5 Changes between 0.9.8r and 0.9.8s [xx XXX xxxx]
7 *) Fix ssl_ciph.c set-up race.
8 [Adam Langley (Google)]
10 *) Fix spurious failures in ecdsatest.c.
11 [Emilia Käsper (Google)]
13 *) Fix the BIO_f_buffer() implementation (which was mixing different
14 interpretations of the '..._len' fields).
15 [Adam Langley (Google)]
17 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
18 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
19 threads won't reuse the same blinding coefficients.
21 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
22 lock to call BN_BLINDING_invert_ex, and avoids one use of
23 BN_BLINDING_update for each BN_BLINDING structure (previously,
24 the last update always remained unused).
25 [Emilia Käsper (Google)]
27 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
28 for multi-threaded use of ECDH.
29 [Adam Langley (Google)]
31 *) Fix x509_name_ex_d2i memory leak on bad inputs.
34 *) Add protection against ECDSA timing attacks as mentioned in the paper
35 by Billy Bob Brumley and Nicola Tuveri, see:
37 http://eprint.iacr.org/2011/232.pdf
39 [Billy Bob Brumley and Nicola Tuveri]
41 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
43 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
44 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
46 *) Fix bug in string printing code: if *any* escaping is enabled we must
47 escape the escape character (backslash) or the resulting string is
51 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
53 *) Disable code workaround for ancient and obsolete Netscape browsers
54 and servers: an attacker can use it in a ciphersuite downgrade attack.
55 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
58 *) Fixed J-PAKE implementation error, originally discovered by
59 Sebastien Martini, further info and confirmation from Stefan
60 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
63 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
65 *) Fix extension code to avoid race conditions which can result in a buffer
66 overrun vulnerability: resumed sessions must not be modified as they can
67 be shared by multiple threads. CVE-2010-3864
70 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
73 *) Don't reencode certificate when calculating signature: cache and use
74 the original encoding instead. This makes signature verification of
75 some broken encodings work correctly.
78 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
79 is also one of the inputs.
80 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
82 *) Don't repeatedly append PBE algorithms to table if they already exist.
83 Sort table on each new add. This effectively makes the table read only
84 after all algorithms are added and subsequent calls to PKCS12_pbe_add
88 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
90 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
93 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
94 access or freeing data twice (CVE-2010-0742)
95 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
97 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
98 common in certificates and some applications which only call
99 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
103 Reduce copying into .apps and .test in makevms.com
104 Don't try to use blank CA certificate in CA.com
105 Allow use of C files from original directories in maketests.com
106 [Steven M. Schweda" <sms@antinode.info>]
108 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
110 *) When rejecting SSL/TLS records due to an incorrect version number, never
111 update s->server with a new major version number. As of
112 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
113 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
114 the previous behavior could result in a read attempt at NULL when
115 receiving specific incorrect SSL/TLS records once record payload
116 protection is active. (CVE-2010-0740)
117 [Bodo Moeller, Adam Langley <agl@chromium.org>]
119 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
120 could be crashed if the relevant tables were not present (e.g. chrooted).
121 [Tomas Hoger <thoger@redhat.com>]
123 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
125 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
126 [Martin Olsson, Neel Mehta]
128 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
129 accommodate for stack sorting, always a write lock!).
132 *) On some versions of WIN32 Heap32Next is very slow. This can cause
133 excessive delays in the RAND_poll(): over a minute. As a workaround
134 include a time check in the inner Heap32Next loop too.
137 *) The code that handled flushing of data in SSL/TLS originally used the
138 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
139 the problem outlined in PR#1949. The fix suggested there however can
140 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
141 of Apache). So instead simplify the code to flush unconditionally.
142 This should be fine since flushing with no data to flush is a no op.
145 *) Handle TLS versions 2.0 and later properly and correctly use the
146 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
147 off ancient servers have a habit of sticking around for a while...
150 *) Modify compression code so it frees up structures without using the
151 ex_data callbacks. This works around a problem where some applications
152 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
153 restarting) then use compression (e.g. SSL with compression) later.
154 This results in significant per-connection memory leaks and
155 has caused some security issues including CVE-2008-1678 and
159 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
160 change when encrypting or decrypting.
163 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
164 connect and renegotiate with servers which do not support RI.
165 Until RI is more widely deployed this option is enabled by default.
168 *) Add "missing" ssl ctrls to clear options and mode.
171 *) If client attempts to renegotiate and doesn't support RI respond with
172 a no_renegotiation alert as required by RFC5746. Some renegotiating
173 TLS clients will continue a connection gracefully when they receive
174 the alert. Unfortunately OpenSSL mishandled this alert and would hang
175 waiting for a server hello which it will never receive. Now we treat a
176 received no_renegotiation alert as a fatal error. This is because
177 applications requesting a renegotiation might well expect it to succeed
178 and would have no code in place to handle the server denying it so the
179 only safe thing to do is to terminate the connection.
182 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
183 peer supports secure renegotiation and 0 otherwise. Print out peer
184 renegotiation support in s_client/s_server.
187 *) Replace the highly broken and deprecated SPKAC certification method with
188 the updated NID creation version. This should correctly handle UTF8.
191 *) Implement RFC5746. Re-enable renegotiation but require the extension
192 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
193 turns out to be a bad idea. It has been replaced by
194 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
195 SSL_CTX_set_options(). This is really not recommended unless you
196 know what you are doing.
197 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
199 *) Fixes to stateless session resumption handling. Use initial_ctx when
200 issuing and attempting to decrypt tickets in case it has changed during
201 servername handling. Use a non-zero length session ID when attempting
202 stateless session resumption: this makes it possible to determine if
203 a resumption has occurred immediately after receiving server hello
204 (several places in OpenSSL subtly assume this) instead of later in
208 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
209 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
210 fixes for a few places where the return code is not checked
212 [Julia Lawall <julia@diku.dk>]
214 *) Add --strict-warnings option to Configure script to include devteam
215 warnings in other configurations.
218 *) Add support for --libdir option and LIBDIR variable in makefiles. This
219 makes it possible to install openssl libraries in locations which
220 have names other than "lib", for example "/usr/lib64" which some
222 [Steve Henson, based on patch from Jeremy Utley]
224 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
225 X690 8.9.12 and can produce some misleading textual output of OIDs.
226 [Steve Henson, reported by Dan Kaminsky]
228 *) Delete MD2 from algorithm tables. This follows the recommendation in
229 several standards that it is not used in new applications due to
230 several cryptographic weaknesses. For binary compatibility reasons
231 the MD2 API is still compiled in by default.
234 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
238 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
239 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
241 [Guenter <lists@gknw.net>]
243 *) Fix the server certificate chain building code to use X509_verify_cert(),
244 it used to have an ad-hoc builder which was unable to cope with anything
245 other than a simple chain.
246 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
248 *) Don't check self signed certificate signatures in X509_verify_cert()
249 by default (a flag can override this): it just wastes time without
250 adding any security. As a useful side effect self signed root CAs
251 with non-FIPS digests are now usable in FIPS mode.
254 *) In dtls1_process_out_of_seq_message() the check if the current message
255 is already buffered was missing. For every new message was memory
256 allocated, allowing an attacker to perform an denial of service attack
257 with sending out of seq handshake messages until there is no memory
258 left. Additionally every future messege was buffered, even if the
259 sequence number made no sense and would be part of another handshake.
260 So only messages with sequence numbers less than 10 in advance will be
261 buffered. (CVE-2009-1378)
262 [Robin Seggelmann, discovered by Daniel Mentz]
264 *) Records are buffered if they arrive with a future epoch to be
265 processed after finishing the corresponding handshake. There is
266 currently no limitation to this buffer allowing an attacker to perform
267 a DOS attack with sending records with future epochs until there is no
268 memory left. This patch adds the pqueue_size() function to detemine
269 the size of a buffer and limits the record buffer to 100 entries.
271 [Robin Seggelmann, discovered by Daniel Mentz]
273 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
274 parent structure is freed. (CVE-2009-1379)
277 *) Handle non-blocking I/O properly in SSL_shutdown() call.
278 [Darryl Miles <darryl-mailinglists@netbauds.net>]
281 [Ilya O. <vrghost@gmail.com>]
283 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
285 *) Disable renegotiation completely - this fixes a severe security
286 problem (CVE-2009-3555) at the cost of breaking all
287 renegotiation. Renegotiation can be re-enabled by setting
288 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
289 run-time. This is really not recommended unless you know what
293 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
295 *) Don't set val to NULL when freeing up structures, it is freed up by
296 underlying code. If sizeof(void *) > sizeof(long) this can result in
297 zeroing past the valid field. (CVE-2009-0789)
298 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
300 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
301 checked correctly. This would allow some invalid signed attributes to
302 appear to verify correctly. (CVE-2009-0591)
303 [Ivan Nestlerode <inestlerode@us.ibm.com>]
305 *) Reject UniversalString and BMPString types with invalid lengths. This
306 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
307 a legal length. (CVE-2009-0590)
310 *) Set S/MIME signing as the default purpose rather than setting it
311 unconditionally. This allows applications to override it at the store
315 *) Permit restricted recursion of ASN1 strings. This is needed in practice
316 to handle some structures.
319 *) Improve efficiency of mem_gets: don't search whole buffer each time
321 [Jeremy Shapiro <jnshapir@us.ibm.com>]
323 *) New -hex option for openssl rand.
326 *) Print out UTF8String and NumericString when parsing ASN1.
329 *) Support NumericString type for name components.
332 *) Allow CC in the environment to override the automatically chosen
333 compiler. Note that nothing is done to ensure flags work with the
337 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
339 *) Properly check EVP_VerifyFinal() and similar return values
341 [Ben Laurie, Bodo Moeller, Google Security Team]
343 *) Enable TLS extensions by default.
346 *) Allow the CHIL engine to be loaded, whether the application is
347 multithreaded or not. (This does not release the developer from the
348 obligation to set up the dynamic locking callbacks.)
349 [Sander Temme <sander@temme.net>]
351 *) Use correct exit code if there is an error in dgst command.
352 [Steve Henson; problem pointed out by Roland Dirlewanger]
354 *) Tweak Configure so that you need to say "experimental-jpake" to enable
355 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
358 *) Add experimental JPAKE support, including demo authentication in
359 s_client and s_server.
362 *) Set the comparison function in v3_addr_canonize().
363 [Rob Austein <sra@hactrn.net>]
365 *) Add support for XMPP STARTTLS in s_client.
366 [Philip Paeps <philip@freebsd.org>]
368 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
369 to ensure that even with this option, only ciphersuites in the
370 server's preference list will be accepted. (Note that the option
371 applies only when resuming a session, so the earlier behavior was
372 just about the algorithm choice for symmetric cryptography.)
375 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
377 *) Fix NULL pointer dereference if a DTLS server received
378 ChangeCipherSpec as first record (CVE-2009-1386).
381 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
382 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
385 *) The fix in 0.9.8c that supposedly got rid of unsafe
386 double-checked locking was incomplete for RSA blinding,
387 addressing just one layer of what turns out to have been
388 doubly unsafe triple-checked locking.
390 So now fix this for real by retiring the MONT_HELPER macro
391 in crypto/rsa/rsa_eay.c.
393 [Bodo Moeller; problem pointed out by Marius Schilder]
395 *) Various precautionary measures:
397 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
399 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
400 (NB: This would require knowledge of the secret session ticket key
401 to exploit, in which case you'd be SOL either way.)
403 - Change bn_nist.c so that it will properly handle input BIGNUMs
404 outside the expected range.
406 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
409 [Neel Mehta, Bodo Moeller]
411 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
412 the load fails. Useful for distros.
413 [Ben Laurie and the FreeBSD team]
415 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
418 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
421 *) Expand ENGINE to support engine supplied SSL client certificate functions.
423 This work was sponsored by Logica.
426 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
427 keystores. Support for SSL/TLS client authentication too.
428 Not compiled unless enable-capieng specified to Configure.
430 This work was sponsored by Logica.
433 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
434 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
435 attribute creation routines such as certifcate requests and PKCS#12
439 Changes between 0.9.8g and 0.9.8h [28 May 2008]
441 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
442 handshake which could lead to a cilent crash as found using the
443 Codenomicon TLS test suite (CVE-2008-1672)
444 [Steve Henson, Mark Cox]
446 *) Fix double free in TLS server name extensions which could lead to
447 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
450 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
452 Clear the error queue to ensure that error entries left from
453 older function calls do not interfere with the correct operation.
454 [Lutz Jaenicke, Erik de Castro Lopo]
456 *) Remove root CA certificates of commercial CAs:
458 The OpenSSL project does not recommend any specific CA and does not
459 have any policy with respect to including or excluding any CA.
460 Therefore it does not make any sense to ship an arbitrary selection
461 of root CA certificates with the OpenSSL software.
464 *) RSA OAEP patches to fix two separate invalid memory reads.
465 The first one involves inputs when 'lzero' is greater than
466 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
467 before the beginning of from). The second one involves inputs where
468 the 'db' section contains nothing but zeroes (there is a one-byte
469 invalid read after the end of 'db').
470 [Ivan Nestlerode <inestlerode@us.ibm.com>]
472 *) Partial backport from 0.9.9-dev:
474 Introduce bn_mul_mont (dedicated Montgomery multiplication
475 procedure) as a candidate for BIGNUM assembler implementation.
476 While 0.9.9-dev uses assembler for various architectures, only
477 x86_64 is available by default here in the 0.9.8 branch, and
478 32-bit x86 is available through a compile-time setting.
480 To try the 32-bit x86 assembler implementation, use Configure
481 option "enable-montasm" (which exists only for this backport).
483 As "enable-montasm" for 32-bit x86 disclaims code stability
484 anyway, in this constellation we activate additional code
485 backported from 0.9.9-dev for further performance improvements,
486 namely BN_from_montgomery_word. (To enable this otherwise,
487 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
489 [Andy Polyakov (backport partially by Bodo Moeller)]
491 *) Add TLS session ticket callback. This allows an application to set
492 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
493 values. This is useful for key rollover for example where several key
494 sets may exist with different names.
497 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
498 This was broken until now in 0.9.8 releases, such that the only way
499 a registered ENGINE could be used (assuming it initialises
500 successfully on the host) was to explicitly set it as the default
501 for the relevant algorithms. This is in contradiction with 0.9.7
502 behaviour and the documentation. With this fix, when an ENGINE is
503 registered into a given algorithm's table of implementations, the
504 'uptodate' flag is reset so that auto-discovery will be used next
505 time a new context for that algorithm attempts to select an
507 [Ian Lister (tweaked by Geoff Thorpe)]
509 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
510 implemention in the following ways:
512 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
515 Lack of BER streaming support means one pass streaming processing is
516 only supported if data is detached: setting the streaming flag is
517 ignored for embedded content.
519 CMS support is disabled by default and must be explicitly enabled
520 with the enable-cms configuration option.
523 *) Update the GMP engine glue to do direct copies between BIGNUM and
524 mpz_t when openssl and GMP use the same limb size. Otherwise the
525 existing "conversion via a text string export" trick is still used.
526 [Paul Sheer <paulsheer@gmail.com>]
528 *) Zlib compression BIO. This is a filter BIO which compressed and
529 uncompresses any data passed through it.
532 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
533 RFC3394 compatible AES key wrapping.
536 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
537 sets string data without copying. X509_ALGOR_set0() and
538 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
539 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
540 from an X509_ATTRIBUTE structure optionally checking it occurs only
541 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
545 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
546 to get the expected BN_FLG_CONSTTIME behavior.
547 [Bodo Moeller (Google)]
551 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
552 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
553 - added some more tests to do_tests.pl
554 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
555 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
556 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
557 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
558 - various changes to netware.pl to enable gcc-cross builds on Win32
560 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
561 - various changes to fix missing prototype warnings
562 - fixed x86nasm.pl to create correct asm files for NASM COFF output
563 - added AES, WHIRLPOOL and CPUID assembler code to build files
564 - added missing AES assembler make rules to mk1mf.pl
565 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
566 [Guenter Knauf <eflash@gmx.net>]
568 *) Implement certificate status request TLS extension defined in RFC3546.
569 A client can set the appropriate parameters and receive the encoded
570 OCSP response via a callback. A server can query the supplied parameters
571 and set the encoded OCSP response in the callback. Add simplified examples
572 to s_client and s_server.
575 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
578 + Binary incompatibility of ssl_ctx_st structure
579 + DTLS interoperation with non-compliant servers
580 + Don't call get_session_cb() without proposed session
581 + Fix ia64 assembler code
582 [Andy Polyakov, Steve Henson]
584 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
586 *) DTLS Handshake overhaul. There were longstanding issues with
587 OpenSSL DTLS implementation, which were making it impossible for
588 RFC 4347 compliant client to communicate with OpenSSL server.
589 Unfortunately just fixing these incompatibilities would "cut off"
590 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
591 server keeps tolerating non RFC compliant syntax. The opposite is
592 not true, 0.9.8f client can not communicate with earlier server.
593 This update even addresses CVE-2007-4995.
596 *) Changes to avoid need for function casts in OpenSSL: some compilers
597 (gcc 4.2 and later) reject their use.
598 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
601 *) Add RFC4507 support to OpenSSL. This includes the corrections in
602 RFC4507bis. The encrypted ticket format is an encrypted encoded
603 SSL_SESSION structure, that way new session features are automatically
606 If a client application caches session in an SSL_SESSION structure
607 support is transparent because tickets are now stored in the encoded
610 The SSL_CTX structure automatically generates keys for ticket
611 protection in servers so again support should be possible
612 with no application modification.
614 If a client or server wishes to disable RFC4507 support then the option
615 SSL_OP_NO_TICKET can be set.
617 Add a TLS extension debugging callback to allow the contents of any client
618 or server extensions to be examined.
620 This work was sponsored by Google.
623 *) Add initial support for TLS extensions, specifically for the server_name
624 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
625 have new members for a host name. The SSL data structure has an
626 additional member SSL_CTX *initial_ctx so that new sessions can be
627 stored in that context to allow for session resumption, even after the
628 SSL has been switched to a new SSL_CTX in reaction to a client's
629 server_name extension.
631 New functions (subject to change):
634 SSL_get_servername_type()
637 New CTRL codes and macros (subject to change):
639 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
640 - SSL_CTX_set_tlsext_servername_callback()
641 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
642 - SSL_CTX_set_tlsext_servername_arg()
643 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
645 openssl s_client has a new '-servername ...' option.
647 openssl s_server has new options '-servername_host ...', '-cert2 ...',
648 '-key2 ...', '-servername_fatal' (subject to change). This allows
649 testing the HostName extension for a specific single host name ('-cert'
650 and '-key' remain fallbacks for handshakes without HostName
651 negotiation). If the unrecogninzed_name alert has to be sent, this by
652 default is a warning; it becomes fatal with the '-servername_fatal'
655 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
657 *) Add AES and SSE2 assembly language support to VC++ build.
660 *) Mitigate attack on final subtraction in Montgomery reduction.
663 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
664 (which previously caused an internal error).
667 *) Squeeze another 10% out of IGE mode when in != out.
670 *) AES IGE mode speedup.
671 [Dean Gaudet (Google)]
673 *) Add the Korean symmetric 128-bit cipher SEED (see
674 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
675 add SEED ciphersuites from RFC 4162:
677 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
678 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
679 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
680 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
682 To minimize changes between patchlevels in the OpenSSL 0.9.8
683 series, SEED remains excluded from compilation unless OpenSSL
684 is configured with 'enable-seed'.
687 *) Mitigate branch prediction attacks, which can be practical if a
688 single processor is shared, allowing a spy process to extract
689 information. For detailed background information, see
690 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
691 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
692 and Necessary Software Countermeasures"). The core of the change
693 are new versions BN_div_no_branch() and
694 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
695 respectively, which are slower, but avoid the security-relevant
696 conditional branches. These are automatically called by BN_div()
697 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
698 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
699 remove a conditional branch.
701 BN_FLG_CONSTTIME is the new name for the previous
702 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
703 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
704 in the exponent causes BN_mod_exp_mont() to use the alternative
705 implementation in BN_mod_exp_mont_consttime().) The old name
706 remains as a deprecated alias.
708 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
709 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
710 constant-time implementations for more than just exponentiation.
711 Here too the old name is kept as a deprecated alias.
713 BN_BLINDING_new() will now use BN_dup() for the modulus so that
714 the BN_BLINDING structure gets an independent copy of the
715 modulus. This means that the previous "BIGNUM *m" argument to
716 BN_BLINDING_new() and to BN_BLINDING_create_param() now
717 essentially becomes "const BIGNUM *m", although we can't actually
718 change this in the header file before 0.9.9. It allows
719 RSA_setup_blinding() to use BN_with_flags() on the modulus to
720 enable BN_FLG_CONSTTIME.
722 [Matthew D Wood (Intel Corp)]
724 *) In the SSL/TLS server implementation, be strict about session ID
725 context matching (which matters if an application uses a single
726 external cache for different purposes). Previously,
727 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
728 set. This did ensure strict client verification, but meant that,
729 with applications using a single external cache for quite
730 different requirements, clients could circumvent ciphersuite
731 restrictions for a given session ID context by starting a session
732 in a different context.
735 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
736 a ciphersuite string such as "DEFAULT:RSA" cannot enable
737 authentication-only ciphersuites.
740 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
741 not complete and could lead to a possible single byte overflow
742 (CVE-2007-5135) [Ben Laurie]
744 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
746 *) Since AES128 and AES256 (and similarly Camellia128 and
747 Camellia256) share a single mask bit in the logic of
748 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
749 kludge to work properly if AES128 is available and AES256 isn't
750 (or if Camellia128 is available and Camellia256 isn't).
