulf Möller]
7465 *) SHA-1 cleanups and performance enhancements.
7466 [Andy Polyakov <appro@fy.chalmers.se>]
7468 *) Sparc v8plus assembler for the bignum library.
7469 [Andy Polyakov <appro@fy.chalmers.se>]
7471 *) Accept any -xxx and +xxx compiler options in Configure.
7474 *) Update HPUX configuration.
7477 *) Add missing sk_<type>_unshift() function to safestack.h
7478 [Ralf S. Engelschall]
7480 *) New function SSL_CTX_use_certificate_chain_file that sets the
7481 "extra_cert"s in addition to the certificate. (This makes sense
7482 only for "PEM" format files, as chains as a whole are not
7486 *) Support verify_depth from the SSL API.
7487 x509_vfy.c had what can be considered an off-by-one-error:
7488 Its depth (which was not part of the external interface)
7489 was actually counting the number of certificates in a chain;
7490 now it really counts the depth.
7493 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
7494 instead of X509err, which often resulted in confusing error
7495 messages since the error codes are not globally unique
7496 (e.g. an alleged error in ssl3_accept when a certificate
7497 didn't match the private key).
7499 *) New function SSL_CTX_set_session_id_context that allows to set a default
7500 value (so that you don't need SSL_set_session_id_context for each
7501 connection using the SSL_CTX).
7504 *) OAEP decoding bug fix.
7507 *) Support INSTALL_PREFIX for package builders, as proposed by
7511 *) New Configure options "threads" and "no-threads". For systems
7512 where the proper compiler options are known (currently Solaris
7513 and Linux), "threads" is the default.
7516 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
7519 *) Install various scripts to $(OPENSSLDIR)/misc, not to
7520 $(INSTALLTOP)/bin -- they shouldn't clutter directories
7521 such as /usr/local/bin.
7524 *) "make linux-shared" to build shared libraries.
7525 [Niels Poppe <niels@netbox.org>]
7527 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
7530 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
7531 extension adding in x509 utility.
7534 *) Remove NOPROTO sections and error code comments.
7537 *) Partial rewrite of the DEF file generator to now parse the ANSI
7541 *) New Configure options --prefix=DIR and --openssldir=DIR.
7544 *) Complete rewrite of the error code script(s). It is all now handled
7545 by one script at the top level which handles error code gathering,
7546 header rewriting and C source file generation. It should be much better
7547 than the old method: it now uses a modified version of Ulf's parser to
7548 read the ANSI prototypes in all header files (thus the old K&R definitions
7549 aren't needed for error creation any more) and do a better job of
7550 translating function codes into names. The old 'ASN1 error code imbedded
7551 in a comment' is no longer necessary and it doesn't use .err files which
7552 have now been deleted. Also the error code call doesn't have to appear all
7553 on one line (which resulted in some large lines...).
7556 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
7559 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
7560 0 (which usually indicates a closed connection), but continue reading.
7563 *) Fix some race conditions.
7566 *) Add support for CRL distribution points extension. Add Certificate
7567 Policies and CRL distribution points documentation.
7570 *) Move the autogenerated header file parts to crypto/opensslconf.h.
7573 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
7574 8 of keying material. Merlin has also confirmed interop with this fix
7575 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
7576 [Merlin Hughes <merlin@baltimore.ie>]
7578 *) Fix lots of warnings.
7579 [Richard Levitte <levitte@stacken.kth.se>]
7581 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
7582 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
7583 [Richard Levitte <levitte@stacken.kth.se>]
7585 *) Fix problems with sizeof(long) == 8.
7586 [Andy Polyakov <appro@fy.chalmers.se>]
7588 *) Change functions to ANSI C.
7591 *) Fix typos in error codes.
7592 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
7594 *) Remove defunct assembler files from Configure.
7597 *) SPARC v8 assembler BIGNUM implementation.