753 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
754 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
755 When a point or a seed is encoded in a BIT STRING, we need to
756 prevent the removal of trailing zero bits to get the proper DER
757 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
758 of a NamedBitList, for which trailing 0 bits need to be removed.)
761 *) Have SSL/TLS server implementation tolerate "mismatched" record
762 protocol version while receiving ClientHello even if the
763 ClientHello is fragmented. (The server can't insist on the
764 particular protocol version it has chosen before the ServerHello
765 message has informed the client about his choice.)
768 *) Add RFC 3779 support.
769 [Rob Austein for ARIN, Ben Laurie]
771 *) Load error codes if they are not already present instead of using a
772 static variable. This allows them to be cleanly unloaded and reloaded.
773 Improve header file function name parsing.
776 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
777 or CAPABILITY handshake as required by RFCs.
780 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
782 *) Introduce limits to prevent malicious keys being able to
783 cause a denial of service. (CVE-2006-2940)
784 [Steve Henson, Bodo Moeller]
786 *) Fix ASN.1 parsing of certain invalid structures that can result
787 in a denial of service. (CVE-2006-2937) [Steve Henson]
789 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
790 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
792 *) Fix SSL client code which could crash if connecting to a
793 malicious SSLv2 server. (CVE-2006-4343)
794 [Tavis Ormandy and Will Drewry, Google Security Team]
796 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
797 match only those. Before that, "AES256-SHA" would be interpreted
798 as a pattern and match "AES128-SHA" too (since AES128-SHA got
799 the same strength classification in 0.9.7h) as we currently only
800 have a single AES bit in the ciphersuite description bitmap.
801 That change, however, also applied to ciphersuite strings such as
802 "RC4-MD5" that intentionally matched multiple ciphersuites --
803 namely, SSL 2.0 ciphersuites in addition to the more common ones
804 from SSL 3.0/TLS 1.0.
806 So we change the selection algorithm again: Naming an explicit
807 ciphersuite selects this one ciphersuite, and any other similar
808 ciphersuite (same bitmap) from *other* protocol versions.
809 Thus, "RC4-MD5" again will properly select both the SSL 2.0
810 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
812 Since SSL 2.0 does not have any ciphersuites for which the
813 128/256 bit distinction would be relevant, this works for now.
814 The proper fix will be to use different bits for AES128 and
815 AES256, which would have avoided the problems from the beginning;
816 however, bits are scarce, so we can only do this in a new release
817 (not just a patchlevel) when we can change the SSL_CIPHER
818 definition to split the single 'unsigned long mask' bitmap into
819 multiple values to extend the available space.
823 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
825 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
826 (CVE-2006-4339) [Ben Laurie and Google Security Team]
828 *) Add AES IGE and biIGE modes.
831 *) Change the Unix randomness entropy gathering to use poll() when
832 possible instead of select(), since the latter has some
833 undesirable limitations.
834 [Darryl Miles via Richard Levitte and Bodo Moeller]
836 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
837 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
838 cannot be implicitly activated as part of, e.g., the "AES" alias.
839 However, please upgrade to OpenSSL 0.9.9[-dev] for
840 non-experimental use of the ECC ciphersuites to get TLS extension
841 support, which is required for curve and point format negotiation
842 to avoid potential handshake problems.
845 *) Disable rogue ciphersuites:
847 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
848 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
849 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
851 The latter two were purportedly from
852 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
855 Also deactivate the remaining ciphersuites from
856 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
857 unofficial, and the ID has long expired.
860 *) Fix RSA blinding Heisenbug (problems sometimes occured on
861 dual-core machines) and other potential thread-safety issues.
864 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
865 versions), which is now available for royalty-free use
866 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
867 Also, add Camellia TLS ciphersuites from RFC 4132.
869 To minimize changes between patchlevels in the OpenSSL 0.9.8
870 series, Camellia remains excluded from compilation unless OpenSSL
871 is configured with 'enable-camellia'.
874 *) Disable the padding bug check when compression is in use. The padding
875 bug check assumes the first packet is of even length, this is not
876 necessarily true if compresssion is enabled and can result in false
877 positives causing handshake failure. The actual bug test is ancient
878 code so it is hoped that implementations will either have fixed it by
879 now or any which still have the bug do not support compression.
882 Changes between 0.9.8a and 0.9.8b [04 May 2006]
884 *) When applying a cipher rule check to see if string match is an explicit
885 cipher suite and only match that one cipher suite if it is.
888 *) Link in manifests for VC++ if needed.
889 [Austin Ziegler <halostatue@gmail.com>]
891 *) Update support for ECC-based TLS ciphersuites according to
892 draft-ietf-tls-ecc-12.txt with proposed changes (but without
893 TLS extensions, which are supported starting with the 0.9.9
894 branch, not in the OpenSSL 0.9.8 branch).
897 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
898 opaque EVP_CIPHER_CTX handling.
901 *) Fixes and enhancements to zlib compression code. We now only use
902 "zlib1.dll" and use the default __cdecl calling convention on Win32
903 to conform with the standards mentioned here:
904 http://www.zlib.net/DLL_FAQ.txt
905 Static zlib linking now works on Windows and the new --with-zlib-include
906 --with-zlib-lib options to Configure can be used to supply the location
907 of the headers and library. Gracefully handle case where zlib library
911 *) Several fixes and enhancements to the OID generation code. The old code
912 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
913 handle numbers larger than ULONG_MAX, truncated printing and had a
914 non standard OBJ_obj2txt() behaviour.
917 *) Add support for building of engines under engine/ as shared libraries
918 under VC++ build system.
921 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
922 Hopefully, we will not see any false combination of paths any more.
925 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
927 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
928 (part of SSL_OP_ALL). This option used to disable the
929 countermeasure against man-in-the-middle protocol-version
930 rollback in the SSL 2.0 server implementation, which is a bad
931 idea. (CVE-2005-2969)
933 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
934 for Information Security, National Institute of Advanced Industrial
935 Science and Technology [AIST], Japan)]
937 *) Add two function to clear and return the verify parameter flags.
940 *) Keep cipherlists sorted in the source instead of sorting them at
941 runtime, thus removing the need for a lock.
944 *) Avoid some small subgroup attacks in Diffie-Hellman.
945 [Nick Mathewson and Ben Laurie]
947 *) Add functions for well-known primes.
950 *) Extended Windows CE support.
951 [Satoshi Nakamura and Andy Polyakov]
953 *) Initialize SSL_METHOD structures at compile time instead of during
954 runtime, thus removing the need for a lock.
957 *) Make PKCS7_decrypt() work even if no certificate is supplied by
958 attempting to decrypt each encrypted key in turn. Add support to
962 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
964 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
967 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
970 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
971 key into the same file any more.
974 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
977 *) Add -utf8 command line and config file option to 'ca'.
978 [Stefan <stf@udoma.org]
980 *) Removed the macro des_crypt(), as it seems to conflict with some
981 libraries. Use DES_crypt().
984 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
985 involves renaming the source and generated shared-libs for
986 both. The engines will accept the corrected or legacy ids
987 ('ncipher' and '4758_cca' respectively) when binding. NB,
988 this only applies when building 'shared'.
989 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
991 *) Add attribute functions to EVP_PKEY structure. Modify
992 PKCS12_create() to recognize a CSP name attribute and
993 use it. Make -CSP option work again in pkcs12 utility.
996 *) Add new functionality to the bn blinding code:
997 - automatic re-creation of the BN_BLINDING parameters after
998 a fixed number of uses (currently 32)
999 - add new function for parameter creation
1000 - introduce flags to control the update behaviour of the
1001 BN_BLINDING parameters
1002 - hide BN_BLINDING structure
1003 Add a second BN_BLINDING slot to the RSA structure to improve
1004 performance when a single RSA object is shared among several
1008 *) Add support for DTLS.
1009 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
1011 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
1012 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
1015 *) Remove buggy and incompletet DH cert support from
1016 ssl/ssl_rsa.c and ssl/s3_both.c
1019 *) Use SHA-1 instead of MD5 as the default digest algorithm for
1020 the apps/openssl applications.
1023 *) Compile clean with "-Wall -Wmissing-prototypes
1024 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
1025 DEBUG_SAFESTACK must also be set.
1028 *) Change ./Configure so that certain algorithms can be disabled by default.
1029 The new counterpiece to "no-xxx" is "enable-xxx".
1031 The patented RC5 and MDC2 algorithms will now be disabled unless
1032 "enable-rc5" and "enable-mdc2", respectively, are specified.
1034 (IDEA remains enabled despite being patented. This is because IDEA
1035 is frequently required for interoperability, and there is no license
1036 fee for non-commercial use. As before, "no-idea" can be used to
1037 avoid this algorithm.)
1041 *) Add processing of proxy certificates (see RFC 3820). This work was
1042 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
1043 EGEE (Enabling Grids for E-science in Europe).
1046 *) RC4 performance overhaul on modern architectures/implementations, such
1047 as Intel P4, IA-64 and AMD64.
1050 *) New utility extract-section.pl. This can be used specify an alternative
1051 section number in a pod file instead of having to treat each file as
1052 a separate case in Makefile. This can be done by adding two lines to the
1055 =for comment openssl_section:XXX
1057 The blank line is mandatory.
1061 *) New arguments -certform, -keyform and -pass for s_client and s_server
1062 to allow alternative format key and certificate files and passphrase
1066 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
1067 update associated structures and add various utility functions.
1069 Add new policy related verify parameters, include policy checking in
1070 standard verify code. Enhance 'smime' application with extra parameters
1071 to support policy checking and print out.
1074 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
1075 Nehemiah processors. These extensions support AES encryption in hardware
1076 as well as RNG (though RNG support is currently disabled).
1077 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
1079 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
1082 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
1083 [Andy Polyakov and a number of other people]
1085 *) Improved PowerPC platform support. Most notably BIGNUM assembler
1086 implementation contributed by IBM.
1087 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
1089 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
1090 exponent rather than 'unsigned long'. There is a corresponding change to
1091 the new 'rsa_keygen' element of the RSA_METHOD structure.
1092 [Jelte Jansen, Geoff Thorpe]
1094 *) Functionality for creating the initial serial number file is now
1095 moved from CA.pl to the 'ca' utility with a new option -create_serial.
1097 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
1098 number file to 1, which is bound to cause problems. To avoid
1099 the problems while respecting compatibility between different 0.9.7
1100 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
1101 CA.pl for serial number initialization. With the new release 0.9.8,
1102 we can fix the problem directly in the 'ca' utility.)
1105 *) Reduced header interdepencies by declaring more opaque objects in
1106 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
1107 give fewer recursive includes, which could break lazy source code - so
1108 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
1109 developers should define this symbol when building and using openssl to
1110 ensure they track the recommended behaviour, interfaces, [etc], but
1111 backwards-compatible behaviour prevails when this isn't defined.
1114 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
1117 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
1118 This will generate a random key of the appropriate length based on the
1119 cipher context. The EVP_CIPHER can provide its own random key generation
1120 routine to support keys of a specific form. This is used in the des and
1121 3des routines to generate a key of the correct parity. Update S/MIME
1122 code to use new functions and hence generate correct parity DES keys.
1123 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
1124 valid (weak or incorrect parity).
1127 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
1128 as looking them up. This is useful when the verified structure may contain
1129 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
1130 present unless the new PKCS7_NO_CRL flag is asserted.
1133 *) Extend ASN1 oid configuration module. It now additionally accepts the
1136 shortName = some long name, 1.2.3.4
1139 *) Reimplemented the BN_CTX implementation. There is now no more static
1140 limitation on the number of variables it can handle nor the depth of the
1141 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
1142 information can now expand as required, and rather than having a single
1143 static array of bignums, BN_CTX now uses a linked-list of such arrays
1144 allowing it to expand on demand whilst maintaining the usefulness of
1145 BN_CTX's "bundling".
1148 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
1149 to allow all RSA operations to function using a single BN_CTX.
1152 *) Preliminary support for certificate policy evaluation and checking. This
1153 is initially intended to pass the tests outlined in "Conformance Testing
1154 of Relying Party Client Certificate Path Processing Logic" v1.07.
1157 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
1158 remained unused and not that useful. A variety of other little bignum
1159 tweaks and fixes have also been made continuing on from the audit (see
1163 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
1164 associated ASN1, EVP and SSL functions and old ASN1 macros.
1167 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
1168 and this should never fail. So the return value from the use of
1169 BN_set_word() (which can fail due to needless expansion) is now deprecated;
1170 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
1173 *) BN_CTX_get() should return zero-valued bignums, providing the same
1174 initialised value as BN_new().
1175 [Geoff Thorpe, suggested by Ulf Möller]
1177 *) Support for inhibitAnyPolicy certificate extension.
1180 *) An audit of the BIGNUM code is underway, for which debugging code is
1181 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
1182 is considered valid when processing BIGNUMs, and causes execution to
1183 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
1184 further steps are taken to deliberately pollute unused data in BIGNUM
1185 structures to try and expose faulty code further on. For now, openssl will
1186 (in its default mode of operation) continue to tolerate the inconsistent
1187 forms that it has tolerated in the past, but authors and packagers should
1188 consider trying openssl and their own applications when compiled with
1189 these debugging symbols defined. It will help highlight potential bugs in
1190 their own code, and will improve the test coverage for OpenSSL itself. At
1191 some point, these tighter rules will become openssl's default to improve
1192 maintainability, though the assert()s and other overheads will remain only
1193 in debugging configurations. See bn.h for more details.
1194 [Geoff Thorpe, Nils Larsch, Ulf Möller]
1196 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
1197 that can only be obtained through BN_CTX_new() (which implicitly
1198 initialises it). The presence of this function only made it possible
1199 to overwrite an existing structure (and cause memory leaks).
1202 *) Because of the callback-based approach for implementing LHASH as a
1203 template type, lh_insert() adds opaque objects to hash-tables and
1204 lh_doall() or lh_doall_arg() are typically used with a destructor callback
1205 to clean up those corresponding objects before destroying the hash table
1206 (and losing the object pointers). So some over-zealous constifications in
1207 LHASH have been relaxed so that lh_insert() does not take (nor store) the
1208 objects as "const" and the lh_doall[_arg] callback wrappers are not
1209 prototyped to have "const" restrictions on the object pointers they are
1210 given (and so aren't required to cast them away any more).
1213 *) The tmdiff.h API was so ugly and minimal that our own timing utility
1214 (speed) prefers to use its own implementation. The two implementations
1215 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
1216 its object type properly exposed (MS_TM) instead of casting to/from "char
1217 *". This may still change yet if someone realises MS_TM and "ms_time_***"
1218 aren't necessarily the greatest nomenclatures - but this is what was used
1219 internally to the implementation so I've used that for now.
1222 *) Ensure that deprecated functions do not get compiled when
1223 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
1224 the self-tests were still using deprecated key-generation functions so
1225 these have been updated also.
1228 *) Reorganise PKCS#7 code to separate the digest location functionality
1229 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
1230 New function PKCS7_set_digest() to set the digest type for PKCS#7
1231 digestedData type. Add additional code to correctly generate the
1232 digestedData type and add support for this type in PKCS7 initialization
1236 *) New function PKCS7_set0_type_other() this initializes a PKCS7
1237 structure of type "other".
1240 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
1241 sure the loop does correctly stop and breaking ("division by zero")
1242 modulus operations are not performed. The (pre-generated) prime
1243 table crypto/bn/bn_prime.h was already correct, but it could not be
1244 re-generated on some platforms because of the "division by zero"
1245 situation in the script.
1246 [Ralf S. Engelschall]
1248 *) Update support for ECC-based TLS ciphersuites according to
1249 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
1250 SHA-1 now is only used for "small" curves (where the
1251 representation of a field element takes up to 24 bytes); for
1252 larger curves, the field element resulting from ECDH is directly
1253 used as premaster secret.
1254 [Douglas Stebila (Sun Microsystems Laboratories)]
1256 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
1257 curve secp160r1 to the tests.
1258 [Douglas Stebila (Sun Microsystems Laboratories)]
1260 *) Add the possibility to load symbols globally with DSO.
1261 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
1263 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
1264 control of the error stack.
1267 *) Add support for STORE in ENGINE.
1270 *) Add the STORE type. The intention is to provide a common interface
1271 to certificate and key stores, be they simple file-based stores, or
1272 HSM-type store, or LDAP stores, or...
1273 NOTE: The code is currently UNTESTED and isn't really used anywhere.
1276 *) Add a generic structure called OPENSSL_ITEM. This can be used to
1277 pass a list of arguments to any function as well as provide a way
1278 for a function to pass data back to the caller.
1281 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
1282 works like BUF_strdup() but can be used to duplicate a portion of
1283 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
1287 *) Add the function sk_find_ex() which works like sk_find(), but will
1288 return an index to an element even if an exact match couldn't be
1289 found. The index is guaranteed to point at the element where the
1290 searched-for key would be inserted to preserve sorting order.
1293 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
1294 takes an extra flags argument for optional functionality. Currently,
1295 the following flags are defined:
1297 OBJ_BSEARCH_VALUE_ON_NOMATCH
1298 This one gets OBJ_bsearch_ex() to return a pointer to the first
1299 element where the comparing function returns a negative or zero
1302 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
1303 This one gets OBJ_bsearch_ex() to return a pointer to the first
1304 element where the comparing function returns zero. This is useful
1305 if there are more than one element where the comparing function
1309 *) Make it possible to create self-signed certificates with 'openssl ca'
1310 in such a way that the self-signed certificate becomes part of the
1311 CA database and uses the same mechanisms for serial number generation
1312 as all other certificate signing. The new flag '-selfsign' enables
1313 this functionality. Adapt CA.sh and CA.pl.in.
1316 *) Add functionality to check the public key of a certificate request
1317 against a given private. This is useful to check that a certificate
1318 request can be signed by that key (self-signing).
1321 *) Make it possible to have multiple active certificates with the same
1322 subject in the CA index file. This is done only if the keyword
1323 'unique_subject' is set to 'no' in the main CA section (default
1324 if 'CA_default') of the configuration file. The value is saved
1325 with the database itself in a separate index attribute file,
1326 named like the index file with '.attr' appended to the name.
1329 *) Generate muti valued AVAs using '+' notation in config files for
1333 *) Support for nameConstraints certificate extension.
1336 *) Support for policyConstraints certificate extension.
1339 *) Support for policyMappings certificate extension.
1342 *) Make sure the default DSA_METHOD implementation only uses its
1343 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
1344 and change its own handlers to be NULL so as to remove unnecessary
1345 indirection. This lets alternative implementations fallback to the
1346 default implementation more easily.
1349 *) Support for directoryName in GeneralName related extensions
1353 *) Make it possible to link applications using Makefile.shared.
1354 Make that possible even when linking against static libraries!
1357 *) Support for single pass processing for S/MIME signing. This now
1358 means that S/MIME signing can be done from a pipe, in addition
1359 cleartext signing (multipart/signed type) is effectively streaming
1360 and the signed data does not need to be all held in memory.
1362 This is done with a new flag PKCS7_STREAM. When this flag is set
1363 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
1364 is done after the data is output (and digests calculated) in
1365 SMIME_write_PKCS7().
1368 *) Add full support for -rpath/-R, both in shared libraries and
1369 applications, at least on the platforms where it's known how
1373 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
1374 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
1375 will now compute a table of multiples of the generator that
1376 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
1377 faster (notably in the case of a single point multiplication,
1378 scalar * generator).
1379 [Nils Larsch, Bodo Moeller]
1381 *) IPv6 support for certificate extensions. The various extensions
1382 which use the IP:a.b.c.d can now take IPv6 addresses using the
1383 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
1387 *) Added an ENGINE that implements RSA by performing private key
1388 exponentiations with the GMP library. The conversions to and from
1389 GMP's mpz_t format aren't optimised nor are any montgomery forms
1390 cached, and on x86 it appears OpenSSL's own performance has caught up.
1391 However there are likely to be other architectures where GMP could
1392 provide a boost. This ENGINE is not built in by default, but it can be
1393 specified at Configure time and should be accompanied by the necessary
1394 linker additions, eg;
1395 ./config -DOPENSSL_USE_GMP -lgmp
1398 *) "openssl engine" will not display ENGINE/DSO load failure errors when
1399 testing availability of engines with "-t" - the old behaviour is
1400 produced by increasing the feature's verbosity with "-tt".
1403 *) ECDSA routines: under certain error conditions uninitialized BN objects
1404 could be freed. Solution: make sure initialization is performed early
1405 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
1409 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
1410 and DH_METHOD (eg. by ENGINE implementations) to override the normal
1411 software implementations. For DSA and DH, parameter generation can
1412 also be overriden by providing the appropriate method callbacks.
1415 *) Change the "progress" mechanism used in key-generation and
1416 primality testing to functions that take a new BN_GENCB pointer in
1417 place of callback/argument pairs. The new API functions have "_ex"
1418 postfixes and the older functions are reimplemented as wrappers for
1419 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
1420 declarations of the old functions to help (graceful) attempts to
1421 migrate to the new functions. Also, the new key-generation API
1422 functions operate on a caller-supplied key-structure and return
1423 success/failure rather than returning a key or NULL - this is to
1424 help make "keygen" another member function of RSA_METHOD etc.
1426 Example for using the new callback interface:
1428 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
1432 BN_GENCB_set(&my_cb, my_callback, my_arg);
1434 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
1435 /* For the meaning of a, b in calls to my_callback(), see the
1436 * documentation of the function that calls the callback.
1437 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
1438 * my_callback should return 1 if it wants BN_is_prime_ex()
1439 * to continue, or 0 to stop.
1444 *) Change the ZLIB compression method to be stateful, and make it
1445 available to TLS with the number defined in
1446 draft-ietf-tls-compression-04.txt.
1449 *) Add the ASN.1 structures and functions for CertificatePair, which
1450 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
1452 CertificatePair ::= SEQUENCE {
1453 forward [0] Certificate OPTIONAL,
1454 reverse [1] Certificate OPTIONAL,
1455 -- at least one of the pair shall be present -- }
1457 Also implement the PEM functions to read and write certificate
1458 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
1460 This needed to be defined, mostly for the sake of the LDAP
1461 attribute crossCertificatePair, but may prove useful elsewhere as
1465 *) Make it possible to inhibit symlinking of shared libraries in
1466 Makefile.shared, for Cygwin's sake.
1469 *) Extend the BIGNUM API by creating a function
1470 void BN_set_negative(BIGNUM *a, int neg);
1471 and a macro that behave like
1472 int BN_is_negative(const BIGNUM *a);
1474 to avoid the need to access 'a->neg' directly in applications.
1477 *) Implement fast modular reduction for pseudo-Mersenne primes
1478 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
1479 EC_GROUP_new_curve_GFp() will now automatically use this
1481 [Nils Larsch <nla@trustcenter.de>]
1483 *) Add new lock type (CRYPTO_LOCK_BN).
1486 *) Change the ENGINE framework to automatically load engines
1487 dynamically from specific directories unless they could be
1488 found to already be built in or loaded. Move all the
1489 current engines except for the cryptodev one to a new
1491 The engines in engines/ are built as shared libraries if
1492 the "shared" options was given to ./Configure or ./config.
1493 Otherwise, they are inserted in libcrypto.a.
1494 /usr/local/ssl/engines is the default directory for dynamic
1495 engines, but that can be overriden at configure time through
1496 the usual use of --prefix and/or --openssldir, and at run
1497 time with the environment variable OPENSSL_ENGINES.
1498 [Geoff Thorpe and Richard Levitte]
1500 *) Add Makefile.shared, a helper makefile to build shared
1501 libraries. Addapt Makefile.org.
1504 *) Add version info to Win32 DLLs.
1505 [Peter 'Luna' Runestig" <peter@runestig.com>]
1507 *) Add new 'medium level' PKCS#12 API. Certificates and keys
1508 can be added using this API to created arbitrary PKCS#12
1509 files while avoiding the low level API.
1511 New options to PKCS12_create(), key or cert can be NULL and
1512 will then be omitted from the output file. The encryption
1513 algorithm NIDs can be set to -1 for no encryption, the mac
1514 iteration count can be set to 0 to omit the mac.
1516 Enhance pkcs12 utility by making the -nokeys and -nocerts
1517 options work when creating a PKCS#12 file. New option -nomac
1518 to omit the mac, NONE can be set for an encryption algorithm.
1519 New code is modified to use the enhanced PKCS12_create()
1520 instead of the low level API.
1523 *) Extend ASN1 encoder to support indefinite length constructed
1524 encoding. This can output sequences tags and octet strings in
1525 this form. Modify pk7_asn1.c to support indefinite length
1526 encoding. This is experimental and needs additional code to
1527 be useful, such as an ASN1 bio and some enhanced streaming
1530 Extend template encode functionality so that tagging is passed
1531 down to the template encoder.
1534 *) Let 'openssl req' fail if an argument to '-newkey' is not
1535 recognized instead of using RSA as a default.
1538 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
1539 As these are not official, they are not included in "ALL";
1540 the "ECCdraft" ciphersuite group alias can be used to select them.
1541 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
1543 *) Add ECDH engine support.
1544 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
1546 *) Add ECDH in new directory crypto/ecdh/.
1547 [Douglas Stebila (Sun Microsystems Laboratories)]
1549 *) Let BN_rand_range() abort with an error after 100 iterations
1550 without success (which indicates a broken PRNG).
1553 *) Change BN_mod_sqrt() so that it verifies that the input value
1554 is really the square of the return value. (Previously,
1555 BN_mod_sqrt would show GIGO behaviour.)
1558 *) Add named elliptic curves over binary fields from X9.62, SECG,
1559 and WAP/WTLS; add OIDs that were still missing.
1561 [Sheueling Chang Shantz and Douglas Stebila
1562 (Sun Microsystems Laboratories)]
1564 *) Extend the EC library for elliptic curves over binary fields
1565 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
1568 EC_GF2m_simple_method
1572 EC_GROUP_new_curve_GF2m
1573 EC_GROUP_set_curve_GF2m
1574 EC_GROUP_get_curve_GF2m
1575 EC_POINT_set_affine_coordinates_GF2m
1576 EC_POINT_get_affine_coordinates_GF2m
1577 EC_POINT_set_compressed_coordinates_GF2m
1579 Point compression for binary fields is disabled by default for
1580 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
1583 As binary polynomials are represented as BIGNUMs, various members
1584 of the EC_GROUP and EC_POINT data structures can be shared
1585 between the implementations for prime fields and binary fields;
1586 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
1587 are essentially identical to their ..._GFp counterparts.
1588 (For simplicity, the '..._GFp' prefix has been dropped from
1589 various internal method names.)
1591 An internal 'field_div' method (similar to 'field_mul' and
1592 'field_sqr') has been added; this is used only for binary fields.
1594 [Sheueling Chang Shantz and Douglas Stebila
1595 (Sun Microsystems Laboratories)]
1597 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
1598 through methods ('mul', 'precompute_mult').
1600 The generic implementations (now internally called 'ec_wNAF_mul'
1601 and 'ec_wNAF_precomputed_mult') remain the default if these
1602 methods are undefined.
1604 [Sheueling Chang Shantz and Douglas Stebila
1605 (Sun Microsystems Laboratories)]
1607 *) New function EC_GROUP_get_degree, which is defined through
1608 EC_METHOD. For curves over prime fields, this returns the bit
1609 length of the modulus.
1611 [Sheueling Chang Shantz and Douglas Stebila
1612 (Sun Microsystems Laboratories)]
1614 *) New functions EC_GROUP_dup, EC_POINT_dup.
1615 (These simply call ..._new and ..._copy).
1617 [Sheueling Chang Shantz and Douglas Stebila
1618 (Sun Microsystems Laboratories)]
1620 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
1621 Polynomials are represented as BIGNUMs (where the sign bit is not
1622 used) in the following functions [macros]:
1625 BN_GF2m_sub [= BN_GF2m_add]
1626 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
1627 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
1628 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
1630 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
1631 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
1632 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
1633 BN_GF2m_cmp [= BN_ucmp]
1635 (Note that only the 'mod' functions are actually for fields GF(2^m).
1636 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
1638 For some functions, an the irreducible polynomial defining a
1639 field can be given as an 'unsigned int[]' with strictly
1640 decreasing elements giving the indices of those bits that are set;
1641 i.e., p[] represents the polynomial
1642 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
1644 p[0] > p[1] > ... > p[k] = 0.
1645 This applies to the following functions:
1650 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
1651 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
1653 BN_GF2m_mod_sqrt_arr
1654 BN_GF2m_mod_solve_quad_arr
1658 Conversion can be performed by the following functions:
1663 bntest.c has additional tests for binary polynomial arithmetic.
1665 Two implementations for BN_GF2m_mod_div() are available.
1666 The default algorithm simply uses BN_GF2m_mod_inv() and
1667 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
1668 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
1669 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
1671 [Sheueling Chang Shantz and Douglas Stebila
1672 (Sun Microsystems Laboratories)]
1674 *) Add new error code 'ERR_R_DISABLED' that can be used when some
1675 functionality is disabled at compile-time.
1676 [Douglas Stebila <douglas.stebila@sun.com>]
1678 *) Change default behaviour of 'openssl asn1parse' so that more
1679 information is visible when viewing, e.g., a certificate:
1681 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
1682 mode the content of non-printable OCTET STRINGs is output in a
1683 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
1684 avoid the appearance of a printable string.
1685 [Nils Larsch <nla@trustcenter.de>]
1687 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
1689 EC_GROUP_set_asn1_flag()
1690 EC_GROUP_get_asn1_flag()
1691 EC_GROUP_set_point_conversion_form()
1692 EC_GROUP_get_point_conversion_form()
1693 These control ASN1 encoding details:
1694 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
1695 has been set to OPENSSL_EC_NAMED_CURVE.
1696 - Points are encoded in uncompressed form by default; options for
1697 asn1_for are as for point2oct, namely
1698 POINT_CONVERSION_COMPRESSED
1699 POINT_CONVERSION_UNCOMPRESSED
1700 POINT_CONVERSION_HYBRID
1702 Also add 'seed' and 'seed_len' members to EC_GROUP with access
1705 EC_GROUP_get0_seed()
1706 EC_GROUP_get_seed_len()
1707 This is used only for ASN1 purposes (so far).
1708 [Nils Larsch <nla@trustcenter.de>]
1710 *) Add 'field_type' member to EC_METHOD, which holds the NID
1711 of the appropriate field type OID. The new function
1712 EC_METHOD_get_field_type() returns this value.
1713 [Nils Larsch <nla@trustcenter.de>]
1718 EC_POINT_point2hex()
1719 EC_POINT_hex2point()
1720 providing useful interfaces to EC_POINT_point2oct() and
1721 EC_POINT_oct2point().
1722 [Nils Larsch <nla@trustcenter.de>]
1724 *) Change internals of the EC library so that the functions
1725 EC_GROUP_set_generator()
1726 EC_GROUP_get_generator()
1727 EC_GROUP_get_order()
1728 EC_GROUP_get_cofactor()
1729 are implemented directly in crypto/ec/ec_lib.c and not dispatched
1730 to methods, which would lead to unnecessary code duplication when
1731 adding different types of curves.
1732 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
1734 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
1735 arithmetic, and such that modified wNAFs are generated
1736 (which avoid length expansion in many cases).
1739 *) Add a function EC_GROUP_check_discriminant() (defined via
1740 EC_METHOD) that verifies that the curve discriminant is non-zero.
1742 Add a function EC_GROUP_check() that makes some sanity tests
1743 on a EC_GROUP, its generator and order. This includes
1744 EC_GROUP_check_discriminant().
1745 [Nils Larsch <nla@trustcenter.de>]
1747 *) Add ECDSA in new directory crypto/ecdsa/.
1749 Add applications 'openssl ecparam' and 'openssl ecdsa'
1750 (these are based on 'openssl dsaparam' and 'openssl dsa').
1752 ECDSA support is also included in various other files across the
1753 library. Most notably,
1754 - 'openssl req' now has a '-newkey ecdsa:file' option;
1755 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
1756 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
1757 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
1758 them suitable for ECDSA where domain parameters must be
1759 extracted before the specific public key;
1760 - ECDSA engine support has been added.
1761 [Nils Larsch <nla@trustcenter.de>]
1763 *) Include some named elliptic curves, and add OIDs from X9.62,
1764 SECG, and WAP/WTLS. Each curve can be obtained from the new
1766 EC_GROUP_new_by_curve_name(),
1767 and the list of available named curves can be obtained with
1768 EC_get_builtin_curves().
1769 Also add a 'curve_name' member to EC_GROUP objects, which can be
1771 EC_GROUP_set_curve_name()
1772 EC_GROUP_get_curve_name()
1773 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
1775 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1776 was actually never needed) and in BN_mul(). The removal in BN_mul()
1777 required a small change in bn_mul_part_recursive() and the addition
1778 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1779 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1780 bn_sub_words() and bn_add_words() except they take arrays with
1784 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
1786 *) Cleanse PEM buffers before freeing them since they may contain
1788 [Benjamin Bennett <ben@psc.edu>]
1790 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1791 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1792 authentication-only ciphersuites.
1795 *) Since AES128 and AES256 share a single mask bit in the logic of
1796 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1797 kludge to work properly if AES128 is available and AES256 isn't.
1800 *) Expand security boundary to match 1.1.1 module.
1803 *) Remove redundant features: hash file source, editing of test vectors
1804 modify fipsld to use external fips_premain.c signature.
1807 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
1808 run algorithm test programs.
1811 *) Make algorithm test programs more tolerant of whitespace.
1814 *) Have SSL/TLS server implementation tolerate "mismatched" record
1815 protocol version while receiving ClientHello even if the
1816 ClientHello is fragmented. (The server can't insist on the
1817 particular protocol version it has chosen before the ServerHello
1818 message has informed the client about his choice.)
1821 *) Load error codes if they are not already present instead of using a
1822 static variable. This allows them to be cleanly unloaded and reloaded.
1825 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
1827 *) Introduce limits to prevent malicious keys being able to
1828 cause a denial of service. (CVE-2006-2940)
1829 [Steve Henson, Bodo Moeller]
1831 *) Fix ASN.1 parsing of certain invalid structures that can result
1832 in a denial of service. (CVE-2006-2937) [Steve Henson]
1834 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1835 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1837 *) Fix SSL client code which could crash if connecting to a
1838 malicious SSLv2 server. (CVE-2006-4343)
1839 [Tavis Ormandy and Will Drewry, Google Security Team]
1841 *) Change ciphersuite string processing so that an explicit
1842 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
1843 will no longer include "AES128-SHA"), and any other similar
1844 ciphersuite (same bitmap) from *other* protocol versions (so that
1845 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
1846 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
1847 changes from 0.9.8b and 0.9.8d.
1850 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
1852 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1853 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1855 *) Change the Unix randomness entropy gathering to use poll() when
1856 possible instead of select(), since the latter has some
1857 undesirable limitations.
1858 [Darryl Miles via Richard Levitte and Bodo Moeller]
1860 *) Disable rogue ciphersuites:
1862 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1863 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1864 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1866 The latter two were purportedly from
1867 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1870 Also deactive the remaining ciphersuites from
1871 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1872 unofficial, and the ID has long expired.
1875 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1876 dual-core machines) and other potential thread-safety issues.
1879 Changes between 0.9.7i and 0.9.7j [04 May 2006]
1881 *) Adapt fipsld and the build system to link against the validated FIPS
1882 module in FIPS mode.
1885 *) Fixes for VC++ 2005 build under Windows.
1888 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
1889 from a Windows bash shell such as MSYS. It is autodetected from the
1890 "config" script when run from a VC++ environment. Modify standard VC++
1891 build to use fipscanister.o from the GNU make build.
1894 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
1896 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
1897 The value now differs depending on if you build for FIPS or not.
1898 BEWARE! A program linked with a shared FIPSed libcrypto can't be
1899 safely run with a non-FIPSed libcrypto, as it may crash because of
1900 the difference induced by this change.
1903 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
1905 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1906 (part of SSL_OP_ALL). This option used to disable the
1907 countermeasure against man-in-the-middle protocol-version
1908 rollback in the SSL 2.0 server implementation, which is a bad
1909 idea. (CVE-2005-2969)
1911 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1912 for Information Security, National Institute of Advanced Industrial
1913 Science and Technology [AIST], Japan)]
1915 *) Minimal support for X9.31 signatures and PSS padding modes. This is
1916 mainly for FIPS compliance and not fully integrated at this stage.
1919 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
1920 the exponentiation using a fixed-length exponent. (Otherwise,
1921 the information leaked through timing could expose the secret key
1922 after many signatures; cf. Bleichenbacher's attack on DSA with
1926 *) Make a new fixed-window mod_exp implementation the default for
1927 RSA, DSA, and DH private-key operations so that the sequence of
1928 squares and multiplies and the memory access pattern are
1929 independent of the particular secret key. This will mitigate
1930 cache-timing and potential related attacks.
1932 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
1933 and this is automatically used by BN_mod_exp_mont() if the new flag
1934 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
1935 will use this BN flag for private exponents unless the flag
1936 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
1937 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
1939 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
1941 *) Change the client implementation for SSLv23_method() and
1942 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
1943 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
1944 (Previously, the SSL 2.0 backwards compatible Client Hello
1945 message format would be used even with SSL_OP_NO_SSLv2.)
1948 *) Add support for smime-type MIME parameter in S/MIME messages which some
1952 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
1953 a threadsafe manner. Modify rsa code to use new function and add calls
1954 to dsa and dh code (which had race conditions before).
1957 *) Include the fixed error library code in the C error file definitions
1958 instead of fixing them up at runtime. This keeps the error code
1959 structures constant.
1962 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
1964 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
1967 *) Fixes for newer kerberos headers. NB: the casts are needed because
1968 the 'length' field is signed on one version and unsigned on another
1969 with no (?) obvious way to tell the difference, without these VC++
1970 complains. Also the "definition" of FAR (blank) is no longer included
1971 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
1972 some needed definitions.
1975 *) Undo Cygwin change.
1978 *) Added support for proxy certificates according to RFC 3820.
1979 Because they may be a security thread to unaware applications,
1980 they must be explicitely allowed in run-time. See
1981 docs/HOWTO/proxy_certificates.txt for further information.
1984 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
1986 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
1987 server and client random values. Previously
1988 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
1989 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
1991 This change has negligible security impact because:
1993 1. Server and client random values still have 24 bytes of pseudo random
1996 2. Server and client random values are sent in the clear in the initial
1999 3. The master secret is derived using the premaster secret (48 bytes in
2000 size for static RSA ciphersuites) as well as client server and random
2003 The OpenSSL team would like to thank the UK NISCC for bringing this issue
2006 [Stephen Henson, reported by UK NISCC]
2008 *) Use Windows randomness collection on Cygwin.
2011 *) Fix hang in EGD/PRNGD query when communication socket is closed
2012 prematurely by EGD/PRNGD.
2013 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
2015 *) Prompt for pass phrases when appropriate for PKCS12 input format.
2018 *) Back-port of selected performance improvements from development
2019 branch, as well as improved support for PowerPC platforms.
2022 *) Add lots of checks for memory allocation failure, error codes to indicate
2023 failure and freeing up memory if a failure occurs.
2024 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
2026 *) Add new -passin argument to dgst.
2029 *) Perform some character comparisons of different types in X509_NAME_cmp:
2030 this is needed for some certificates that reencode DNs into UTF8Strings
2031 (in violation of RFC3280) and can't or wont issue name rollover
2035 *) Make an explicit check during certificate validation to see that
2036 the CA setting in each certificate on the chain is correct. As a
2037 side effect always do the following basic checks on extensions,
2038 not just when there's an associated purpose to the check:
2040 - if there is an unhandled critical extension (unless the user
2041 has chosen to ignore this fault)
2042 - if the path length has been exceeded (if one is set at all)
2043 - that certain extensions fit the associated purpose (if one has
2047 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
2049 *) Avoid a race condition when CRLs are checked in a multi threaded
2050 environment. This would happen due to the reordering of the revoked
2051 entries during signature checking and serial number lookup. Now the
2052 encoding is cached and the serial number sort performed under a lock.
2053 Add new STACK function sk_is_sorted().
2056 *) Add Delta CRL to the extension code.
2059 *) Various fixes to s3_pkt.c so alerts are sent properly.
2060 [David Holmes <d.holmes@f5.com>]
2062 *) Reduce the chances of duplicate issuer name and serial numbers (in
2063 violation of RFC3280) using the OpenSSL certificate creation utilities.
2064 This is done by creating a random 64 bit value for the initial serial
2065 number when a serial number file is created or when a self signed
2066 certificate is created using 'openssl req -x509'. The initial serial
2067 number file is created using 'openssl x509 -next_serial' in CA.pl
2068 rather than being initialized to 1.
2071 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
2073 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
2074 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
2075 [Joe Orton, Steve Henson]
2077 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
2079 [Joe Orton, Steve Henson]
2081 *) Make it possible to have multiple active certificates with the same
2082 subject in the CA index file. This is done only if the keyword
2083 'unique_subject' is set to 'no' in the main CA section (default
2084 if 'CA_default') of the configuration file. The value is saved
2085 with the database itself in a separate index attribute file,
2086 named like the index file with '.attr' appended to the name.
2089 *) X509 verify fixes. Disable broken certificate workarounds when
2090 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
2091 keyUsage extension present. Don't accept CRLs with unhandled critical
2092 extensions: since verify currently doesn't process CRL extensions this
2093 rejects a CRL with *any* critical extensions. Add new verify error codes
2097 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
2098 A clarification of RFC2560 will require the use of OCTET STRINGs and
2099 some implementations cannot handle the current raw format. Since OpenSSL
2100 copies and compares OCSP nonces as opaque blobs without any attempt at
2101 parsing them this should not create any compatibility issues.
2104 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
2105 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
2106 this HMAC (and other) operations are several times slower than OpenSSL
2110 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
2111 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2113 *) Use the correct content when signing type "other".
2116 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
2118 *) Fix various bugs revealed by running the NISCC test suite:
2120 Stop out of bounds reads in the ASN1 code when presented with
2121 invalid tags (CVE-2003-0543 and CVE-2003-0544).
2123 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
2125 If verify callback ignores invalid public key errors don't try to check
2126 certificate signature with the NULL public key.
2130 *) New -ignore_err option in ocsp application to stop the server
2131 exiting on the first error in a request.
2134 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
2135 if the server requested one: as stated in TLS 1.0 and SSL 3.0
2139 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
2140 extra data after the compression methods not only for TLS 1.0
2141 but also for SSL 3.0 (as required by the specification).
2142 [Bodo Moeller; problem pointed out by Matthias Loepfe]
2144 *) Change X509_certificate_type() to mark the key as exported/exportable
2145 when it's 512 *bits* long, not 512 bytes.
2148 *) Change AES_cbc_encrypt() so it outputs exact multiple of
2149 blocks during encryption.
2152 *) Various fixes to base64 BIO and non blocking I/O. On write
2153 flushes were not handled properly if the BIO retried. On read
2154 data was not being buffered properly and had various logic bugs.