7598 [Andy Polyakov <appro@fy.chalmers.se>]
7600 *) Support for Certificate Policies extension: both print and set.
7601 Various additions to support the r2i method this uses.
7604 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
7605 return a const string when you are expecting an allocated buffer.
7608 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
7609 types DirectoryString and DisplayText.
7612 *) Add code to allow r2i extensions to access the configuration database,
7613 add an LHASH database driver and add several ctx helper functions.
7616 *) Fix an evil bug in bn_expand2() which caused various BN functions to
7617 fail when they extended the size of a BIGNUM.
7620 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
7621 support typesafe stack.
7624 *) Fix typo in SSL_[gs]et_options().
7625 [Nils Frostberg <nils@medcom.se>]
7627 *) Delete various functions and files that belonged to the (now obsolete)
7628 old X509V3 handling code.
7631 *) New Configure option "rsaref".
7634 *) Don't auto-generate pem.h.
7637 *) Introduce type-safe ASN.1 SETs.
7640 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
7641 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
7643 *) Introduce type-safe STACKs. This will almost certainly break lots of code
7644 that links with OpenSSL (well at least cause lots of warnings), but fear
7645 not: the conversion is trivial, and it eliminates loads of evil casts. A
7646 few STACKed things have been converted already. Feel free to convert more.
7647 In the fullness of time, I'll do away with the STACK type altogether.
7650 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
7651 specified in <certfile> by updating the entry in the index.txt file.
7652 This way one no longer has to edit the index.txt file manually for
7653 revoking a certificate. The -revoke option does the gory details now.
7654 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
7656 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
7657 `-text' option at all and this way the `-noout -text' combination was
7658 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
7659 [Ralf S. Engelschall]
7661 *) Make sure a corresponding plain text error message exists for the
7662 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
7663 verify callback function determined that a certificate was revoked.
7664 [Ralf S. Engelschall]
7666 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
7667 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
7668 all available cipers including rc5, which was forgotten until now.
7669 In order to let the testing shell script know which algorithms
7670 are available, a new (up to now undocumented) command
7671 "openssl list-cipher-commands" is used.
7674 *) Bugfix: s_client occasionally would sleep in select() when
7675 it should have checked SSL_pending() first.
7678 *) New functions DSA_do_sign and DSA_do_verify to provide access to
7679 the raw DSA values prior to ASN.1 encoding.
7682 *) Tweaks to Configure
7683 [Niels Poppe <niels@netbox.org>]
7685 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
7689 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
7692 *) New config option to avoid instructions that are illegal on the 80386.
7693 The default code is faster, but requires at least a 486.
7696 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
7697 SSL2_SERVER_VERSION (not used at all) macros, which are now the
7698 same as SSL2_VERSION anyway.
7701 *) New "-showcerts" option for s_client.
7704 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
7705 application. Various cleanups and fixes.
7708 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
7709 modify error routines to work internally. Add error codes and PBE init
7710 to library startup routines.
7713 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
7714 packing functions to asn1 and evp. Changed function names and error
7715 codes along the way.
7718 *) PKCS12 integration: and so it begins... First of several patches to
7719 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
7720 objects to objects.h
7723 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
7724 and display support for Thawte strong extranet extension.
7727 *) Add LinuxPPC support.
7728 [Jeff Dubrule <igor@pobox.org>]
7730 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
7731 bn_div_words in alpha.s.
7732 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
7734 *) Make sure the RSA OAEP test is skipped under -DRSAref because
7735 OAEP isn't supported when OpenSSL is built with RSAref.
7736 [Ulf Moeller <ulf@fitug.de>]
7738 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
7739 so they no longer are missing under -DNOPROTO.
7740 [Soren S. Jorvang <soren@t.dk>]
7743 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
7745 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
7746 doesn't work when the session is reused. Coming soon!
7749 *) Fix a security hole, that allows sessions to be reused in the wrong
7750 context thus bypassing client cert protection! All software that uses
7751 client certs and session caches in multiple contexts NEEDS PATCHING to
7752 allow session reuse! A fuller solution is in the works.