2155 This also affects blocking I/O when the data being decoded is a
2159 *) Various S/MIME bugfixes and compatibility changes:
2160 output correct application/pkcs7 MIME type if
2161 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
2162 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
2163 of files as .eml work). Correctly handle very long lines in MIME
2167 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
2169 *) Countermeasure against the Klima-Pokorny-Rosa extension of
2170 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
2171 a protocol version number mismatch like a decryption error
2172 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
2175 *) Turn on RSA blinding by default in the default implementation
2176 to avoid a timing attack. Applications that don't want it can call
2177 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
2178 They would be ill-advised to do so in most cases.
2179 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
2181 *) Change RSA blinding code so that it works when the PRNG is not
2182 seeded (in this case, the secret RSA exponent is abused as
2183 an unpredictable seed -- if it is not unpredictable, there
2184 is no point in blinding anyway). Make RSA blinding thread-safe
2185 by remembering the creator's thread ID in rsa->blinding and
2186 having all other threads use local one-time blinding factors
2187 (this requires more computation than sharing rsa->blinding, but
2188 avoids excessive locking; and if an RSA object is not shared
2189 between threads, blinding will still be very fast).
2192 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
2193 ENGINE as defaults for all supported algorithms irrespective of
2194 the 'flags' parameter. 'flags' is now honoured, so applications
2195 should make sure they are passing it correctly.
2198 *) Target "mingw" now allows native Windows code to be generated in
2199 the Cygwin environment as well as with the MinGW compiler.
2202 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
2204 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
2205 via timing by performing a MAC computation even if incorrrect
2206 block cipher padding has been found. This is a countermeasure
2207 against active attacks where the attacker has to distinguish
2208 between bad padding and a MAC verification error. (CVE-2003-0078)
2210 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
2211 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
2212 Martin Vuagnoux (EPFL, Ilion)]
2214 *) Make the no-err option work as intended. The intention with no-err
2215 is not to have the whole error stack handling routines removed from
2216 libcrypto, it's only intended to remove all the function name and
2217 reason texts, thereby removing some of the footprint that may not
2218 be interesting if those errors aren't displayed anyway.
2220 NOTE: it's still possible for any application or module to have it's
2221 own set of error texts inserted. The routines are there, just not
2222 used by default when no-err is given.
2225 *) Add support for FreeBSD on IA64.
2226 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
2228 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
2229 Kerberos function mit_des_cbc_cksum(). Before this change,
2230 the value returned by DES_cbc_cksum() was like the one from
2231 mit_des_cbc_cksum(), except the bytes were swapped.
2232 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
2234 *) Allow an application to disable the automatic SSL chain building.
2235 Before this a rather primitive chain build was always performed in
2236 ssl3_output_cert_chain(): an application had no way to send the
2237 correct chain if the automatic operation produced an incorrect result.
2239 Now the chain builder is disabled if either:
2241 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
2243 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
2245 The reasoning behind this is that an application would not want the
2246 auto chain building to take place if extra chain certificates are
2247 present and it might also want a means of sending no additional
2248 certificates (for example the chain has two certificates and the
2252 *) Add the possibility to build without the ENGINE framework.
2253 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2255 *) Under Win32 gmtime() can return NULL: check return value in
2256 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
2259 *) DSA routines: under certain error conditions uninitialized BN objects
2260 could be freed. Solution: make sure initialization is performed early
2261 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
2262 Nils Larsch <nla@trustcenter.de> via PR#459)
2265 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
2266 checked on reconnect on the client side, therefore session resumption
2267 could still fail with a "ssl session id is different" error. This
2268 behaviour is masked when SSL_OP_ALL is used due to
2269 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
2270 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2271 followup to PR #377.
2274 *) IA-32 assembler support enhancements: unified ELF targets, support
2275 for SCO/Caldera platforms, fix for Cygwin shared build.
2278 *) Add support for FreeBSD on sparc64. As a consequence, support for
2279 FreeBSD on non-x86 processors is separate from x86 processors on
2280 the config script, much like the NetBSD support.
2281 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
2283 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
2285 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
2288 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
2289 code (06) was taken as the first octet of the session ID and the last
2290 octet was ignored consequently. As a result SSLv2 client side session
2291 caching could not have worked due to the session ID mismatch between
2293 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2297 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
2298 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
2302 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
2303 seems that in spite of existing for more than a year, many application
2304 author have done nothing to provide the necessary callbacks, which
2305 means that this particular engine will not work properly anywhere.
2306 This is a very unfortunate situation which forces us, in the name
2307 of usability, to give the hw_ncipher.c a static lock, which is part
2309 NOTE: This is for the 0.9.7 series ONLY. This hack will never
2310 appear in 0.9.8 or later. We EXPECT application authors to have
2311 dealt properly with this when 0.9.8 is released (unless we actually
2312 make such changes in the libcrypto locking code that changes will
2313 have to be made anyway).
2316 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
2317 octets have been read, EOF or an error occurs. Without this change
2318 some truncated ASN1 structures will not produce an error.
2321 *) Disable Heimdal support, since it hasn't been fully implemented.
2322 Still give the possibility to force the use of Heimdal, but with
2323 warnings and a request that patches get sent to openssl-dev.
2326 *) Add the VC-CE target, introduce the WINCE sysname, and add
2327 INSTALL.WCE and appropriate conditionals to make it build.
2328 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2330 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
2331 cygssl-x.y.z.dll, where x, y and z are the major, minor and
2332 edit numbers of the version.
2333 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2335 *) Introduce safe string copy and catenation functions
2336 (BUF_strlcpy() and BUF_strlcat()).
2337 [Ben Laurie (CHATS) and Richard Levitte]
2339 *) Avoid using fixed-size buffers for one-line DNs.
2340 [Ben Laurie (CHATS)]
2342 *) Add BUF_MEM_grow_clean() to avoid information leakage when
2343 resizing buffers containing secrets, and use where appropriate.
2344 [Ben Laurie (CHATS)]
2346 *) Avoid using fixed size buffers for configuration file location.
2347 [Ben Laurie (CHATS)]
2349 *) Avoid filename truncation for various CA files.
2350 [Ben Laurie (CHATS)]
2352 *) Use sizeof in preference to magic numbers.
2353 [Ben Laurie (CHATS)]
2355 *) Avoid filename truncation in cert requests.
2356 [Ben Laurie (CHATS)]
2358 *) Add assertions to check for (supposedly impossible) buffer
2360 [Ben Laurie (CHATS)]
2362 *) Don't cache truncated DNS entries in the local cache (this could
2363 potentially lead to a spoofing attack).
2364 [Ben Laurie (CHATS)]
2366 *) Fix various buffers to be large enough for hex/decimal
2367 representations in a platform independent manner.
2368 [Ben Laurie (CHATS)]
2370 *) Add CRYPTO_realloc_clean() to avoid information leakage when
2371 resizing buffers containing secrets, and use where appropriate.
2372 [Ben Laurie (CHATS)]
2374 *) Add BIO_indent() to avoid much slightly worrying code to do
2376 [Ben Laurie (CHATS)]
2378 *) Convert sprintf()/BIO_puts() to BIO_printf().
2379 [Ben Laurie (CHATS)]
2381 *) buffer_gets() could terminate with the buffer only half
2383 [Ben Laurie (CHATS)]
2385 *) Add assertions to prevent user-supplied crypto functions from
2386 overflowing internal buffers by having large block sizes, etc.
2387 [Ben Laurie (CHATS)]
2389 *) New OPENSSL_assert() macro (similar to assert(), but enabled
2391 [Ben Laurie (CHATS)]
2393 *) Eliminate unused copy of key in RC4.
2394 [Ben Laurie (CHATS)]
2396 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
2397 [Ben Laurie (CHATS)]
2399 *) Fix off-by-one error in EGD path.
2400 [Ben Laurie (CHATS)]
2402 *) If RANDFILE path is too long, ignore instead of truncating.
2403 [Ben Laurie (CHATS)]
2405 *) Eliminate unused and incorrectly sized X.509 structure
2407 [Ben Laurie (CHATS)]
2409 *) Eliminate unused and dangerous function knumber().
2410 [Ben Laurie (CHATS)]
2412 *) Eliminate unused and dangerous structure, KSSL_ERR.
2413 [Ben Laurie (CHATS)]
2415 *) Protect against overlong session ID context length in an encoded
2416 session object. Since these are local, this does not appear to be
2418 [Ben Laurie (CHATS)]
2420 *) Change from security patch (see 0.9.6e below) that did not affect
2421 the 0.9.6 release series:
2423 Remote buffer overflow in SSL3 protocol - an attacker could
2424 supply an oversized master key in Kerberos-enabled versions.
2426 [Ben Laurie (CHATS)]
2428 *) Change the SSL kerb5 codes to match RFC 2712.
2431 *) Make -nameopt work fully for req and add -reqopt switch.
2432 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
2434 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
2435 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
2437 *) Make sure tests can be performed even if the corresponding algorithms
2438 have been removed entirely. This was also the last step to make
2439 OpenSSL compilable with DJGPP under all reasonable conditions.
2440 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
2442 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
2443 to allow version independent disabling of normally unselected ciphers,
2444 which may be activated as a side-effect of selecting a single cipher.
2446 (E.g., cipher list string "RSA" enables ciphersuites that are left
2447 out of "ALL" because they do not provide symmetric encryption.
2448 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
2449 [Lutz Jaenicke, Bodo Moeller]
2451 *) Add appropriate support for separate platform-dependent build
2452 directories. The recommended way to make a platform-dependent
2453 build directory is the following (tested on Linux), maybe with
2456 # Place yourself outside of the OpenSSL source tree. In
2457 # this example, the environment variable OPENSSL_SOURCE
2458 # is assumed to contain the absolute OpenSSL source directory.
2459 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
2460 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
2461 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
2462 mkdir -p `dirname $F`
2463 ln -s $OPENSSL_SOURCE/$F $F
2466 To be absolutely sure not to disturb the source tree, a "make clean"
2467 is a good thing. If it isn't successfull, don't worry about it,
2468 it probably means the source directory is very clean.
2471 *) Make sure any ENGINE control commands make local copies of string
2472 pointers passed to them whenever necessary. Otherwise it is possible
2473 the caller may have overwritten (or deallocated) the original string
2474 data when a later ENGINE operation tries to use the stored values.
2475 [Götz Babin-Ebell <babinebell@trustcenter.de>]
2477 *) Improve diagnostics in file reading and command-line digests.
2478 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
2480 *) Add AES modes CFB and OFB to the object database. Correct an
2481 error in AES-CFB decryption.
2484 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
2485 allows existing EVP_CIPHER_CTX structures to be reused after
2486 calling EVP_*Final(). This behaviour is used by encryption
2487 BIOs and some applications. This has the side effect that
2488 applications must explicitly clean up cipher contexts with
2489 EVP_CIPHER_CTX_cleanup() or they will leak memory.
2492 *) Check the values of dna and dnb in bn_mul_recursive before calling
2493 bn_mul_comba (a non zero value means the a or b arrays do not contain
2494 n2 elements) and fallback to bn_mul_normal if either is not zero.
2497 *) Fix escaping of non-ASCII characters when using the -subj option
2498 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
2501 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
2502 form for "surname", serialNumber has no short form.
2503 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
2504 therefore remove "mail" short name for "internet 7".
2505 The OID for unique identifiers in X509 certificates is
2506 x500UniqueIdentifier, not uniqueIdentifier.
2507 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
2510 *) Add an "init" command to the ENGINE config module and auto initialize
2511 ENGINEs. Without any "init" command the ENGINE will be initialized
2512 after all ctrl commands have been executed on it. If init=1 the
2513 ENGINE is initailized at that point (ctrls before that point are run
2514 on the uninitialized ENGINE and after on the initialized one). If
2515 init=0 then the ENGINE will not be iniatialized at all.
2518 *) Fix the 'app_verify_callback' interface so that the user-defined
2519 argument is actually passed to the callback: In the
2520 SSL_CTX_set_cert_verify_callback() prototype, the callback
2521 declaration has been changed from
2524 int (*cb)(X509_STORE_CTX *,void *);
2525 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
2526 i=s->ctx->app_verify_callback(&ctx)
2527 has been changed into
2528 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
2530 To update applications using SSL_CTX_set_cert_verify_callback(),
2531 a dummy argument can be added to their callback functions.
2532 [D. K. Smetters <smetters@parc.xerox.com>]
2534 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
2535 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
2537 *) Add and OPENSSL_LOAD_CONF define which will cause
2538 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
2539 This allows older applications to transparently support certain
2540 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
2541 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
2542 load the config file and OPENSSL_add_all_algorithms_conf() which will
2543 always load it have also been added.
2546 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
2547 Adjust NIDs and EVP layer.
2548 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2550 *) Config modules support in openssl utility.
2552 Most commands now load modules from the config file,
2553 though in a few (such as version) this isn't done
2554 because it couldn't be used for anything.
2556 In the case of ca and req the config file used is
2557 the same as the utility itself: that is the -config
2558 command line option can be used to specify an
2562 *) Move default behaviour from OPENSSL_config(). If appname is NULL
2563 use "openssl_conf" if filename is NULL use default openssl config file.
2566 *) Add an argument to OPENSSL_config() to allow the use of an alternative
2567 config section name. Add a new flag to tolerate a missing config file
2568 and move code to CONF_modules_load_file().
2571 *) Support for crypto accelerator cards from Accelerated Encryption
2572 Processing, www.aep.ie. (Use engine 'aep')
2573 The support was copied from 0.9.6c [engine] and adapted/corrected
2574 to work with the new engine framework.
2575 [AEP Inc. and Richard Levitte]
2577 *) Support for SureWare crypto accelerator cards from Baltimore
2578 Technologies. (Use engine 'sureware')
2579 The support was copied from 0.9.6c [engine] and adapted
2580 to work with the new engine framework.
2583 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
2584 make the newer ENGINE framework commands for the CHIL engine work.
2585 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
2587 *) Make it possible to produce shared libraries on ReliantUNIX.
2588 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
2590 *) Add the configuration target debug-linux-ppro.
2591 Make 'openssl rsa' use the general key loading routines
2592 implemented in apps.c, and make those routines able to
2593 handle the key format FORMAT_NETSCAPE and the variant
2595 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2597 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2598 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2600 *) Add -keyform to rsautl, and document -engine.
2601 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
2603 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
2604 BIO_R_NO_SUCH_FILE error code rather than the generic
2605 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
2608 *) Add new functions
2610 ERR_peek_last_error_line
2611 ERR_peek_last_error_line_data.
2612 These are similar to
2615 ERR_peek_error_line_data,
2616 but report on the latest error recorded rather than the first one
2617 still in the error queue.
2618 [Ben Laurie, Bodo Moeller]
2620 *) default_algorithms option in ENGINE config module. This allows things
2622 default_algorithms = ALL
2623 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
2626 *) Prelminary ENGINE config module.
2629 *) New experimental application configuration code.
2632 *) Change the AES code to follow the same name structure as all other
2633 symmetric ciphers, and behave the same way. Move everything to
2634 the directory crypto/aes, thereby obsoleting crypto/rijndael.
2635 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2637 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
2638 [Ben Laurie and Theo de Raadt]
2640 *) Add option to output public keys in req command.
2641 [Massimiliano Pala madwolf@openca.org]
2643 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
2644 (up to about 10% better than before for P-192 and P-224).
2647 *) New functions/macros
2649 SSL_CTX_set_msg_callback(ctx, cb)
2650 SSL_CTX_set_msg_callback_arg(ctx, arg)
2651 SSL_set_msg_callback(ssl, cb)
2652 SSL_set_msg_callback_arg(ssl, arg)
2654 to request calling a callback function
2656 void cb(int write_p, int version, int content_type,
2657 const void *buf, size_t len, SSL *ssl, void *arg)
2659 whenever a protocol message has been completely received
2660 (write_p == 0) or sent (write_p == 1). Here 'version' is the
2661 protocol version according to which the SSL library interprets
2662 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
2663 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
2664 the content type as defined in the SSL 3.0/TLS 1.0 protocol
2665 specification (change_cipher_spec(20), alert(21), handshake(22)).
2666 'buf' and 'len' point to the actual message, 'ssl' to the
2667 SSL object, and 'arg' is the application-defined value set by
2668 SSL[_CTX]_set_msg_callback_arg().
2670 'openssl s_client' and 'openssl s_server' have new '-msg' options
2671 to enable a callback that displays all protocol messages.
2674 *) Change the shared library support so shared libraries are built as
2675 soon as the corresponding static library is finished, and thereby get
2676 openssl and the test programs linked against the shared library.
2677 This still only happens when the keyword "shard" has been given to
2678 the configuration scripts.
2680 NOTE: shared library support is still an experimental thing, and
2681 backward binary compatibility is still not guaranteed.
2682 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
2684 *) Add support for Subject Information Access extension.
2685 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2687 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
2688 additional bytes when new memory had to be allocated, not just
2689 when reusing an existing buffer.
2692 *) New command line and configuration option 'utf8' for the req command.
2693 This allows field values to be specified as UTF8 strings.
2696 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
2697 runs for the former and machine-readable output for the latter.
2700 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
2701 of the e-mail address in the DN (i.e., it will go into a certificate
2702 extension only). The new configuration file option 'email_in_dn = no'
2703 has the same effect.
2704 [Massimiliano Pala madwolf@openca.org]
2706 *) Change all functions with names starting with des_ to be starting
2707 with DES_ instead. Add wrappers that are compatible with libdes,
2708 but are named _ossl_old_des_*. Finally, add macros that map the
2709 des_* symbols to the corresponding _ossl_old_des_* if libdes
2710 compatibility is desired. If OpenSSL 0.9.6c compatibility is
2711 desired, the des_* symbols will be mapped to DES_*, with one
2714 Since we provide two compatibility mappings, the user needs to
2715 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
2716 compatibility is desired. The default (i.e., when that macro
2717 isn't defined) is OpenSSL 0.9.6c compatibility.
2719 There are also macros that enable and disable the support of old
2720 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
2721 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
2722 are defined, the default will apply: to support the old des routines.
2724 In either case, one must include openssl/des.h to get the correct
2725 definitions. Do not try to just include openssl/des_old.h, that
2728 NOTE: This is a major break of an old API into a new one. Software
2729 authors are encouraged to switch to the DES_ style functions. Some
2730 time in the future, des_old.h and the libdes compatibility functions
2731 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
2732 default), and then completely removed.
2735 *) Test for certificates which contain unsupported critical extensions.
2736 If such a certificate is found during a verify operation it is
2737 rejected by default: this behaviour can be overridden by either
2738 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
2739 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
2740 X509_supported_extension() has also been added which returns 1 if a
2741 particular extension is supported.
2744 *) Modify the behaviour of EVP cipher functions in similar way to digests
2745 to retain compatibility with existing code.
2748 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
2749 compatibility with existing code. In particular the 'ctx' parameter does
2750 not have to be to be initialized before the call to EVP_DigestInit() and
2751 it is tidied up after a call to EVP_DigestFinal(). New function
2752 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
2753 EVP_MD_CTX_copy() changed to not require the destination to be
2754 initialized valid and new function EVP_MD_CTX_copy_ex() added which
2755 requires the destination to be valid.
2757 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
2758 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
2761 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
2762 so that complete 'Handshake' protocol structures are kept in memory
2763 instead of overwriting 'msg_type' and 'length' with 'body' data.
2766 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
2767 [Massimo Santin via Richard Levitte]
2769 *) Major restructuring to the underlying ENGINE code. This includes
2770 reduction of linker bloat, separation of pure "ENGINE" manipulation
2771 (initialisation, etc) from functionality dealing with implementations
2772 of specific crypto iterfaces. This change also introduces integrated
2773 support for symmetric ciphers and digest implementations - so ENGINEs
2774 can now accelerate these by providing EVP_CIPHER and EVP_MD
2775 implementations of their own. This is detailed in crypto/engine/README
2776 as it couldn't be adequately described here. However, there are a few
2777 API changes worth noting - some RSA, DSA, DH, and RAND functions that
2778 were changed in the original introduction of ENGINE code have now
2779 reverted back - the hooking from this code to ENGINE is now a good
2780 deal more passive and at run-time, operations deal directly with
2781 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
2782 dereferencing through an ENGINE pointer any more. Also, the ENGINE
2783 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
2784 they were not being used by the framework as there is no concept of a
2785 BIGNUM_METHOD and they could not be generalised to the new
2786 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
2787 ENGINE_cpy() has been removed as it cannot be consistently defined in
2791 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
2794 *) Change mkdef.pl to sort symbols that get the same entry number,
2795 and make sure the automatically generated functions ERR_load_*
2796 become part of libeay.num as well.
2799 *) New function SSL_renegotiate_pending(). This returns true once
2800 renegotiation has been requested (either SSL_renegotiate() call
2801 or HelloRequest/ClientHello receveived from the peer) and becomes
2802 false once a handshake has been completed.
2803 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
2804 sends a HelloRequest, but does not ensure that a handshake takes
2805 place. SSL_renegotiate_pending() is useful for checking if the
2806 client has followed the request.)
2809 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
2810 By default, clients may request session resumption even during
2811 renegotiation (if session ID contexts permit); with this option,
2812 session resumption is possible only in the first handshake.
2814 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
2815 more bits available for options that should not be part of
2816 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
2819 *) Add some demos for certificate and certificate request creation.
2822 *) Make maximum certificate chain size accepted from the peer application
2823 settable (SSL*_get/set_max_cert_list()), as proposed by
2824 "Douglas E. Engert" <deengert@anl.gov>.