7753 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
7755 *) Some more source tree cleanups (removed obsolete files
7756 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
7757 permission on "config" script to be executable) and a fix for the INSTALL
7759 [Ulf Moeller <ulf@fitug.de>]
7761 *) Remove some legacy and erroneous uses of malloc, free instead of
7763 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
7765 *) Make rsa_oaep_test return non-zero on error.
7766 [Ulf Moeller <ulf@fitug.de>]
7768 *) Add support for native Solaris shared libraries. Configure
7769 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
7770 if someone would make that last step automatic.
7771 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
7773 *) ctx_size was not built with the right compiler during "make links". Fixed.
7776 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
7777 except NULL ciphers". This means the default cipher list will no longer
7778 enable NULL ciphers. They need to be specifically enabled e.g. with
7779 the string "DEFAULT:eNULL".
7782 *) Fix to RSA private encryption routines: if p < q then it would
7783 occasionally produce an invalid result. This will only happen with
7784 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
7787 *) Be less restrictive and allow also `perl util/perlpath.pl
7788 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
7789 because this way one can also use an interpreter named `perl5' (which is
7790 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
7791 installed as `perl').
7792 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7794 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
7795 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7797 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
7798 advapi32.lib to Win32 build and change the pem test comparision
7799 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
7800 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
7801 and crypto/des/ede_cbcm_enc.c.
7804 *) DES quad checksum was broken on big-endian architectures. Fixed.
7807 *) Comment out two functions in bio.h that aren't implemented. Fix up the
7808 Win32 test batch file so it (might) work again. The Win32 test batch file
7809 is horrible: I feel ill....
7812 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
7813 in e_os.h. Audit of header files to check ANSI and non ANSI
7814 sections: 10 functions were absent from non ANSI section and not exported
7815 from Windows DLLs. Fixed up libeay.num for new functions.
7818 *) Make `openssl version' output lines consistent.
7819 [Ralf S. Engelschall]
7821 *) Fix Win32 symbol export lists for BIO functions: Added
7822 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
7823 to ms/libeay{16,32}.def.
7824 [Ralf S. Engelschall]
7826 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
7827 fine under Unix and passes some trivial tests I've now added. But the
7828 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
7829 added to make sure no one expects that this stuff really works in the
7830 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
7831 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
7833 [Ralf S. Engelschall]
7835 *) Fix the generation of two part addresses in perl.
7836 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
7838 *) Add config entry for Linux on MIPS.
7839 [John Tobey <jtobey@channel1.com>]
7841 *) Make links whenever Configure is run, unless we are on Windoze.
7844 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
7845 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
7849 *) Add a useful kludge to allow package maintainers to specify compiler and
7850 other platforms details on the command line without having to patch the
7851 Configure script everytime: One now can use ``perl Configure
7852 <id>:<details>'', i.e. platform ids are allowed to have details appended
7853 to them (seperated by colons). This is treated as there would be a static
7854 pre-configured entry in Configure's %table under key <id> with value
7855 <details> and ``perl Configure <id>'' is called. So, when you want to
7856 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
7857 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
7858 now, which overrides the FreeBSD-elf entry on-the-fly.
7859 [Ralf S. Engelschall]
7861 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
7864 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
7865 on the `perl Configure ...' command line. This way one can compile
7866 OpenSSL libraries with Position Independent Code (PIC) which is needed
7867 for linking it into DSOs.
7868 [Ralf S. Engelschall]
7870 *) Remarkably, export ciphers were totally broken and no-one had noticed!
7874 *) Cleaned up the LICENSE document: The official contact for any license
7875 questions now is the OpenSSL core team under openssl-core@openssl.org.
7876 And add a paragraph about the dual-license situation to make sure people
7877 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
7878 to the OpenSSL toolkit.
7879 [Ralf S. Engelschall]
7881 *) General source tree makefile cleanups: Made `making xxx in yyy...'
7882 display consistent in the source tree and replaced `/bin/rm' by `rm'.
7883 Additonally cleaned up the `make links' target: Remove unnecessary
7884 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
7885 to speed processing and no longer clutter the display with confusing
7886 stuff. Instead only the actually done links are displayed.
7887 [Ralf S. Engelschall]
7889 *) Permit null encryption ciphersuites, used for authentication only. It used
7890 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
7891 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
7895 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
7896 signed attributes when verifying signatures (this would break them),
7897 the detached data encoding was wrong and public keys obtained using
7898 X509_get_pubkey() weren't freed.
7901 *) Add text documentation for the BUFFER functions. Also added a work around
7902 to a Win95 console bug. This was triggered by the password read stuff: the
7903 last character typed gets carried over to the next fread(). If you were
7904 generating a new cert request using 'req' for example then the last
7905 character of the passphrase would be CR which would then enter the first
7909 *) Added the new `Includes OpenSSL Cryptography Software' button as
7910 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
7911 button and can be used by applications based on OpenSSL to show the
7912 relationship to the OpenSSL project.
7913 [Ralf S. Engelschall]
7915 *) Remove confusing variables in function signatures in files
7916 ssl/ssl_lib.c and ssl/ssl.h.
7917 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7919 *) Don't install bss_file.c under PREFIX/include/
7920 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7922 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
7923 functions that return function pointers and has support for NT specific
7924 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
7925 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
7926 unsigned to signed types: this was killing the Win32 compile.
7929 *) Add new certificate file to stack functions,
7930 SSL_add_dir_cert_subjects_to_stack() and
7931 SSL_add_file_cert_subjects_to_stack(). These largely supplant
7932 SSL_load_client_CA_file(), and can be used to add multiple certs easily
7933 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
7934 This means that Apache-SSL and similar packages don't have to mess around
7935 to add as many CAs as they want to the preferred list.
7938 *) Experiment with doxygen documentation. Currently only partially applied to
7940 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
7941 openssl.doxy as the configuration file.
7944 *) Get rid of remaining C++-style comments which strict C compilers hate.
7945 [Ralf S. Engelschall, pointed out by Carlos Amengual]
7947 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
7948 compiled in by default: it has problems with large keys.
7951 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
7952 DH private keys and/or callback functions which directly correspond to
7953 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
7954 is needed for applications which have to configure certificates on a
7955 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
7957 For the RSA certificate situation is makes no difference, but
7958 for the DSA certificate situation this fixes the "no shared cipher"
7959 problem where the OpenSSL cipher selection procedure failed because the
7960 temporary keys were not overtaken from the context and the API provided
7961 no way to reconfigure them.
7962 The new functions now let applications reconfigure the stuff and they
7963 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
7964 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
7965 non-public-API function ssl_cert_instantiate() is used as a helper
7966 function and also to reduce code redundancy inside ssl_rsa.c.
7967 [Ralf S. Engelschall]
7969 *) Move s_server -dcert and -dkey options out of the undocumented feature
7970 area because they are useful for the DSA situation and should be
7971 recognized by the users.
7972 [Ralf S. Engelschall]
7974 *) Fix the cipher decision scheme for export ciphers: the export bits are
7975 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
7976 SSL_EXP_MASK. So, the original variable has to be used instead of the
7977 already masked variable.
7978 [Richard Levitte <levitte@stacken.kth.se>]
7980 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
7981 [Richard Levitte <levitte@stacken.kth.se>]
7983 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
7984 from `int' to `unsigned int' because it's a length and initialized by
7985 EVP_DigestFinal() which expects an `unsigned int *'.
7986 [Richard Levitte <levitte@stacken.kth.se>]
7988 *) Don't hard-code path to Perl interpreter on shebang line of Configure
7989 script. Instead use the usual Shell->Perl transition trick.