2827 *) Add support for shared libraries for Unixware-7
2828 (Boyd Lynn Gerber <gerberb@zenez.com>).
2831 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
2832 be done prior to destruction. Use this to unload error strings from
2833 ENGINEs that load their own error strings. NB: This adds two new API
2834 functions to "get" and "set" this destroy handler in an ENGINE.
2837 *) Alter all existing ENGINE implementations (except "openssl" and
2838 "openbsd") to dynamically instantiate their own error strings. This
2839 makes them more flexible to be built both as statically-linked ENGINEs
2840 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
2841 Also, add stub code to each that makes building them as self-contained
2842 shared-libraries easier (see README.ENGINE).
2845 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
2846 implementations into applications that are completely implemented in
2847 self-contained shared-libraries. The "dynamic" ENGINE exposes control
2848 commands that can be used to configure what shared-library to load and
2849 to control aspects of the way it is handled. Also, made an update to
2850 the README.ENGINE file that brings its information up-to-date and
2851 provides some information and instructions on the "dynamic" ENGINE
2852 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
2855 *) Make it possible to unload ranges of ERR strings with a new
2856 "ERR_unload_strings" function.
2859 *) Add a copy() function to EVP_MD.
2862 *) Make EVP_MD routines take a context pointer instead of just the
2863 md_data void pointer.
2866 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
2867 that the digest can only process a single chunk of data
2868 (typically because it is provided by a piece of
2869 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
2870 is only going to provide a single chunk of data, and hence the
2871 framework needn't accumulate the data for oneshot drivers.
2874 *) As with "ERR", make it possible to replace the underlying "ex_data"
2875 functions. This change also alters the storage and management of global
2876 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
2877 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
2878 index counters. The API functions that use this state have been changed
2879 to take a "class_index" rather than pointers to the class's local STACK
2880 and counter, and there is now an API function to dynamically create new
2881 classes. This centralisation allows us to (a) plug a lot of the
2882 thread-safety problems that existed, and (b) makes it possible to clean
2883 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
2884 such data would previously have always leaked in application code and
2885 workarounds were in place to make the memory debugging turn a blind eye
2886 to it. Application code that doesn't use this new function will still
2887 leak as before, but their memory debugging output will announce it now
2888 rather than letting it slide.
2890 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
2891 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
2892 has a return value to indicate success or failure.
2895 *) Make it possible to replace the underlying "ERR" functions such that the
2896 global state (2 LHASH tables and 2 locks) is only used by the "default"
2897 implementation. This change also adds two functions to "get" and "set"
2898 the implementation prior to it being automatically set the first time
2899 any other ERR function takes place. Ie. an application can call "get",
2900 pass the return value to a module it has just loaded, and that module
2901 can call its own "set" function using that value. This means the
2902 module's "ERR" operations will use (and modify) the error state in the
2903 application and not in its own statically linked copy of OpenSSL code.
2906 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
2907 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
2908 the operation, and provides a more encapsulated way for external code
2909 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
2910 to use these functions rather than manually incrementing the counts.
2912 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
2915 *) Add EVP test program.
2918 *) Add symmetric cipher support to ENGINE. Expect the API to change!
2921 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
2922 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
2923 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
2924 These allow a CRL to be built without having to access X509_CRL fields
2925 directly. Modify 'ca' application to use new functions.
2928 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
2929 bug workarounds. Rollback attack detection is a security feature.
2930 The problem will only arise on OpenSSL servers when TLSv1 is not
2931 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
2932 Software authors not wanting to support TLSv1 will have special reasons
2933 for their choice and can explicitly enable this option.
2934 [Bodo Moeller, Lutz Jaenicke]
2936 *) Rationalise EVP so it can be extended: don't include a union of
2937 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
2938 (similar to those existing for EVP_CIPHER_CTX).
2943 EVP_MD_CTX_init(&md); /* new function call */
2944 EVP_DigestInit(&md, EVP_sha1());
2945 EVP_DigestUpdate(&md, in, len);
2946 EVP_DigestFinal(&md, out, NULL);
2947 EVP_MD_CTX_cleanup(&md); /* new function call */
2951 *) Make DES key schedule conform to the usual scheme, as well as
2952 correcting its structure. This means that calls to DES functions
2953 now have to pass a pointer to a des_key_schedule instead of a
2954 plain des_key_schedule (which was actually always a pointer
2957 des_key_schedule ks;
2959 des_set_key_checked(..., &ks);
2960 des_ncbc_encrypt(..., &ks, ...);
2962 (Note that a later change renames 'des_...' into 'DES_...'.)
2965 *) Initial reduction of linker bloat: the use of some functions, such as
2966 PEM causes large amounts of unused functions to be linked in due to
2967 poor organisation. For example pem_all.c contains every PEM function
2968 which has a knock on effect of linking in large amounts of (unused)
2969 ASN1 code. Grouping together similar functions and splitting unrelated
2970 functions prevents this.
2973 *) Cleanup of EVP macros.
2976 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
2977 correct _ecb suffix.
2980 *) Add initial OCSP responder support to ocsp application. The
2981 revocation information is handled using the text based index
2982 use by the ca application. The responder can either handle
2983 requests generated internally, supplied in files (for example
2984 via a CGI script) or using an internal minimal server.
2987 *) Add configuration choices to get zlib compression for TLS.
2990 *) Changes to Kerberos SSL for RFC 2712 compliance:
2991 1. Implemented real KerberosWrapper, instead of just using
2992 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
2993 2. Implemented optional authenticator field of KerberosWrapper.
2995 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
2996 and authenticator structs; see crypto/krb5/.
2998 Generalized Kerberos calls to support multiple Kerberos libraries.
2999 [Vern Staats <staatsvr@asc.hpc.mil>,
3000 Jeffrey Altman <jaltman@columbia.edu>
3001 via Richard Levitte]
3003 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
3004 already does with RSA. testdsa.h now has 'priv_key/pub_key'
3005 values for each of the key sizes rather than having just
3006 parameters (and 'speed' generating keys each time).
3009 *) Speed up EVP routines.
3012 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
3013 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
3014 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
3015 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
3017 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
3018 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
3019 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
3022 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
3024 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
3027 *) Added the OS2-EMX target.
3028 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
3030 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
3031 to support NCONF routines in extension code. New function CONF_set_nconf()
3032 to allow functions which take an NCONF to also handle the old LHASH
3033 structure: this means that the old CONF compatible routines can be
3034 retained (in particular wrt extensions) without having to duplicate the
3035 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
3038 *) Enhance the general user interface with mechanisms for inner control
3039 and with possibilities to have yes/no kind of prompts.
3042 *) Change all calls to low level digest routines in the library and
3043 applications to use EVP. Add missing calls to HMAC_cleanup() and
3044 don't assume HMAC_CTX can be copied using memcpy().
3045 [Verdon Walker <VWalker@novell.com>, Steve Henson]
3047 *) Add the possibility to control engines through control names but with
3048 arbitrary arguments instead of just a string.
3049 Change the key loaders to take a UI_METHOD instead of a callback
3050 function pointer. NOTE: this breaks binary compatibility with earlier
3051 versions of OpenSSL [engine].
3052 Adapt the nCipher code for these new conditions and add a card insertion
3056 *) Enhance the general user interface with mechanisms to better support
3057 dialog box interfaces, application-defined prompts, the possibility
3058 to use defaults (for example default passwords from somewhere else)
3059 and interrupts/cancellations.
3062 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
3063 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
3066 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
3067 tidy up some unnecessarily weird code in 'sk_new()').
3068 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
3070 *) Change the key loading routines for ENGINEs to use the same kind
3071 callback (pem_password_cb) as all other routines that need this
3075 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
3076 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
3077 than this minimum value is recommended.
3080 *) New random seeder for OpenVMS, using the system process statistics
3081 that are easily reachable.
3084 *) Windows apparently can't transparently handle global
3085 variables defined in DLLs. Initialisations such as:
3087 const ASN1_ITEM *it = &ASN1_INTEGER_it;
3089 wont compile. This is used by the any applications that need to
3090 declare their own ASN1 modules. This was fixed by adding the option
3091 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
3092 needed for static libraries under Win32.
3095 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
3096 setting of purpose and trust fields. New X509_STORE trust and
3097 purpose functions and tidy up setting in other SSL functions.
3100 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
3101 structure. These are inherited by X509_STORE_CTX when it is
3102 initialised. This allows various defaults to be set in the
3103 X509_STORE structure (such as flags for CRL checking and custom
3104 purpose or trust settings) for functions which only use X509_STORE_CTX
3105 internally such as S/MIME.
3107 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
3108 trust settings if they are not set in X509_STORE. This allows X509_STORE
3109 purposes and trust (in S/MIME for example) to override any set by default.
3111 Add command line options for CRL checking to smime, s_client and s_server
3115 *) Initial CRL based revocation checking. If the CRL checking flag(s)
3116 are set then the CRL is looked up in the X509_STORE structure and
3117 its validity and signature checked, then if the certificate is found
3118 in the CRL the verify fails with a revoked error.
3120 Various new CRL related callbacks added to X509_STORE_CTX structure.
3122 Command line options added to 'verify' application to support this.
3124 This needs some additional work, such as being able to handle multiple
3125 CRLs with different times, extension based lookup (rather than just
3126 by subject name) and ultimately more complete V2 CRL extension
3130 *) Add a general user interface API (crypto/ui/). This is designed
3131 to replace things like des_read_password and friends (backward
3132 compatibility functions using this new API are provided).
3133 The purpose is to remove prompting functions from the DES code
3134 section as well as provide for prompting through dialog boxes in
3135 a window system and the like.
3138 *) Add "ex_data" support to ENGINE so implementations can add state at a
3139 per-structure level rather than having to store it globally.
3142 *) Make it possible for ENGINE structures to be copied when retrieved by
3143 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
3144 This causes the "original" ENGINE structure to act like a template,
3145 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
3146 operational state can be localised to each ENGINE structure, despite the
3147 fact they all share the same "methods". New ENGINE structures returned in
3148 this case have no functional references and the return value is the single
3149 structural reference. This matches the single structural reference returned
3150 by ENGINE_by_id() normally, when it is incremented on the pre-existing
3154 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
3155 needs to match any other type at all we need to manually clear the
3159 *) Changes to the "openssl engine" utility to include;
3160 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
3161 about an ENGINE's available control commands.
3162 - executing control commands from command line arguments using the
3163 '-pre' and '-post' switches. '-post' is only used if '-t' is
3164 specified and the ENGINE is successfully initialised. The syntax for
3165 the individual commands are colon-separated, for example;
3166 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
3169 *) New dynamic control command support for ENGINEs. ENGINEs can now
3170 declare their own commands (numbers), names (strings), descriptions,
3171 and input types for run-time discovery by calling applications. A
3172 subset of these commands are implicitly classed as "executable"
3173 depending on their input type, and only these can be invoked through
3174 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
3175 can be based on user input, config files, etc). The distinction is
3176 that "executable" commands cannot return anything other than a boolean
3177 result and can only support numeric or string input, whereas some
3178 discoverable commands may only be for direct use through
3179 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
3180 pointers, or other custom uses. The "executable" commands are to
3181 support parameterisations of ENGINE behaviour that can be
3182 unambiguously defined by ENGINEs and used consistently across any
3183 OpenSSL-based application. Commands have been added to all the
3184 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
3185 control over shared-library paths without source code alterations.
3188 *) Changed all ENGINE implementations to dynamically allocate their
3189 ENGINEs rather than declaring them statically. Apart from this being
3190 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
3191 this also allows the implementations to compile without using the
3192 internal engine_int.h header.
3195 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
3196 'const' value. Any code that should be able to modify a RAND_METHOD
3197 should already have non-const pointers to it (ie. they should only
3198 modify their own ones).
3201 *) Made a variety of little tweaks to the ENGINE code.
3202 - "atalla" and "ubsec" string definitions were moved from header files
3203 to C code. "nuron" string definitions were placed in variables
3204 rather than hard-coded - allowing parameterisation of these values
3205 later on via ctrl() commands.
3206 - Removed unused "#if 0"'d code.
3207 - Fixed engine list iteration code so it uses ENGINE_free() to release
3208 structural references.
3209 - Constified the RAND_METHOD element of ENGINE structures.
3210 - Constified various get/set functions as appropriate and added
3211 missing functions (including a catch-all ENGINE_cpy that duplicates
3212 all ENGINE values onto a new ENGINE except reference counts/state).
3213 - Removed NULL parameter checks in get/set functions. Setting a method
3214 or function to NULL is a way of cancelling out a previously set
3215 value. Passing a NULL ENGINE parameter is just plain stupid anyway
3216 and doesn't justify the extra error symbols and code.
3217 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
3218 flags from engine_int.h to engine.h.
3219 - Changed prototypes for ENGINE handler functions (init(), finish(),
3220 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
3223 *) Implement binary inversion algorithm for BN_mod_inverse in addition
3224 to the algorithm using long division. The binary algorithm can be
3225 used only if the modulus is odd. On 32-bit systems, it is faster
3226 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
3227 roughly 5-15% for 256-bit moduli), so we use it only for moduli
3228 up to 450 bits. In 64-bit environments, the binary algorithm
3229 appears to be advantageous for much longer moduli; here we use it
3230 for moduli up to 2048 bits.
3233 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
3234 could not support the combine flag in choice fields.
3237 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
3238 extensions from a certificate request to the certificate.
3241 *) Allow multiple 'certopt' and 'nameopt' options to be separated
3242 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
3243 file: this allows the display of the certificate about to be
3244 signed to be customised, to allow certain fields to be included
3245 or excluded and extension details. The old system didn't display
3246 multicharacter strings properly, omitted fields not in the policy
3247 and couldn't display additional details such as extensions.
3250 *) Function EC_POINTs_mul for multiple scalar multiplication
3251 of an arbitrary number of elliptic curve points
3252 \sum scalars[i]*points[i],
3253 optionally including the generator defined for the EC_GROUP:
3254 scalar*generator + \sum scalars[i]*points[i].
3256 EC_POINT_mul is a simple wrapper function for the typical case
3257 that the point list has just one item (besides the optional
3261 *) First EC_METHODs for curves over GF(p):
3263 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
3264 operations and provides various method functions that can also
3265 operate with faster implementations of modular arithmetic.
3267 EC_GFp_mont_method() reuses most functions that are part of
3268 EC_GFp_simple_method, but uses Montgomery arithmetic.
3270 [Bodo Moeller; point addition and point doubling
3271 implementation directly derived from source code provided by
3272 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
3274 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
3275 crypto/ec/ec_lib.c):
3277 Curves are EC_GROUP objects (with an optional group generator)
3278 based on EC_METHODs that are built into the library.
3280 Points are EC_POINT objects based on EC_GROUP objects.
3282 Most of the framework would be able to handle curves over arbitrary
3283 finite fields, but as there are no obvious types for fields other
3284 than GF(p), some functions are limited to that for now.
3287 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
3288 that the file contains a complete HTTP response.
3291 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
3292 change the def and num file printf format specifier from "%-40sXXX"
3293 to "%-39s XXX". The latter will always guarantee a space after the
3294 field while the former will cause them to run together if the field
3295 is 40 of more characters long.
3298 *) Constify the cipher and digest 'method' functions and structures
3299 and modify related functions to take constant EVP_MD and EVP_CIPHER
3303 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
3304 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
3307 *) Modify EVP_Digest*() routines so they now return values. Although the
3308 internal software routines can never fail additional hardware versions
3312 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
3314 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
3315 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
3318 ERR_R_NESTED_ASN1_ERROR
3320 ERR_R_MISSING_ASN1_EOS
3321 were 4 .. 9, conflicting with
3322 ERR_LIB_RSA (= ERR_R_RSA_LIB)
3324 ERR_LIB_PEM (= ERR_R_PEM_LIB).
3325 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
3327 Add new error code 'ERR_R_INTERNAL_ERROR'.
3330 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
3334 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
3335 sets the subject name for a new request or supersedes the
3336 subject name in a given request. Formats that can be parsed are
3337 'CN=Some Name, OU=myOU, C=IT'
3339 'CN=Some Name/OU=myOU/C=IT'.
3341 Add options '-batch' and '-verbose' to 'openssl req'.
3342 [Massimiliano Pala <madwolf@hackmasters.net>]
3344 *) Introduce the possibility to access global variables through
3345 functions on platform were that's the best way to handle exporting
3346 global variables in shared libraries. To enable this functionality,
3347 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
3348 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
3349 is normally done by Configure or something similar).
3351 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
3352 in the source file (foo.c) like this:
3354 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
3355 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
3357 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
3358 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
3360 OPENSSL_DECLARE_GLOBAL(int,foo);
3361 #define foo OPENSSL_GLOBAL_REF(foo)
3362 OPENSSL_DECLARE_GLOBAL(double,bar);
3363 #define bar OPENSSL_GLOBAL_REF(bar)
3365 The #defines are very important, and therefore so is including the
3366 header file everywhere where the defined globals are used.
3368 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
3369 of ASN.1 items, but that structure is a bit different.
3371 The largest change is in util/mkdef.pl which has been enhanced with
3372 better and easier to understand logic to choose which symbols should
3373 go into the Windows .def files as well as a number of fixes and code
3374 cleanup (among others, algorithm keywords are now sorted
3375 lexicographically to avoid constant rewrites).
3378 *) In BN_div() keep a copy of the sign of 'num' before writing the
3379 result to 'rm' because if rm==num the value will be overwritten
3380 and produce the wrong result if 'num' is negative: this caused
3381 problems with BN_mod() and BN_nnmod().
3384 *) Function OCSP_request_verify(). This checks the signature on an
3385 OCSP request and verifies the signer certificate. The signer
3386 certificate is just checked for a generic purpose and OCSP request
3390 *) Add OCSP_check_validity() function to check the validity of OCSP
3391 responses. OCSP responses are prepared in real time and may only
3392 be a few seconds old. Simply checking that the current time lies
3393 between thisUpdate and nextUpdate max reject otherwise valid responses
3394 caused by either OCSP responder or client clock inaccuracy. Instead
3395 we allow thisUpdate and nextUpdate to fall within a certain period of
3396 the current time. The age of the response can also optionally be
3397 checked. Two new options -validity_period and -status_age added to
3401 *) If signature or public key algorithm is unrecognized print out its
3402 OID rather that just UNKNOWN.
3405 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
3406 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
3407 ID to be generated from the issuer certificate alone which can then be
3408 passed to OCSP_id_issuer_cmp().
3411 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
3412 ASN1 modules to export functions returning ASN1_ITEM pointers
3413 instead of the ASN1_ITEM structures themselves. This adds several
3414 new macros which allow the underlying ASN1 function/structure to
3415 be accessed transparently. As a result code should not use ASN1_ITEM
3416 references directly (such as &X509_it) but instead use the relevant
3417 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
3418 use of the new ASN1 code on platforms where exporting structures
3419 is problematical (for example in shared libraries) but exporting
3420 functions returning pointers to structures is not.
3423 *) Add support for overriding the generation of SSL/TLS session IDs.
3424 These callbacks can be registered either in an SSL_CTX or per SSL.
3425 The purpose of this is to allow applications to control, if they wish,
3426 the arbitrary values chosen for use as session IDs, particularly as it
3427 can be useful for session caching in multiple-server environments. A
3428 command-line switch for testing this (and any client code that wishes
3429 to use such a feature) has been added to "s_server".
3430 [Geoff Thorpe, Lutz Jaenicke]
3432 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
3433 of the form '#if defined(...) || defined(...) || ...' and
3434 '#if !defined(...) && !defined(...) && ...'. This also avoids
3435 the growing number of special cases it was previously handling.
3438 *) Make all configuration macros available for application by making
3439 sure they are available in opensslconf.h, by giving them names starting
3440 with "OPENSSL_" to avoid conflicts with other packages and by making
3441 sure e_os2.h will cover all platform-specific cases together with
3443 Additionally, it is now possible to define configuration/platform-
3444 specific names (called "system identities"). In the C code, these
3445 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
3446 macro with the name beginning with "OPENSSL_SYS_", which is determined
3447 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
3451 *) New option -set_serial to 'req' and 'x509' this allows the serial
3452 number to use to be specified on the command line. Previously self
3453 signed certificates were hard coded with serial number 0 and the
3454 CA options of 'x509' had to use a serial number in a file which was
3458 *) New options to 'ca' utility to support V2 CRL entry extensions.
3459 Currently CRL reason, invalidity date and hold instruction are
3460 supported. Add new CRL extensions to V3 code and some new objects.
3463 *) New function EVP_CIPHER_CTX_set_padding() this is used to
3464 disable standard block padding (aka PKCS#5 padding) in the EVP
3465 API, which was previously mandatory. This means that the data is
3466 not padded in any way and so the total length much be a multiple
3467 of the block size, otherwise an error occurs.
3470 *) Initial (incomplete) OCSP SSL support.
3473 *) New function OCSP_parse_url(). This splits up a URL into its host,
3474 port and path components: primarily to parse OCSP URLs. New -url
3475 option to ocsp utility.
3478 *) New nonce behavior. The return value of OCSP_check_nonce() now
3479 reflects the various checks performed. Applications can decide
3480 whether to tolerate certain situations such as an absent nonce
3481 in a response when one was present in a request: the ocsp application
3482 just prints out a warning. New function OCSP_add1_basic_nonce()
3483 this is to allow responders to include a nonce in a response even if
3484 the request is nonce-less.
3487 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
3488 skipped when using openssl x509 multiple times on a single input file,
3489 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
3492 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
3493 set string type: to handle setting ASN1_TIME structures. Fix ca
3494 utility to correctly initialize revocation date of CRLs.