7990 [Ralf S. Engelschall]
7992 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
7993 (in addition to RSA certificates) to match the behaviour of `openssl dsa
7994 -noout -modulus' as it's already the case for `openssl rsa -noout
7995 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
7996 currently the public key is printed (a decision which was already done by
7997 `openssl dsa -modulus' in the past) which serves a similar purpose.
7998 Additionally the NO_RSA no longer completely removes the whole -modulus
7999 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
8001 [Ralf S. Engelschall]
8003 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
8004 BIO. See the source (crypto/evp/bio_ok.c) for more info.
8005 [Arne Ansper <arne@ats.cyber.ee>]
8007 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
8008 to be added. Now both 'req' and 'ca' can use new objects defined in the
8012 *) Add cool BIO that does syslog (or event log on NT).
8013 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
8015 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
8016 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
8017 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
8018 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
8021 *) Add preliminary config info for new extension code.
8024 *) Make RSA_NO_PADDING really use no padding.
8025 [Ulf Moeller <ulf@fitug.de>]
8027 *) Generate errors when private/public key check is done.
8030 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
8031 for some CRL extensions and new objects added.
8034 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
8035 key usage extension and fuller support for authority key id.
8038 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
8039 padding method for RSA, which is recommended for new applications in PKCS
8040 #1 v2.0 (RFC 2437, October 1998).
8041 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
8042 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
8043 against Bleichbacher's attack on RSA.
8044 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
8047 *) Updates to the new SSL compression code
8048 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8050 *) Fix so that the version number in the master secret, when passed
8051 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
8052 (because the server will not accept higher), that the version number
8053 is 0x03,0x01, not 0x03,0x00
8054 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8056 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
8057 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
8058 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
8061 *) Support for RAW extensions where an arbitrary extension can be
8062 created by including its DER encoding. See apps/openssl.cnf for
8066 *) Make sure latest Perl versions don't interpret some generated C array
8067 code as Perl array code in the crypto/err/err_genc.pl script.
8068 [Lars Weber <3weber@informatik.uni-hamburg.de>]
8070 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
8071 not many people have the assembler. Various Win32 compilation fixes and
8072 update to the INSTALL.W32 file with (hopefully) more accurate Win32
8076 *) Modify configure script 'Configure' to automatically create crypto/date.h
8077 file under Win32 and also build pem.h from pem.org. New script
8078 util/mkfiles.pl to create the MINFO file on environments that can't do a
8079 'make files': perl util/mkfiles.pl >MINFO should work.
8082 *) Major rework of DES function declarations, in the pursuit of correctness
8083 and purity. As a result, many evil casts evaporated, and some weirdness,
8084 too. You may find this causes warnings in your code. Zapping your evil
8085 casts will probably fix them. Mostly.
8088 *) Fix for a typo in asn1.h. Bug fix to object creation script
8089 obj_dat.pl. It considered a zero in an object definition to mean
8090 "end of object": none of the objects in objects.h have any zeros
8091 so it wasn't spotted.
8092 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
8094 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
8095 Masking (CBCM). In the absence of test vectors, the best I have been able
8096 to do is check that the decrypt undoes the encrypt, so far. Send me test
8097 vectors if you have them.
8100 *) Correct calculation of key length for export ciphers (too much space was
8101 allocated for null ciphers). This has not been tested!
8104 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
8105 message is now correct (it understands "crypto" and "ssl" on its
8106 command line). There is also now an "update" option. This will update
8107 the util/ssleay.num and util/libeay.num files with any new functions.
8109 perl util/mkdef.pl crypto ssl update
8110 it will update them.
8113 *) Overhauled the Perl interface (perl/*):
8114 - ported BN stuff to OpenSSL's different BN library
8115 - made the perl/ source tree CVS-aware
8116 - renamed the package from SSLeay to OpenSSL (the files still contain
8117 their history because I've copied them in the repository)
8118 - removed obsolete files (the test scripts will be replaced
8119 by better Test::Harness variants in the future)
8120 [Ralf S. Engelschall]
8122 *) First cut for a very conservative source tree cleanup:
8123 1. merge various obsolete readme texts into doc/ssleay.txt
8124 where we collect the old documents and readme texts.