3497 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
3498 the clients preferred ciphersuites and rather use its own preferences.
3499 Should help to work around M$ SGC (Server Gated Cryptography) bug in
3500 Internet Explorer by ensuring unchanged hash method during stepup.
3501 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
3504 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
3505 to aes and add a new 'exist' option to print out symbols that don't
3509 *) Additional options to ocsp utility to allow flags to be set and
3510 additional certificates supplied.
3513 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
3514 OCSP client a number of certificate to only verify the response
3518 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
3519 handle the new API. Currently only ECB, CBC modes supported. Add new
3522 Add TLS AES ciphersuites as described in RFC3268, "Advanced
3523 Encryption Standard (AES) Ciphersuites for Transport Layer
3524 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
3525 not enabled by default and were not part of the "ALL" ciphersuite
3526 alias because they were not yet official; they could be
3527 explicitly requested by specifying the "AESdraft" ciphersuite
3528 group alias. In the final release of OpenSSL 0.9.7, the group
3529 alias is called "AES" and is part of "ALL".)
3530 [Ben Laurie, Steve Henson, Bodo Moeller]
3532 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
3533 request to response.
3536 *) Functions for OCSP responders. OCSP_request_onereq_count(),
3537 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
3538 extract information from a certificate request. OCSP_response_create()
3539 creates a response and optionally adds a basic response structure.
3540 OCSP_basic_add1_status() adds a complete single response to a basic
3541 response and returns the OCSP_SINGLERESP structure just added (to allow
3542 extensions to be included for example). OCSP_basic_add1_cert() adds a
3543 certificate to a basic response and OCSP_basic_sign() signs a basic
3544 response with various flags. New helper functions ASN1_TIME_check()
3545 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
3546 (converts ASN1_TIME to GeneralizedTime).
3549 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
3550 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
3551 structure from a certificate. X509_pubkey_digest() digests the public_key
3552 contents: this is used in various key identifiers.
3555 *) Make sk_sort() tolerate a NULL argument.
3556 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
3558 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
3559 passed by the function are trusted implicitly. If any of them signed the
3560 response then it is assumed to be valid and is not verified.
3563 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
3564 to data. This was previously part of the PKCS7 ASN1 code. This
3565 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
3566 [Steve Henson, reported by Kenneth R. Robinette
3567 <support@securenetterm.com>]
3569 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
3570 routines: without these tracing memory leaks is very painful.
3571 Fix leaks in PKCS12 and PKCS7 routines.
3574 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
3575 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
3576 effectively meant GeneralizedTime would never be used. Now it
3577 is initialised to -1 but X509_time_adj() now has to check the value
3578 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
3579 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
3580 [Steve Henson, reported by Kenneth R. Robinette
3581 <support@securenetterm.com>]
3583 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
3584 result in a zero length in the ASN1_INTEGER structure which was
3585 not consistent with the structure when d2i_ASN1_INTEGER() was used
3586 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
3587 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
3588 where it did not print out a minus for negative ASN1_INTEGER.
3591 *) Add summary printout to ocsp utility. The various functions which
3592 convert status values to strings have been renamed to:
3593 OCSP_response_status_str(), OCSP_cert_status_str() and
3594 OCSP_crl_reason_str() and are no longer static. New options
3595 to verify nonce values and to disable verification. OCSP response
3596 printout format cleaned up.
3599 *) Add additional OCSP certificate checks. These are those specified
3600 in RFC2560. This consists of two separate checks: the CA of the
3601 certificate being checked must either be the OCSP signer certificate
3602 or the issuer of the OCSP signer certificate. In the latter case the
3603 OCSP signer certificate must contain the OCSP signing extended key
3604 usage. This check is performed by attempting to match the OCSP
3605 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
3606 in the OCSP_CERTID structures of the response.
3609 *) Initial OCSP certificate verification added to OCSP_basic_verify()
3610 and related routines. This uses the standard OpenSSL certificate
3611 verify routines to perform initial checks (just CA validity) and
3612 to obtain the certificate chain. Then additional checks will be
3613 performed on the chain. Currently the root CA is checked to see
3614 if it is explicitly trusted for OCSP signing. This is used to set
3615 a root CA as a global signing root: that is any certificate that
3616 chains to that CA is an acceptable OCSP signing certificate.
3619 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
3620 extensions from a separate configuration file.
3621 As when reading extensions from the main configuration file,
3622 the '-extensions ...' option may be used for specifying the
3624 [Massimiliano Pala <madwolf@comune.modena.it>]
3626 *) New OCSP utility. Allows OCSP requests to be generated or
3627 read. The request can be sent to a responder and the output
3628 parsed, outputed or printed in text form. Not complete yet:
3629 still needs to check the OCSP response validity.
3632 *) New subcommands for 'openssl ca':
3633 'openssl ca -status <serial>' prints the status of the cert with
3634 the given serial number (according to the index file).
3635 'openssl ca -updatedb' updates the expiry status of certificates
3637 [Massimiliano Pala <madwolf@comune.modena.it>]
3639 *) New '-newreq-nodes' command option to CA.pl. This is like
3640 '-newreq', but calls 'openssl req' with the '-nodes' option
3641 so that the resulting key is not encrypted.
3642 [Damien Miller <djm@mindrot.org>]
3644 *) New configuration for the GNU Hurd.
3645 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
3647 *) Initial code to implement OCSP basic response verify. This
3648 is currently incomplete. Currently just finds the signer's
3649 certificate and verifies the signature on the response.
3652 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
3653 value of OPENSSLDIR. This is available via the new '-d' option
3654 to 'openssl version', and is also included in 'openssl version -a'.
3657 *) Allowing defining memory allocation callbacks that will be given
3658 file name and line number information in additional arguments
3659 (a const char* and an int). The basic functionality remains, as
3660 well as the original possibility to just replace malloc(),
3661 realloc() and free() by functions that do not know about these
3662 additional arguments. To register and find out the current
3663 settings for extended allocation functions, the following
3664 functions are provided:
3666 CRYPTO_set_mem_ex_functions
3667 CRYPTO_set_locked_mem_ex_functions
3668 CRYPTO_get_mem_ex_functions
3669 CRYPTO_get_locked_mem_ex_functions
3671 These work the same way as CRYPTO_set_mem_functions and friends.
3672 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
3673 extended allocation function is enabled.
3674 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
3675 a conventional allocation function is enabled.
3676 [Richard Levitte, Bodo Moeller]
3678 *) Finish off removing the remaining LHASH function pointer casts.
3679 There should no longer be any prototype-casting required when using
3680 the LHASH abstraction, and any casts that remain are "bugs". See
3681 the callback types and macros at the head of lhash.h for details
3682 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
3685 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
3686 If /dev/[u]random devices are not available or do not return enough
3687 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
3689 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
3690 /etc/entropy will be queried once each in this sequence, quering stops
3691 when enough entropy was collected without querying more sockets.
3694 *) Change the Unix RAND_poll() variant to be able to poll several
3695 random devices, as specified by DEVRANDOM, until a sufficient amount
3696 of data has been collected. We spend at most 10 ms on each file
3697 (select timeout) and read in non-blocking mode. DEVRANDOM now
3698 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
3699 (previously it was just the string "/dev/urandom"), so on typical
3700 platforms the 10 ms delay will never occur.
3701 Also separate out the Unix variant to its own file, rand_unix.c.
3702 For VMS, there's a currently-empty rand_vms.c.
3705 *) Move OCSP client related routines to ocsp_cl.c. These
3706 provide utility functions which an application needing
3707 to issue a request to an OCSP responder and analyse the
3708 response will typically need: as opposed to those which an
3709 OCSP responder itself would need which will be added later.
3711 OCSP_request_sign() signs an OCSP request with an API similar
3712 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
3713 response. OCSP_response_get1_basic() extracts basic response
3714 from response. OCSP_resp_find_status(): finds and extracts status
3715 information from an OCSP_CERTID structure (which will be created
3716 when the request structure is built). These are built from lower
3717 level functions which work on OCSP_SINGLERESP structures but
3718 wont normally be used unless the application wishes to examine
3719 extensions in the OCSP response for example.
3721 Replace nonce routines with a pair of functions.
3722 OCSP_request_add1_nonce() adds a nonce value and optionally
3723 generates a random value. OCSP_check_nonce() checks the
3724 validity of the nonce in an OCSP response.
3727 *) Change function OCSP_request_add() to OCSP_request_add0_id().
3728 This doesn't copy the supplied OCSP_CERTID and avoids the
3729 need to free up the newly created id. Change return type
3730 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
3731 This can then be used to add extensions to the request.
3732 Deleted OCSP_request_new(), since most of its functionality
3733 is now in OCSP_REQUEST_new() (and the case insensitive name
3734 clash) apart from the ability to set the request name which
3735 will be added elsewhere.
3738 *) Update OCSP API. Remove obsolete extensions argument from
3739 various functions. Extensions are now handled using the new
3740 OCSP extension code. New simple OCSP HTTP function which
3741 can be used to send requests and parse the response.
3744 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
3745 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
3746 uses the special reorder version of SET OF to sort the attributes
3747 and reorder them to match the encoded order. This resolves a long
3748 standing problem: a verify on a PKCS7 structure just after signing
3749 it used to fail because the attribute order did not match the
3750 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
3751 it uses the received order. This is necessary to tolerate some broken
3752 software that does not order SET OF. This is handled by encoding
3753 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
3754 to produce the required SET OF.
3757 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
3758 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
3759 files to get correct declarations of the ASN.1 item variables.
3762 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
3763 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
3764 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
3765 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
3766 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
3767 ASN1_ITEM and no wrapper functions.
3770 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
3771 replace the old function pointer based I/O routines. Change most of
3772 the *_d2i_bio() and *_d2i_fp() functions to use these.
3775 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
3776 lines, recognice more "algorithms" that can be deselected, and make
3777 it complain about algorithm deselection that isn't recognised.
3780 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
3781 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
3782 to use new functions. Add NO_ASN1_OLD which can be set to remove
3783 some old style ASN1 functions: this can be used to determine if old
3784 code will still work when these eventually go away.
3787 *) New extension functions for OCSP structures, these follow the
3788 same conventions as certificates and CRLs.
3791 *) New function X509V3_add1_i2d(). This automatically encodes and
3792 adds an extension. Its behaviour can be customised with various
3793 flags to append, replace or delete. Various wrappers added for
3794 certifcates and CRLs.
3797 *) Fix to avoid calling the underlying ASN1 print routine when
3798 an extension cannot be parsed. Correct a typo in the
3799 OCSP_SERVICELOC extension. Tidy up print OCSP format.
3802 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
3803 entries for variables.
3806 *) Add functionality to apps/openssl.c for detecting locking
3807 problems: As the program is single-threaded, all we have
3808 to do is register a locking callback using an array for
3809 storing which locks are currently held by the program.
3812 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
3813 SSL_get_ex_data_X509_STORE_idx(), which is used in
3814 ssl_verify_cert_chain() and thus can be called at any time
3815 during TLS/SSL handshakes so that thread-safety is essential.
3816 Unfortunately, the ex_data design is not at all suited
3817 for multi-threaded use, so it probably should be abolished.
3820 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
3821 [Broadcom, tweaked and integrated by Geoff Thorpe]
3823 *) Move common extension printing code to new function
3824 X509V3_print_extensions(). Reorganise OCSP print routines and
3825 implement some needed OCSP ASN1 functions. Add OCSP extensions.
3828 *) New function X509_signature_print() to remove duplication in some
3832 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
3833 set (this was treated exactly the same as SET OF previously). This
3834 is used to reorder the STACK representing the structure to match the
3835 encoding. This will be used to get round a problem where a PKCS7
3836 structure which was signed could not be verified because the STACK
3837 order did not reflect the encoded order.
3840 *) Reimplement the OCSP ASN1 module using the new code.
3843 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
3844 for its ASN1 operations. The old style function pointers still exist
3845 for now but they will eventually go away.
3848 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
3849 completely replaces the old ASN1 functionality with a table driven
3850 encoder and decoder which interprets an ASN1_ITEM structure describing
3851 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
3852 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
3853 has also been converted to the new form.
3856 *) Change BN_mod_exp_recp so that negative moduli are tolerated
3857 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
3858 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
3859 for negative moduli.
3862 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
3863 of not touching the result's sign bit.
3866 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
3870 *) Changed the LHASH code to use prototypes for callbacks, and created
3871 macros to declare and implement thin (optionally static) functions
3872 that provide type-safety and avoid function pointer casting for the
3873 type-specific callbacks.
3876 *) Added Kerberos Cipher Suites to be used with TLS, as written in
3878 [Veers Staats <staatsvr@asc.hpc.mil>,
3879 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
3881 *) Reformat the FAQ so the different questions and answers can be divided
3882 in sections depending on the subject.
3885 *) Have the zlib compression code load ZLIB.DLL dynamically under
3889 *) New function BN_mod_sqrt for computing square roots modulo a prime
3890 (using the probabilistic Tonelli-Shanks algorithm unless
3891 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
3892 be handled deterministically).
3893 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3895 *) Make BN_mod_inverse faster by explicitly handling small quotients
3896 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
3897 512 bits], about 30% for larger ones [1024 or 2048 bits].)
3900 *) New function BN_kronecker.
3903 *) Fix BN_gcd so that it works on negative inputs; the result is
3904 positive unless both parameters are zero.
3905 Previously something reasonably close to an infinite loop was
3906 possible because numbers could be growing instead of shrinking
3907 in the implementation of Euclid's algorithm.
3910 *) Fix BN_is_word() and BN_is_one() macros to take into account the
3911 sign of the number in question.
3913 Fix BN_is_word(a,w) to work correctly for w == 0.
3915 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
3916 because its test if the absolute value of 'a' equals 'w'.
3917 Note that BN_abs_is_word does *not* handle w == 0 reliably;
3918 it exists mostly for use in the implementations of BN_is_zero(),
3919 BN_is_one(), and BN_is_word().
3922 *) New function BN_swap.
3925 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
3926 the exponentiation functions are more likely to produce reasonable
3927 results on negative inputs.
3930 *) Change BN_mod_mul so that the result is always non-negative.
3931 Previously, it could be negative if one of the factors was negative;
3932 I don't think anyone really wanted that behaviour.
3935 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
3936 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
3937 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
3938 and add new functions:
3947 BN_mod_lshift1_quick
3951 These functions always generate non-negative results.
3953 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
3954 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
3956 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
3957 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
3958 be reduced modulo m.
3959 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3962 The following entry accidentily appeared in the CHANGES file
3963 distributed with OpenSSL 0.9.7. The modifications described in
3964 it do *not* apply to OpenSSL 0.9.7.
3966 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3967 was actually never needed) and in BN_mul(). The removal in BN_mul()
3968 required a small change in bn_mul_part_recursive() and the addition
3969 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3970 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3971 bn_sub_words() and bn_add_words() except they take arrays with
3976 *) In 'openssl passwd', verify passwords read from the terminal
3977 unless the '-salt' option is used (which usually means that
3978 verification would just waste user's time since the resulting
3979 hash is going to be compared with some given password hash)
3980 or the new '-noverify' option is used.
3982 This is an incompatible change, but it does not affect
3983 non-interactive use of 'openssl passwd' (passwords on the command
3984 line, '-stdin' option, '-in ...' option) and thus should not
3988 *) Remove all references to RSAref, since there's no more need for it.
3991 *) Make DSO load along a path given through an environment variable
3992 (SHLIB_PATH) with shl_load().
3995 *) Constify the ENGINE code as a result of BIGNUM constification.
3996 Also constify the RSA code and most things related to it. In a
3997 few places, most notable in the depth of the ASN.1 code, ugly
3998 casts back to non-const were required (to be solved at a later
4002 *) Make it so the openssl application has all engines loaded by default.
4005 *) Constify the BIGNUM routines a little more.
4008 *) Add the following functions:
4010 ENGINE_load_cswift()
4012 ENGINE_load_atalla()
4014 ENGINE_load_builtin_engines()
4016 That way, an application can itself choose if external engines that
4017 are built-in in OpenSSL shall ever be used or not. The benefit is
4018 that applications won't have to be linked with libdl or other dso
4019 libraries unless it's really needed.
4021 Changed 'openssl engine' to load all engines on demand.
4022 Changed the engine header files to avoid the duplication of some
4023 declarations (they differed!).
4026 *) 'openssl engine' can now list capabilities.
4029 *) Better error reporting in 'openssl engine'.
4032 *) Never call load_dh_param(NULL) in s_server.
4035 *) Add engine application. It can currently list engines by name and
4036 identity, and test if they are actually available.
4039 *) Improve RPM specification file by forcing symbolic linking and making
4040 sure the installed documentation is also owned by root.root.
4041 [Damien Miller <djm@mindrot.org>]
4043 *) Give the OpenSSL applications more possibilities to make use of
4044 keys (public as well as private) handled by engines.
4047 *) Add OCSP code that comes from CertCo.
4050 *) Add VMS support for the Rijndael code.
4053 *) Added untested support for Nuron crypto accelerator.
4056 *) Add support for external cryptographic devices. This code was
4057 previously distributed separately as the "engine" branch.
4058 [Geoff Thorpe, Richard Levitte]
4060 *) Rework the filename-translation in the DSO code. It is now possible to
4061 have far greater control over how a "name" is turned into a filename
4062 depending on the operating environment and any oddities about the
4063 different shared library filenames on each system.
4066 *) Support threads on FreeBSD-elf in Configure.
4069 *) Fix for SHA1 assembly problem with MASM: it produces
4070 warnings about corrupt line number information when assembling
4071 with debugging information. This is caused by the overlapping
4073 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
4076 NCONF_get_number() has no error checking at all. As a replacement,
4077 NCONF_get_number_e() is defined (_e for "error checking") and is
4078 promoted strongly. The old NCONF_get_number is kept around for
4079 binary backward compatibility.
4080 Make it possible for methods to load from something other than a BIO,
4081 by providing a function pointer that is given a name instead of a BIO.
4082 For example, this could be used to load configuration data from an
4086 *) Fix for non blocking accept BIOs. Added new I/O special reason
4087 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
4088 with non blocking I/O was not possible because no retry code was
4089 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
4093 *) Added the beginnings of Rijndael support.
4096 *) Fix for bug in DirectoryString mask setting. Add support for
4097 X509_NAME_print_ex() in 'req' and X509_print_ex() function
4098 to allow certificate printing to more controllable, additional
4099 'certopt' option to 'x509' to allow new printing options to be
4103 *) Clean old EAY MD5 hack from e_os.h.
4106 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
4108 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4109 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4110 [Joe Orton, Steve Henson]
4112 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
4114 *) Fix additional bug revealed by the NISCC test suite:
4116 Stop bug triggering large recursion when presented with
4117 certain ASN.1 tags (CVE-2003-0851)
4120 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
4122 *) Fix various bugs revealed by running the NISCC test suite:
4124 Stop out of bounds reads in the ASN1 code when presented with
4125 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4127 If verify callback ignores invalid public key errors don't try to check
4128 certificate signature with the NULL public key.
4132 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4133 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4137 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4138 extra data after the compression methods not only for TLS 1.0
4139 but also for SSL 3.0 (as required by the specification).
4140 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4142 *) Change X509_certificate_type() to mark the key as exported/exportable
4143 when it's 512 *bits* long, not 512 bytes.
4146 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
4148 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4149 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4150 a protocol version number mismatch like a decryption error
4151 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4154 *) Turn on RSA blinding by default in the default implementation
4155 to avoid a timing attack. Applications that don't want it can call
4156 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4157 They would be ill-advised to do so in most cases.
4158 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4160 *) Change RSA blinding code so that it works when the PRNG is not
4161 seeded (in this case, the secret RSA exponent is abused as
4162 an unpredictable seed -- if it is not unpredictable, there
4163 is no point in blinding anyway). Make RSA blinding thread-safe
4164 by remembering the creator's thread ID in rsa->blinding and
4165 having all other threads use local one-time blinding factors
4166 (this requires more computation than sharing rsa->blinding, but
4167 avoids excessive locking; and if an RSA object is not shared
4168 between threads, blinding will still be very fast).
4171 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
4173 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4174 via timing by performing a MAC computation even if incorrrect
4175 block cipher padding has been found. This is a countermeasure
4176 against active attacks where the attacker has to distinguish
4177 between bad padding and a MAC verification error. (CVE-2003-0078)
4179 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4180 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4181 Martin Vuagnoux (EPFL, Ilion)]
4183 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
4185 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
4186 memory from it's contents. This is done with a counter that will
4187 place alternating values in each byte. This can be used to solve
4188 two issues: 1) the removal of calls to memset() by highly optimizing
4189 compilers, and 2) cleansing with other values than 0, since those can
4190 be read through on certain media, for example a swap space on disk.
4193 *) Bugfix: client side session caching did not work with external caching,
4194 because the session->cipher setting was not restored when reloading
4195 from the external cache. This problem was masked, when
4196 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
4197 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
4200 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
4201 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
4202 [Zeev Lieber <zeev-l@yahoo.com>]
4204 *) Undo an undocumented change introduced in 0.9.6e which caused
4205 repeated calls to OpenSSL_add_all_ciphers() and
4206 OpenSSL_add_all_digests() to be ignored, even after calling
4210 *) Change the default configuration reader to deal with last line not
4211 being properly terminated.
4214 *) Change X509_NAME_cmp() so it applies the special rules on handling
4215 DN values that are of type PrintableString, as well as RDNs of type
4216 emailAddress where the value has the type ia5String.