8125 2. remove the first part of files where I'm already sure that we no
8126 longer need them because of three reasons: either they are just temporary
8127 files which were left by Eric or they are preserved original files where
8128 I've verified that the diff is also available in the CVS via "cvs diff
8129 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
8130 the crypto/md/ stuff).
8131 [Ralf S. Engelschall]
8133 *) More extension code. Incomplete support for subject and issuer alt
8134 name, issuer and authority key id. Change the i2v function parameters
8135 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
8136 what that's for :-) Fix to ASN1 macro which messed up
8137 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
8140 *) Preliminary support for ENUMERATED type. This is largely copied from the
8144 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
8145 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8147 *) Make sure `make rehash' target really finds the `openssl' program.
8148 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8150 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
8151 like to hear about it if this slows down other processors.
8154 *) Add CygWin32 platform information to Configure script.
8155 [Alan Batie <batie@aahz.jf.intel.com>]
8157 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
8158 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
8160 *) New program nseq to manipulate netscape certificate sequences
8163 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
8167 *) Fixes to BN code. Previously the default was to define BN_RECURSION
8168 but the BN code had some problems that would cause failures when
8169 doing certificate verification and some other functions.
8170 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8172 *) Add ASN1 and PEM code to support netscape certificate sequences.
8175 *) Add ASN1 and PEM code to support netscape certificate sequences.
8178 *) Add several PKIX and private extended key usage OIDs.
8181 *) Modify the 'ca' program to handle the new extension code. Modify
8182 openssl.cnf for new extension format, add comments.
8185 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
8186 and add a sample to openssl.cnf so req -x509 now adds appropriate
8190 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
8191 error code, add initial support to X509_print() and x509 application.
8194 *) Takes a deep breath and start addding X509 V3 extension support code. Add
8195 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
8196 stuff is currently isolated and isn't even compiled yet.
8199 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
8200 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
8201 Removed the versions check from X509 routines when loading extensions:
8202 this allows certain broken certificates that don't set the version
8203 properly to be processed.
8206 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
8207 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
8208 can still be regenerated with "make depend".
8211 *) Spelling mistake in C version of CAST-128.
8212 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
8214 *) Changes to the error generation code. The perl script err-code.pl
8215 now reads in the old error codes and retains the old numbers, only
8216 adding new ones if necessary. It also only changes the .err files if new
8217 codes are added. The makefiles have been modified to only insert errors
8218 when needed (to avoid needlessly modifying header files). This is done
8219 by only inserting errors if the .err file is newer than the auto generated
8220 C file. To rebuild all the error codes from scratch (the old behaviour)
8221 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
8222 or delete all the .err files.
8225 *) CAST-128 was incorrectly implemented for short keys. The C version has
8226 been fixed, but is untested. The assembler versions are also fixed, but
8227 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
8228 to regenerate it if needed.
8229 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
8230 Hagino <itojun@kame.net>]
8232 *) File was opened incorrectly in randfile.c.
8233 [Ulf Möller <ulf@fitug.de>]
8235 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
8236 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
8237 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
8238 al: it's just almost always a UTCTime. Note this patch adds new error
8239 codes so do a "make errors" if there are problems.
8242 *) Correct Linux 1 recognition in config.
8243 [Ulf Möller <ulf@fitug.de>]
8245 *) Remove pointless MD5 hash when using DSA keys in ca.
8246 [Anonymous <nobody@replay.com>]
8248 *) Generate an error if given an empty string as a cert directory. Also
8249 generate an error if handed NULL (previously returned 0 to indicate an
8250 error, but didn't set one).
8251 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
8253 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
8256 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
8257 parameters. This was causing a warning which killed off the Win32 compile.
8260 *) Remove C++ style comments from crypto/bn/bn_local.h.
8261 [Neil Costigan <neil.costigan@celocom.com>]
8263 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
8264 based on a text string, looking up short and long names and finally
8265 "dot" format. The "dot" format stuff didn't work. Added new function
8266 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
8267 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
8268 OID is not part of the table.
8271 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
8272 X509_LOOKUP_by_alias().
8275 *) Sort openssl functions by name.
8278 *) Get the gendsa program working (hopefully) and add it to app list. Remove
8279 encryption from sample DSA keys (in case anyone is interested the password
8283 *) Make _all_ *_free functions accept a NULL pointer.
8284 [Frans Heymans <fheymans@isaserver.be>]
8286 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
8288 [Anonymous <nobody@replay.com>]
8290 *) s_server should send the CAfile as acceptable CAs, not its own cert.
8291 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8293 *) Don't blow it for numeric -newkey arguments to apps/req.
8294 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8296 *) Temp key "for export" tests were wrong in s3_srvr.c.
8297 [Anonymous <nobody@replay.com>]
8299 *) Add prototype for temp key callback functions
8300 SSL_CTX_set_tmp_{rsa,dh}_callback().
8303 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
8304 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
8307 *) X509_name_add_entry() freed the wrong thing after an error.
8308 [Arne Ansper <arne@ats.cyber.ee>]
8310 *) rsa_eay.c would attempt to free a NULL context.
8311 [Arne Ansper <arne@ats.cyber.ee>]
8313 *) BIO_s_socket() had a broken should_retry() on Windoze.
8314 [Arne Ansper <arne@ats.cyber.ee>]
8316 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
8317 [Arne Ansper <arne@ats.cyber.ee>]
8319 *) Make sure the already existing X509_STORE->depth variable is initialized
8320 in X509_STORE_new(), but document the fact that this variable is still
8321 unused in the certificate verification process.
8322 [Ralf S. Engelschall]
8324 *) Fix the various library and apps files to free up pkeys obtained from
8325 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
8328 *) Fix reference counting in X509_PUBKEY_get(). This makes
8329 demos/maurice/example2.c work, amongst others, probably.
8330 [Steve Henson and Ben Laurie]
8332 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
8333 `openssl' and second, the shortcut symlinks for the `openssl <command>'
8334 are no longer created. This way we have a single and consistent command
8335 line interface `openssl <command>', similar to `cvs <command>'.
8336 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
8338 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
8339 BIT STRING wrapper always have zero unused bits.
8342 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
8345 *) Make the top-level INSTALL documentation easier to understand.
8348 *) Makefiles updated to exit if an error occurs in a sub-directory
8349 make (including if user presses ^C) [Paul Sutton]
8351 *) Make Montgomery context stuff explicit in RSA data structure.
8354 *) Fix build order of pem and err to allow for generated pem.h.
8357 *) Fix renumbering bug in X509_NAME_delete_entry().
8360 *) Enhanced the err-ins.pl script so it makes the error library number
8361 global and can add a library name. This is needed for external ASN1 and
8362 other error libraries.
8365 *) Fixed sk_insert which never worked properly.
8368 *) Fix ASN1 macros so they can handle indefinite length construted
8369 EXPLICIT tags. Some non standard certificates use these: they can now
8373 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
8374 into a single doc/ssleay.txt bundle. This way the information is still
8375 preserved but no longer messes up this directory. Now it's new room for
8376 the new set of documenation files.
8377 [Ralf S. Engelschall]
8379 *) SETs were incorrectly DER encoded. This was a major pain, because they
8380 shared code with SEQUENCEs, which aren't coded the same. This means that
8381 almost everything to do with SETs or SEQUENCEs has either changed name or
8382 number of arguments.
8383 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
8385 *) Fix test data to work with the above.
8388 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
8389 was already fixed by Eric for 0.9.1 it seems.