4217 [stefank@valicert.com via Richard Levitte]
4219 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
4220 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
4221 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
4222 the bitwise-OR of the two for use by the majority of applications
4223 wanting this behaviour, and update the docs. The documented
4224 behaviour and actual behaviour were inconsistent and had been
4225 changing anyway, so this is more a bug-fix than a behavioural
4227 [Geoff Thorpe, diagnosed by Nadav Har'El]
4229 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
4230 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
4233 *) Fix initialization code race conditions in
4234 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
4235 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
4236 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
4237 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
4238 ssl2_get_cipher_by_char(),
4239 ssl3_get_cipher_by_char().
4240 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
4242 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
4243 the cached sessions are flushed, as the remove_cb() might use ex_data
4244 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
4245 (see [openssl.org #212]).
4246 [Geoff Thorpe, Lutz Jaenicke]
4248 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
4249 length, instead of the encoding length to d2i_ASN1_OBJECT.
4252 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
4254 *) [In 0.9.6g-engine release:]
4255 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
4256 [Lynn Gazis <lgazis@rainbow.com>]
4258 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
4260 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
4261 and get fix the header length calculation.
4262 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
4263 Alon Kantor <alonk@checkpoint.com> (and others),
4266 *) Use proper error handling instead of 'assertions' in buffer
4267 overflow checks added in 0.9.6e. This prevents DoS (the
4268 assertions could call abort()).
4269 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
4271 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
4273 *) Add various sanity checks to asn1_get_length() to reject
4274 the ASN1 length bytes if they exceed sizeof(long), will appear
4275 negative or the content length exceeds the length of the
4277 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4279 *) Fix cipher selection routines: ciphers without encryption had no flags
4280 for the cipher strength set and where therefore not handled correctly
4281 by the selection routines (PR #130).
4284 *) Fix EVP_dsa_sha macro.
4288 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
4289 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
4290 that was added in OpenSSL 0.9.6d.
4292 As the countermeasure turned out to be incompatible with some
4293 broken SSL implementations, the new option is part of SSL_OP_ALL.
4294 SSL_OP_ALL is usually employed when compatibility with weird SSL
4295 implementations is desired (e.g. '-bugs' option to 's_client' and
4296 's_server'), so the new option is automatically set in many
4300 *) Changes in security patch:
4302 Changes marked "(CHATS)" were sponsored by the Defense Advanced
4303 Research Projects Agency (DARPA) and Air Force Research Laboratory,
4304 Air Force Materiel Command, USAF, under agreement number
4307 *) Add various sanity checks to asn1_get_length() to reject
4308 the ASN1 length bytes if they exceed sizeof(long), will appear
4309 negative or the content length exceeds the length of the
4310 supplied buffer. (CVE-2002-0659)
4311 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4313 *) Assertions for various potential buffer overflows, not known to
4315 [Ben Laurie (CHATS)]
4317 *) Various temporary buffers to hold ASCII versions of integers were
4318 too small for 64 bit platforms. (CVE-2002-0655)
4319 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
4321 *) Remote buffer overflow in SSL3 protocol - an attacker could
4322 supply an oversized session ID to a client. (CVE-2002-0656)
4323 [Ben Laurie (CHATS)]
4325 *) Remote buffer overflow in SSL2 protocol - an attacker could
4326 supply an oversized client master key. (CVE-2002-0656)
4327 [Ben Laurie (CHATS)]
4329 Changes between 0.9.6c and 0.9.6d [9 May 2002]
4331 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
4332 encoded as NULL) with id-dsa-with-sha1.
4333 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
4335 *) Check various X509_...() return values in apps/req.c.
4336 [Nils Larsch <nla@trustcenter.de>]
4338 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
4339 an end-of-file condition would erronously be flagged, when the CRLF
4340 was just at the end of a processed block. The bug was discovered when
4341 processing data through a buffering memory BIO handing the data to a
4342 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
4343 <ptsekov@syntrex.com> and Nedelcho Stanev.
4346 *) Implement a countermeasure against a vulnerability recently found
4347 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
4348 before application data chunks to avoid the use of known IVs
4349 with data potentially chosen by the attacker.
4352 *) Fix length checks in ssl3_get_client_hello().
4355 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
4356 to prevent ssl3_read_internal() from incorrectly assuming that
4357 ssl3_read_bytes() found application data while handshake
4358 processing was enabled when in fact s->s3->in_read_app_data was
4359 merely automatically cleared during the initial handshake.
4360 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
4362 *) Fix object definitions for Private and Enterprise: they were not
4363 recognized in their shortname (=lowercase) representation. Extend
4364 obj_dat.pl to issue an error when using undefined keywords instead
4365 of silently ignoring the problem (Svenning Sorensen
4366 <sss@sss.dnsalias.net>).
4369 *) Fix DH_generate_parameters() so that it works for 'non-standard'
4370 generators, i.e. generators other than 2 and 5. (Previously, the
4371 code did not properly initialise the 'add' and 'rem' values to
4372 BN_generate_prime().)
4374 In the new general case, we do not insist that 'generator' is
4375 actually a primitive root: This requirement is rather pointless;
4376 a generator of the order-q subgroup is just as good, if not
4380 *) Map new X509 verification errors to alerts. Discovered and submitted by
4381 Tom Wu <tom@arcot.com>.
4384 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
4385 returning non-zero before the data has been completely received
4386 when using non-blocking I/O.
4387 [Bodo Moeller; problem pointed out by John Hughes]
4389 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
4390 [Ben Laurie, Lutz Jaenicke]
4392 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
4393 Yoram Zahavi <YoramZ@gilian.com>).
4396 *) Add information about CygWin 1.3 and on, and preserve proper
4397 configuration for the versions before that.
4398 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4400 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
4401 check whether we deal with a copy of a session and do not delete from
4402 the cache in this case. Problem reported by "Izhar Shoshani Levi"
4403 <izhar@checkpoint.com>.
4406 *) Do not store session data into the internal session cache, if it
4407 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
4408 flag is set). Proposed by Aslam <aslam@funk.com>.
4411 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
4415 *) [In 0.9.6d-engine release:]
4416 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4417 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4419 *) Add the configuration target linux-s390x.
4420 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
4422 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
4423 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
4424 variable as an indication that a ClientHello message has been
4425 received. As the flag value will be lost between multiple
4426 invocations of ssl3_accept when using non-blocking I/O, the
4427 function may not be aware that a handshake has actually taken
4428 place, thus preventing a new session from being added to the
4431 To avoid this problem, we now set s->new_session to 2 instead of
4432 using a local variable.
4433 [Lutz Jaenicke, Bodo Moeller]
4435 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
4436 if the SSL_R_LENGTH_MISMATCH error is detected.
4437 [Geoff Thorpe, Bodo Moeller]
4439 *) New 'shared_ldflag' column in Configure platform table.
4442 *) Fix EVP_CIPHER_mode macro.
4443 ["Dan S. Camper" <dan@bti.net>]
4445 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
4446 type, we must throw them away by setting rr->length to 0.
4447 [D P Chang <dpc@qualys.com>]
4449 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
4451 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
4452 <Dominikus.Scherkl@biodata.com>. (The previous implementation
4453 worked incorrectly for those cases where range = 10..._2 and
4454 3*range is two bits longer than range.)
4457 *) Only add signing time to PKCS7 structures if it is not already
4461 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
4462 OBJ_ld_ce should be OBJ_id_ce.
4463 Also some ip-pda OIDs in crypto/objects/objects.txt were
4464 incorrect (cf. RFC 3039).
4465 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
4467 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
4468 returns early because it has nothing to do.
4469 [Andy Schneider <andy.schneider@bjss.co.uk>]
4471 *) [In 0.9.6c-engine release:]
4472 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
4473 [Andy Schneider <andy.schneider@bjss.co.uk>]
4475 *) [In 0.9.6c-engine release:]
4476 Add support for Cryptographic Appliance's keyserver technology.
4477 (Use engine 'keyclient')
4478 [Cryptographic Appliances and Geoff Thorpe]
4480 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
4481 is called via tools/c89.sh because arguments have to be
4482 rearranged (all '-L' options must appear before the first object
4484 [Richard Shapiro <rshapiro@abinitio.com>]
4486 *) [In 0.9.6c-engine release:]
4487 Add support for Broadcom crypto accelerator cards, backported
4489 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
4491 *) [In 0.9.6c-engine release:]
4492 Add support for SureWare crypto accelerator cards from
4493 Baltimore Technologies. (Use engine 'sureware')
4494 [Baltimore Technologies and Mark Cox]
4496 *) [In 0.9.6c-engine release:]
4497 Add support for crypto accelerator cards from Accelerated
4498 Encryption Processing, www.aep.ie. (Use engine 'aep')
4499 [AEP Inc. and Mark Cox]
4501 *) Add a configuration entry for gcc on UnixWare.
4502 [Gary Benson <gbenson@redhat.com>]
4504 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
4505 messages are stored in a single piece (fixed-length part and
4506 variable-length part combined) and fix various bugs found on the way.
4509 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
4510 instead. BIO_gethostbyname() does not know what timeouts are
4511 appropriate, so entries would stay in cache even when they have
4513 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
4515 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
4516 faced with a pathologically small ClientHello fragment that does
4517 not contain client_version: Instead of aborting with an error,
4518 simply choose the highest available protocol version (i.e.,
4519 TLS 1.0 unless it is disabled). In practice, ClientHello
4520 messages are never sent like this, but this change gives us
4521 strictly correct behaviour at least for TLS.
4524 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
4525 never resets s->method to s->ctx->method when called from within
4526 one of the SSL handshake functions.
4527 [Bodo Moeller; problem pointed out by Niko Baric]
4529 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
4530 (sent using the client's version number) if client_version is
4531 smaller than the protocol version in use. Also change
4532 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
4533 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
4534 the client will at least see that alert.
4537 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
4541 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
4542 client receives HelloRequest while in a handshake.
4543 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
4545 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
4546 should end in 'break', not 'goto end' which circuments various
4547 cleanups done in state SSL_ST_OK. But session related stuff
4548 must be disabled for SSL_ST_OK in the case that we just sent a
4551 Also avoid some overhead by not calling ssl_init_wbio_buffer()
4552 before just sending a HelloRequest.
4553 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
4555 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
4556 reveal whether illegal block cipher padding was found or a MAC
4557 verification error occured. (Neither SSLerr() codes nor alerts
4558 are directly visible to potential attackers, but the information
4559 may leak via logfiles.)
4561 Similar changes are not required for the SSL 2.0 implementation
4562 because the number of padding bytes is sent in clear for SSL 2.0,
4563 and the extra bytes are just ignored. However ssl/s2_pkt.c
4564 failed to verify that the purported number of padding bytes is in
4568 *) Add OpenUNIX-8 support including shared libraries
4569 (Boyd Lynn Gerber <gerberb@zenez.com>).
4572 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
4573 'wristwatch attack' using huge encoding parameters (cf.
4574 James H. Manger's CRYPTO 2001 paper). Note that the
4575 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
4576 encoding parameters and hence was not vulnerable.
4579 *) BN_sqr() bug fix.
4580 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
4582 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
4583 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
4584 followed by modular reduction.
4585 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
4587 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
4588 equivalent based on BN_pseudo_rand() instead of BN_rand().
4591 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
4592 This function was broken, as the check for a new client hello message
4593 to handle SGC did not allow these large messages.
4594 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
4597 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
4600 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
4601 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
4604 *) Rework the configuration and shared library support for Tru64 Unix.
4605 The configuration part makes use of modern compiler features and
4606 still retains old compiler behavior for those that run older versions
4607 of the OS. The shared library support part includes a variant that
4608 uses the RPATH feature, and is available through the special
4609 configuration target "alpha-cc-rpath", which will never be selected
4611 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
4613 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
4614 with the same message size as in ssl3_get_certificate_request().
4615 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
4616 messages might inadvertently be reject as too long.
4617 [Petr Lampa <lampa@fee.vutbr.cz>]
4619 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
4622 *) Modified SSL library such that the verify_callback that has been set
4623 specificly for an SSL object with SSL_set_verify() is actually being
4624 used. Before the change, a verify_callback set with this function was
4625 ignored and the verify_callback() set in the SSL_CTX at the time of
4626 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
4627 to allow the necessary settings.
4630 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
4631 explicitly to NULL, as at least on Solaris 8 this seems not always to be
4632 done automatically (in contradiction to the requirements of the C
4633 standard). This made problems when used from OpenSSH.
4636 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
4637 dh->length and always used
4639 BN_rand_range(priv_key, dh->p).
4641 BN_rand_range() is not necessary for Diffie-Hellman, and this
4642 specific range makes Diffie-Hellman unnecessarily inefficient if
4643 dh->length (recommended exponent length) is much smaller than the
4644 length of dh->p. We could use BN_rand_range() if the order of
4645 the subgroup was stored in the DH structure, but we only have
4650 BN_rand(priv_key, l, ...)
4652 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
4658 RSA_eay_public_encrypt
4659 RSA_eay_private_decrypt
4660 RSA_eay_private_encrypt (signing)
4661 RSA_eay_public_decrypt (signature verification)
4663 (default implementations for RSA_public_encrypt,
4664 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
4665 always reject numbers >= n.
4668 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
4669 to synchronize access to 'locking_thread'. This is necessary on
4670 systems where access to 'locking_thread' (an 'unsigned long'
4671 variable) is not atomic.
4674 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
4675 *before* setting the 'crypto_lock_rand' flag. The previous code had
4676 a race condition if 0 is a valid thread ID.
4677 [Travis Vitek <vitek@roguewave.com>]
4679 *) Add support for shared libraries under Irix.
4680 [Albert Chin-A-Young <china@thewrittenword.com>]
4682 *) Add configuration option to build on Linux on both big-endian and
4684 [Ralf Baechle <ralf@uni-koblenz.de>]
4686 *) Add the possibility to create shared libraries on HP-UX.
4689 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
4691 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
4692 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
4693 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
4694 PRNG state recovery was possible based on the output of
4695 one PRNG request appropriately sized to gain knowledge on
4696 'md' followed by enough consecutive 1-byte PRNG requests
4697 to traverse all of 'state'.
4699 1. When updating 'md_local' (the current thread's copy of 'md')
4700 during PRNG output generation, hash all of the previous
4701 'md_local' value, not just the half used for PRNG output.
4703 2. Make the number of bytes from 'state' included into the hash
4704 independent from the number of PRNG bytes requested.
4706 The first measure alone would be sufficient to avoid
4707 Markku-Juhani's attack. (Actually it had never occurred
4708 to me that the half of 'md_local' used for chaining was the
4709 half from which PRNG output bytes were taken -- I had always
4710 assumed that the secret half would be used.) The second
4711 measure makes sure that additional data from 'state' is never
4712 mixed into 'md_local' in small portions; this heuristically
4713 further strengthens the PRNG.
4716 *) Fix crypto/bn/asm/mips3.s.
4719 *) When only the key is given to "enc", the IV is undefined. Print out
4720 an error message in this case.
4723 *) Handle special case when X509_NAME is empty in X509 printing routines.
4726 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
4727 positive and less than q.
4730 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
4731 used: it isn't thread safe and the add_lock_callback should handle
4733 [Paul Rose <Paul.Rose@bridge.com>]
4735 *) Verify that incoming data obeys the block size in
4736 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
4740 [Ulf Möller, Bodo Möller]
4742 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
4743 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
4744 when fixing the server behaviour for backwards-compatible 'client
4745 hello' messages. (Note that the attack is impractical against
4746 SSL 3.0 and TLS 1.0 anyway because length and version checking
4747 means that the probability of guessing a valid ciphertext is
4748 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
4751 Before 0.9.5, the countermeasure (hide the error by generating a
4752 random 'decryption result') did not work properly because
4753 ERR_clear_error() was missing, meaning that SSL_get_error() would
4754 detect the supposedly ignored error.
4756 Both problems are now fixed.
4759 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
4760 (previously it was 1024).
4763 *) Fix for compatibility mode trust settings: ignore trust settings
4764 unless some valid trust or reject settings are present.
4767 *) Fix for blowfish EVP: its a variable length cipher.
4770 *) Fix various bugs related to DSA S/MIME verification. Handle missing
4771 parameters in DSA public key structures and return an error in the
4772 DSA routines if parameters are absent.
4775 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
4776 in the current directory if neither $RANDFILE nor $HOME was set.
4777 RAND_file_name() in 0.9.6a returned NULL in this case. This has
4778 caused some confusion to Windows users who haven't defined $HOME.
4779 Thus RAND_file_name() is changed again: e_os.h can define a
4780 DEFAULT_HOME, which will be used if $HOME is not set.
4781 For Windows, we use "C:"; on other platforms, we still require
4782 environment variables.
4784 *) Move 'if (!initialized) RAND_poll()' into regions protected by
4785 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
4786 having multiple threads call RAND_poll() concurrently.
4789 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
4790 combination of a flag and a thread ID variable.
4791 Otherwise while one thread is in ssleay_rand_bytes (which sets the
4792 flag), *other* threads can enter ssleay_add_bytes without obeying
4793 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
4794 that they do not hold after the first thread unsets add_do_not_lock).
4797 *) Change bctest again: '-x' expressions are not available in all
4801 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
4803 *) Fix a couple of memory leaks in PKCS7_dataDecode()
4804 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
4806 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
4807 the default extension for executables, if any. Also, make the perl
4808 scripts that use symlink() to test if it really exists and use "cp"
4809 if it doesn't. All this made OpenSSL compilable and installable in
4813 *) Fix for asn1_GetSequence() for indefinite length constructed data.
4814 If SEQUENCE is length is indefinite just set c->slen to the total
4815 amount of data available.
4816 [Steve Henson, reported by shige@FreeBSD.org]
4817 [This change does not apply to 0.9.7.]
4819 *) Change bctest to avoid here-documents inside command substitution
4820 (workaround for FreeBSD /bin/sh bug).
4821 For compatibility with Ultrix, avoid shell functions (introduced
4822 in the bctest version that searches along $PATH).
4825 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
4826 with des_encrypt() defined on some operating systems, like Solaris
4830 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
4831 On the Importance of Eliminating Errors in Cryptographic
4832 Computations, J. Cryptology 14 (2001) 2, 101-119,
4833 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
4836 *) MIPS assembler BIGNUM division bug fix.
4839 *) Disabled incorrect Alpha assembler code.
4842 *) Fix PKCS#7 decode routines so they correctly update the length
4843 after reading an EOC for the EXPLICIT tag.
4845 [This change does not apply to 0.9.7.]
4847 *) Fix bug in PKCS#12 key generation routines. This was triggered
4848 if a 3DES key was generated with a 0 initial byte. Include
4849 PKCS12_BROKEN_KEYGEN compilation option to retain the old
4850 (but broken) behaviour.
4853 *) Enhance bctest to search for a working bc along $PATH and print
4855 [Tim Rice <tim@multitalents.net> via Richard Levitte]
4857 *) Fix memory leaks in err.c: free err_data string if necessary;
4858 don't write to the wrong index in ERR_set_error_data.
4861 *) Implement ssl23_peek (analogous to ssl23_read), which previously
4865 *) Replace rdtsc with _emit statements for VC++ version 5.
4866 [Jeremy Cooper <jeremy@baymoo.org>]
4868 *) Make it possible to reuse SSLv2 sessions.
4871 *) In copy_email() check for >= 0 as a return value for
4872 X509_NAME_get_index_by_NID() since 0 is a valid index.
4873 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
4875 *) Avoid coredump with unsupported or invalid public keys by checking if
4876 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
4877 PKCS7_verify() fails with non detached data.
4880 *) Don't use getenv in library functions when run as setuid/setgid.
4881 New function OPENSSL_issetugid().
4884 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
4885 due to incorrect handling of multi-threading:
4887 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
4889 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
4891 3. Count how many times MemCheck_off() has been called so that
4892 nested use can be treated correctly. This also avoids
4893 inband-signalling in the previous code (which relied on the
4894 assumption that thread ID 0 is impossible).
4897 *) Add "-rand" option also to s_client and s_server.
4900 *) Fix CPU detection on Irix 6.x.
4901 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
4902 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4904 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
4907 [This change does not apply to 0.9.7.]
4909 *) Use the cached encoding of an X509_NAME structure rather than
4910 copying it. This is apparently the reason for the libsafe "errors"
4911 but the code is actually correct.
4914 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
4915 Bleichenbacher's DSA attack.
4916 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
4917 to be set and top=0 forces the highest bit to be set; top=-1 is new
4918 and leaves the highest bit random.
4919 [Ulf Moeller, Bodo Moeller]
4921 *) In the NCONF_...-based implementations for CONF_... queries
4922 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
4923 a temporary CONF structure with the data component set to NULL
4924 (which gives segmentation faults in lh_retrieve).
4925 Instead, use NULL for the CONF pointer in CONF_get_string and
4926 CONF_get_number (which may use environment variables) and directly
4927 return NULL from CONF_get_section.
4930 *) Fix potential buffer overrun for EBCDIC.
4933 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
4934 keyUsage if basicConstraints absent for a CA.
4937 *) Make SMIME_write_PKCS7() write mail header values with a format that
4938 is more generally accepted (no spaces before the semicolon), since
4939 some programs can't parse those values properly otherwise. Also make
4940 sure BIO's that break lines after each write do not create invalid
4944 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
4945 macros previously used would not encode an empty SEQUENCE OF
4946 and break the signature.
4948 [This change does not apply to 0.9.7.]
4950 *) Zero the premaster secret after deriving the master secret in
4954 *) Add some EVP_add_digest_alias registrations (as found in
4955 OpenSSL_add_all_digests()) to SSL_library_init()
4956 aka OpenSSL_add_ssl_algorithms(). This provides improved
4957 compatibility with peers using X.509 certificates
4958 with unconventional AlgorithmIdentifier OIDs.