8390 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
8392 *) Autodetect FreeBSD3.
8395 *) Fix various bugs in Configure. This affects the following platforms:
8399 unixware-2.0-pentium
8403 *) Eliminate generated files from CVS. Reorder tests to regenerate files
8404 before they are needed.
8407 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
8411 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
8413 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
8414 changed SSLeay to OpenSSL in version strings.
8415 [Ralf S. Engelschall]
8417 *) Some fixups to the top-level documents.
8420 *) Fixed the nasty bug where rsaref.h was not found under compile-time
8421 because the symlink to include/ was missing.
8422 [Ralf S. Engelschall]
8424 *) Incorporated the popular no-RSA/DSA-only patches
8425 which allow to compile a RSA-free SSLeay.
8426 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
8428 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
8429 when "ssleay" is still not found.
8430 [Ralf S. Engelschall]
8432 *) Added more platforms to Configure: Cray T3E, HPUX 11,
8433 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
8435 *) Updated the README file.
8436 [Ralf S. Engelschall]
8438 *) Added various .cvsignore files in the CVS repository subdirs
8439 to make a "cvs update" really silent.
8440 [Ralf S. Engelschall]
8442 *) Recompiled the error-definition header files and added
8443 missing symbols to the Win32 linker tables.
8444 [Ralf S. Engelschall]
8446 *) Cleaned up the top-level documents;
8447 o new files: CHANGES and LICENSE
8448 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
8449 o merged COPYRIGHT into LICENSE
8450 o removed obsolete TODO file
8451 o renamed MICROSOFT to INSTALL.W32
8452 [Ralf S. Engelschall]
8454 *) Removed dummy files from the 0.9.1b source tree:
8455 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
8456 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
8457 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
8458 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
8459 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
8460 [Ralf S. Engelschall]
8462 *) Added various platform portability fixes.
8465 *) The Genesis of the OpenSSL rpject:
8466 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
8467 Young and Tim J. Hudson created while they were working for C2Net until
8469 [The OpenSSL Project]
8472 Changes between 0.9.0b and 0.9.1b [not released]
8474 *) Updated a few CA certificates under certs/
8477 *) Changed some BIGNUM api stuff.
8480 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
8481 DGUX x86, Linux Alpha, etc.
8484 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
8485 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
8489 *) Add -strparse option to asn1pars program which parses nested
8491 [Dr Stephen Henson <shenson@bigfoot.com>]
8493 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
8496 *) DSA fix for "ca" program.
8499 *) Added "-genkey" option to "dsaparam" program.
8502 *) Added RIPE MD160 (rmd160) message digest.
8505 *) Added -a (all) option to "ssleay version" command.
8508 *) Added PLATFORM define which is the id given to Configure.
8511 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
8514 *) Extended the ASN.1 parser routines.
8517 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
8520 *) Added a BN_CTX to the BN library.
8523 *) Fixed the weak key values in DES library
8526 *) Changed API in EVP library for cipher aliases.
8529 *) Added support for RC2/64bit cipher.
8532 *) Converted the lhash library to the crypto/mem.c functions.
8535 *) Added more recognized ASN.1 object ids.
8538 *) Added more RSA padding checks for SSL/TLS.
8541 *) Added BIO proxy/filter functionality.
8544 *) Added extra_certs to SSL_CTX which can be used
8545 send extra CA certificates to the client in the CA cert chain sending
8546 process. It can be configured with SSL_CTX_add_extra_chain_cert().
8549 *) Now Fortezza is denied in the authentication phase because
8550 this is key exchange mechanism is not supported by SSLeay at all.
8553 *) Additional PKCS1 checks.
8556 *) Support the string "TLSv1" for all TLS v1 ciphers.
8559 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
8560 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
8563 *) Fixed a few memory leaks.
8566 *) Fixed various code and comment typos.
8569 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
8570 bytes sent in the client random.
8571 [Edward Bishop <ebishop@spyglass.com>]