4961 *) Fix for Irix with NO_ASM.
4962 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4964 *) ./config script fixes.
4965 [Ulf Moeller, Richard Levitte]
4967 *) Fix 'openssl passwd -1'.
4970 *) Change PKCS12_key_gen_asc() so it can cope with non null
4971 terminated strings whose length is passed in the passlen
4972 parameter, for example from PEM callbacks. This was done
4973 by adding an extra length parameter to asc2uni().
4974 [Steve Henson, reported by <oddissey@samsung.co.kr>]
4976 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
4977 call failed, free the DSA structure.
4980 *) Fix to uni2asc() to cope with zero length Unicode strings.
4981 These are present in some PKCS#12 files.
4984 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
4985 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
4986 when writing a 32767 byte record.
4987 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
4989 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
4990 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
4992 (RSA objects have a reference count access to which is protected
4993 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
4994 so they are meant to be shared between threads.)
4995 [Bodo Moeller, Geoff Thorpe; original patch submitted by
4996 "Reddie, Steven" <Steven.Reddie@ca.com>]
4998 *) Fix a deadlock in CRYPTO_mem_leaks().
5001 *) Use better test patterns in bntest.
5004 *) rand_win.c fix for Borland C.
5007 *) BN_rshift bugfix for n == 0.
5010 *) Add a 'bctest' script that checks for some known 'bc' bugs
5011 so that 'make test' does not abort just because 'bc' is broken.
5014 *) Store verify_result within SSL_SESSION also for client side to
5015 avoid potential security hole. (Re-used sessions on the client side
5016 always resulted in verify_result==X509_V_OK, not using the original
5017 result of the server certificate verification.)
5020 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
5021 SSL3_RT_APPLICATION_DATA, return 0.
5022 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
5026 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
5027 releases, have been re-implemented by renaming the previous
5028 implementations of ssl2_read and ssl3_read to ssl2_read_internal
5029 and ssl3_read_internal, respectively, and adding 'peek' parameters
5030 to them. The new ssl[23]_{read,peek} functions are calls to
5031 ssl[23]_read_internal with the 'peek' flag set appropriately.
5032 A 'peek' parameter has also been added to ssl3_read_bytes, which
5033 does the actual work for ssl3_read_internal.
5036 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
5037 the method-specific "init()" handler. Also clean up ex_data after
5038 calling the method-specific "finish()" handler. Previously, this was
5039 happening the other way round.
5042 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
5043 The previous value, 12, was not always sufficient for BN_mod_exp().
5046 *) Make sure that shared libraries get the internal name engine with
5047 the full version number and not just 0. This should mark the
5048 shared libraries as not backward compatible. Of course, this should
5049 be changed again when we can guarantee backward binary compatibility.
5052 *) Fix typo in get_cert_by_subject() in by_dir.c
5053 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
5055 *) Rework the system to generate shared libraries:
5057 - Make note of the expected extension for the shared libraries and
5058 if there is a need for symbolic links from for example libcrypto.so.0
5059 to libcrypto.so.0.9.7. There is extended info in Configure for
5062 - Make as few rebuilds of the shared libraries as possible.
5064 - Still avoid linking the OpenSSL programs with the shared libraries.
5066 - When installing, install the shared libraries separately from the
5070 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
5072 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
5073 and not in SSL_clear because the latter is also used by the
5074 accept/connect functions; previously, the settings made by
5075 SSL_set_read_ahead would be lost during the handshake.
5076 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
5078 *) Correct util/mkdef.pl to be selective about disabled algorithms.
5079 Previously, it would create entries for disableed algorithms no
5083 *) Added several new manual pages for SSL_* function.
5086 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
5088 *) In ssl23_get_client_hello, generate an error message when faced
5089 with an initial SSL 3.0/TLS record that is too small to contain the
5090 first two bytes of the ClientHello message, i.e. client_version.
5091 (Note that this is a pathologic case that probably has never happened
5092 in real life.) The previous approach was to use the version number
5093 from the record header as a substitute; but our protocol choice
5094 should not depend on that one because it is not authenticated
5095 by the Finished messages.
5098 *) More robust randomness gathering functions for Windows.
5099 [Jeffrey Altman <jaltman@columbia.edu>]
5101 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
5102 not set then we don't setup the error code for issuer check errors
5103 to avoid possibly overwriting other errors which the callback does
5104 handle. If an application does set the flag then we assume it knows
5105 what it is doing and can handle the new informational codes
5109 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
5110 a general "ANY" type, as such it should be able to decode anything
5111 including tagged types. However it didn't check the class so it would
5112 wrongly interpret tagged types in the same way as their universal
5113 counterpart and unknown types were just rejected. Changed so that the
5114 tagged and unknown types are handled in the same way as a SEQUENCE:
5115 that is the encoding is stored intact. There is also a new type
5116 "V_ASN1_OTHER" which is used when the class is not universal, in this
5117 case we have no idea what the actual type is so we just lump them all
5121 *) On VMS, stdout may very well lead to a file that is written to
5122 in a record-oriented fashion. That means that every write() will
5123 write a separate record, which will be read separately by the
5124 programs trying to read from it. This can be very confusing.
5126 The solution is to put a BIO filter in the way that will buffer
5127 text until a linefeed is reached, and then write everything a
5128 line at a time, so every record written will be an actual line,
5129 not chunks of lines and not (usually doesn't happen, but I've
5130 seen it once) several lines in one record. BIO_f_linebuffer() is
5133 Currently, it's a VMS-only method, because that's where it has
5134 been tested well enough.
5137 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
5138 it can return incorrect results.
5139 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
5140 but it was in 0.9.6-beta[12].)
5143 *) Disable the check for content being present when verifying detached
5144 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
5145 include zero length content when signing messages.
5148 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
5149 BIO_ctrl (for BIO pairs).
5152 *) Add DSO method for VMS.
5155 *) Bug fix: Montgomery multiplication could produce results with the
5159 *) Add RPM specification openssl.spec and modify it to build three
5160 packages. The default package contains applications, application
5161 documentation and run-time libraries. The devel package contains
5162 include files, static libraries and function documentation. The
5163 doc package contains the contents of the doc directory. The original
5164 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
5167 *) Add a large number of documentation files for many SSL routines.
5168 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5170 *) Add a configuration entry for Sony News 4.
5171 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
5173 *) Don't set the two most significant bits to one when generating a
5174 random number < q in the DSA library.
5177 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
5178 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
5179 the underlying transport is blocking) if a handshake took place.
5180 (The default behaviour is needed by applications such as s_client
5181 and s_server that use select() to determine when to use SSL_read;
5182 but for applications that know in advance when to expect data, it
5183 just makes things more complicated.)
5186 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
5190 *) Add a few more EBCDIC conditionals that make `req' and `x509'
5191 work better on such systems.
5192 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5194 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
5195 Update PKCS12_parse() so it copies the friendlyName and the
5196 keyid to the certificates aux info.
5199 *) Fix bug in PKCS7_verify() which caused an infinite loop
5200 if there was more than one signature.
5201 [Sven Uszpelkat <su@celocom.de>]
5203 *) Major change in util/mkdef.pl to include extra information
5204 about each symbol, as well as presentig variables as well
5205 as functions. This change means that there's n more need
5206 to rebuild the .num files when some algorithms are excluded.
5209 *) Allow the verify time to be set by an application,
5210 rather than always using the current time.
5213 *) Phase 2 verify code reorganisation. The certificate
5214 verify code now looks up an issuer certificate by a
5215 number of criteria: subject name, authority key id
5216 and key usage. It also verifies self signed certificates
5217 by the same criteria. The main comparison function is
5218 X509_check_issued() which performs these checks.
5220 Lot of changes were necessary in order to support this
5221 without completely rewriting the lookup code.
5223 Authority and subject key identifier are now cached.
5225 The LHASH 'certs' is X509_STORE has now been replaced
5226 by a STACK_OF(X509_OBJECT). This is mainly because an
5227 LHASH can't store or retrieve multiple objects with
5228 the same hash value.
5230 As a result various functions (which were all internal
5231 use only) have changed to handle the new X509_STORE
5232 structure. This will break anything that messed round
5233 with X509_STORE internally.
5235 The functions X509_STORE_add_cert() now checks for an
5236 exact match, rather than just subject name.
5238 The X509_STORE API doesn't directly support the retrieval
5239 of multiple certificates matching a given criteria, however
5240 this can be worked round by performing a lookup first
5241 (which will fill the cache with candidate certificates)
5242 and then examining the cache for matches. This is probably
5243 the best we can do without throwing out X509_LOOKUP
5244 entirely (maybe later...).
5246 The X509_VERIFY_CTX structure has been enhanced considerably.
5248 All certificate lookup operations now go via a get_issuer()
5249 callback. Although this currently uses an X509_STORE it
5250 can be replaced by custom lookups. This is a simple way
5251 to bypass the X509_STORE hackery necessary to make this
5252 work and makes it possible to use more efficient techniques
5253 in future. A very simple version which uses a simple
5254 STACK for its trusted certificate store is also provided
5255 using X509_STORE_CTX_trusted_stack().
5257 The verify_cb() and verify() callbacks now have equivalents
5258 in the X509_STORE_CTX structure.
5260 X509_STORE_CTX also has a 'flags' field which can be used
5261 to customise the verify behaviour.
5264 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
5265 excludes S/MIME capabilities.
5268 *) When a certificate request is read in keep a copy of the
5269 original encoding of the signed data and use it when outputing
5270 again. Signatures then use the original encoding rather than
5271 a decoded, encoded version which may cause problems if the
5272 request is improperly encoded.
5275 *) For consistency with other BIO_puts implementations, call
5276 buffer_write(b, ...) directly in buffer_puts instead of calling
5279 In BIO_puts, increment b->num_write as in BIO_write.
5280 [Peter.Sylvester@EdelWeb.fr]
5282 *) Fix BN_mul_word for the case where the word is 0. (We have to use
5283 BN_zero, we may not return a BIGNUM with an array consisting of
5287 *) Avoid calling abort() from within the library when problems are
5288 detected, except if preprocessor symbols have been defined
5289 (such as REF_CHECK, BN_DEBUG etc.).
5292 *) New openssl application 'rsautl'. This utility can be
5293 used for low level RSA operations. DER public key
5294 BIO/fp routines also added.
5297 *) New Configure entry and patches for compiling on QNX 4.
5298 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
5300 *) A demo state-machine implementation was sponsored by
5301 Nuron (http://www.nuron.com/) and is now available in
5302 demos/state_machine.
5305 *) New options added to the 'dgst' utility for signature
5306 generation and verification.
5309 *) Unrecognized PKCS#7 content types are now handled via a
5310 catch all ASN1_TYPE structure. This allows unsupported
5311 types to be stored as a "blob" and an application can
5312 encode and decode it manually.
5315 *) Fix various signed/unsigned issues to make a_strex.c
5317 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
5319 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
5320 length if passed a buffer. ASN1_INTEGER_to_BN failed
5321 if passed a NULL BN and its argument was negative.
5322 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
5324 *) Modification to PKCS#7 encoding routines to output definite
5325 length encoding. Since currently the whole structures are in
5326 memory there's not real point in using indefinite length
5327 constructed encoding. However if OpenSSL is compiled with
5328 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
5331 *) Added BIO_vprintf() and BIO_vsnprintf().
5334 *) Added more prefixes to parse for in the the strings written
5335 through a logging bio, to cover all the levels that are available
5336 through syslog. The prefixes are now:
5338 PANIC, EMERG, EMR => LOG_EMERG
5339 ALERT, ALR => LOG_ALERT
5340 CRIT, CRI => LOG_CRIT
5341 ERROR, ERR => LOG_ERR
5342 WARNING, WARN, WAR => LOG_WARNING
5343 NOTICE, NOTE, NOT => LOG_NOTICE
5344 INFO, INF => LOG_INFO
5345 DEBUG, DBG => LOG_DEBUG
5347 and as before, if none of those prefixes are present at the
5348 beginning of the string, LOG_ERR is chosen.
5350 On Win32, the LOG_* levels are mapped according to this:
5352 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
5353 LOG_WARNING => EVENTLOG_WARNING_TYPE
5354 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
5358 *) Made it possible to reconfigure with just the configuration
5359 argument "reconf" or "reconfigure". The command line arguments
5360 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
5361 and are retrieved from there when reconfiguring.
5365 [Assar Westerlund <assar@sics.se>, Richard Levitte]
5367 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
5370 *) The obj_dat.pl script was messing up the sorting of object
5371 names. The reason was that it compared the quoted version
5372 of strings as a result "OCSP" > "OCSP Signing" because
5373 " > SPACE. Changed script to store unquoted versions of
5374 names and add quotes on output. It was also omitting some
5375 names from the lookup table if they were given a default
5376 value (that is if SN is missing it is given the same
5377 value as LN and vice versa), these are now added on the
5378 grounds that if an object has a name we should be able to
5379 look it up. Finally added warning output when duplicate
5380 short or long names are found.
5383 *) Changes needed for Tandem NSK.
5384 [Scott Uroff <scott@xypro.com>]
5386 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
5387 RSA_padding_check_SSLv23(), special padding was never detected
5388 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
5389 version rollback attacks was not effective.
5391 In s23_clnt.c, don't use special rollback-attack detection padding
5392 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
5393 client; similarly, in s23_srvr.c, don't do the rollback check if
5394 SSL 2.0 is the only protocol enabled in the server.
5397 *) Make it possible to get hexdumps of unprintable data with 'openssl
5398 asn1parse'. By implication, the functions ASN1_parse_dump() and
5399 BIO_dump_indent() are added.
5402 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
5403 these print out strings and name structures based on various
5404 flags including RFC2253 support and proper handling of
5405 multibyte characters. Added options to the 'x509' utility
5406 to allow the various flags to be set.
5409 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
5410 Also change the functions X509_cmp_current_time() and
5411 X509_gmtime_adj() work with an ASN1_TIME structure,
5412 this will enable certificates using GeneralizedTime in validity
5413 dates to be checked.
5416 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
5417 negative public key encodings) on by default,
5418 NO_NEG_PUBKEY_BUG can be set to disable it.
5421 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
5422 content octets. An i2c_ASN1_OBJECT is unnecessary because
5423 the encoding can be trivially obtained from the structure.
5426 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
5427 not read locks (CRYPTO_r_[un]lock).
5430 *) A first attempt at creating official support for shared
5431 libraries through configuration. I've kept it so the
5432 default is static libraries only, and the OpenSSL programs
5433 are always statically linked for now, but there are
5434 preparations for dynamic linking in place.
5435 This has been tested on Linux and Tru64.
5438 *) Randomness polling function for Win9x, as described in:
5439 Peter Gutmann, Software Generation of Practically Strong
5443 *) Fix so PRNG is seeded in req if using an already existing
5447 *) New options to smime application. -inform and -outform
5448 allow alternative formats for the S/MIME message including
5449 PEM and DER. The -content option allows the content to be
5450 specified separately. This should allow things like Netscape
5451 form signing output easier to verify.
5454 *) Fix the ASN1 encoding of tags using the 'long form'.
5457 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
5458 STRING types. These convert content octets to and from the
5459 underlying type. The actual tag and length octets are
5460 already assumed to have been read in and checked. These
5461 are needed because all other string types have virtually
5462 identical handling apart from the tag. By having versions
5463 of the ASN1 functions that just operate on content octets
5464 IMPLICIT tagging can be handled properly. It also allows
5465 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
5466 and ASN1_INTEGER are identical apart from the tag.
5469 *) Change the handling of OID objects as follows:
5471 - New object identifiers are inserted in objects.txt, following
5472 the syntax given in objects.README.
5473 - objects.pl is used to process obj_mac.num and create a new
5475 - obj_dat.pl is used to create a new obj_dat.h, using the data in
5478 This is currently kind of a hack, and the perl code in objects.pl
5479 isn't very elegant, but it works as I intended. The simplest way
5480 to check that it worked correctly is to look in obj_dat.h and
5481 check the array nid_objs and make sure the objects haven't moved
5482 around (this is important!). Additions are OK, as well as
5483 consistent name changes.
5486 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
5489 *) Addition of the command line parameter '-rand file' to 'openssl req'.
5490 The given file adds to whatever has already been seeded into the
5491 random pool through the RANDFILE configuration file option or
5492 environment variable, or the default random state file.
5495 *) mkstack.pl now sorts each macro group into lexical order.
5496 Previously the output order depended on the order the files
5497 appeared in the directory, resulting in needless rewriting
5501 *) Patches to make OpenSSL compile under Win32 again. Mostly
5502 work arounds for the VC++ problem that it treats func() as
5503 func(void). Also stripped out the parts of mkdef.pl that
5504 added extra typesafe functions: these no longer exist.
5507 *) Reorganisation of the stack code. The macros are now all
5508 collected in safestack.h . Each macro is defined in terms of
5509 a "stack macro" of the form SKM_<name>(type, a, b). The
5510 DEBUG_SAFESTACK is now handled in terms of function casts,
5511 this has the advantage of retaining type safety without the
5512 use of additional functions. If DEBUG_SAFESTACK is not defined
5513 then the non typesafe macros are used instead. Also modified the
5514 mkstack.pl script to handle the new form. Needs testing to see
5515 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
5516 the default if no major problems. Similar behaviour for ASN1_SET_OF
5517 and PKCS12_STACK_OF.
5520 *) When some versions of IIS use the 'NET' form of private key the
5521 key derivation algorithm is different. Normally MD5(password) is
5522 used as a 128 bit RC4 key. In the modified case
5523 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
5524 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
5525 as the old Netscape_RSA functions except they have an additional
5526 'sgckey' parameter which uses the modified algorithm. Also added
5527 an -sgckey command line option to the rsa utility. Thanks to
5528 Adrian Peck <bertie@ncipher.com> for posting details of the modified
5529 algorithm to openssl-dev.
5532 *) The evp_local.h macros were using 'c.##kname' which resulted in
5533 invalid expansion on some systems (SCO 5.0.5 for example).
5534 Corrected to 'c.kname'.
5535 [Phillip Porch <root@theporch.com>]
5537 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
5538 a STACK of email addresses from a certificate or request, these look
5539 in the subject name and the subject alternative name extensions and
5540 omit any duplicate addresses.
5543 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
5544 This makes DSA verification about 2 % faster.
5547 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
5548 (meaning that now 2^5 values will be precomputed, which is only 4 KB
5549 plus overhead for 1024 bit moduli).
5550 This makes exponentiations about 0.5 % faster for 1024 bit
5551 exponents (as measured by "openssl speed rsa2048").
5554 *) Rename memory handling macros to avoid conflicts with other
5556 Malloc => OPENSSL_malloc
5557 Malloc_locked => OPENSSL_malloc_locked
5558 Realloc => OPENSSL_realloc
5559 Free => OPENSSL_free
5562 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
5563 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
5566 *) CygWin32 support.
5567 [John Jarvie <jjarvie@newsguy.com>]
5569 *) The type-safe stack code has been rejigged. It is now only compiled
5570 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
5571 by default all type-specific stack functions are "#define"d back to
5572 standard stack functions. This results in more streamlined output
5573 but retains the type-safety checking possibilities of the original
5577 *) The STACK code has been cleaned up, and certain type declarations
5578 that didn't make a lot of sense have been brought in line. This has
5579 also involved a cleanup of sorts in safestack.h to more correctly
5580 map type-safe stack functions onto their plain stack counterparts.
5581 This work has also resulted in a variety of "const"ifications of
5582 lots of the code, especially "_cmp" operations which should normally
5583 be prototyped with "const" parameters anyway.
5586 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
5587 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
5588 (The PRNG state consists of two parts, the large pool 'state' and 'md',
5589 where all of 'md' is used each time the PRNG is used, but 'state'
5590 is used only indexed by a cyclic counter. As entropy may not be
5591 well distributed from the beginning, 'md' is important as a
5592 chaining variable. However, the output function chains only half
5593 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
5594 all of 'md', and seeding with STATE_SIZE dummy bytes will result
5595 in all of 'state' being rewritten, with the new values depending
5596 on virtually all of 'md'. This overcomes the 80 bit limitation.)
5599 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
5600 the handshake is continued after ssl_verify_cert_chain();
5601 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
5602 can lead to 'unexplainable' connection aborts later.
5603 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
5605 *) Major EVP API cipher revision.
5606 Add hooks for extra EVP features. This allows various cipher
5607 parameters to be set in the EVP interface. Support added for variable
5608 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
5609 setting of RC2 and RC5 parameters.
5611 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
5614 Remove lots of duplicated code from the EVP library. For example *every*
5615 cipher init() function handles the 'iv' in the same way according to the
5616 cipher mode. They also all do nothing if the 'key' parameter is NULL and
5617 for CFB and OFB modes they zero ctx->num.
5619 New functionality allows removal of S/MIME code RC2 hack.
5621 Most of the routines have the same form and so can be declared in terms
5624 By shifting this to the top level EVP_CipherInit() it can be removed from
5625 all individual ciphers. If the cipher wants to handle IVs or keys
5626 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
5629 Change lots of functions like EVP_EncryptUpdate() to now return a
5630 value: although software versions of the algorithms cannot fail
5631 any installed hardware versions can.
5634 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
5635 this option is set, tolerate broken clients that send the negotiated
5636 protocol version number instead of the requested protocol version
5640 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
5641 i.e. non-zero for export ciphersuites, zero otherwise.
5642 Previous versions had this flag inverted, inconsistent with
5643 rsa_tmp_cb (..._TMP_RSA_CB).
5644 [Bodo Moeller; problem reported by Amit Chopra]
5646 *) Add missing DSA library text string. Work around for some IIS
5647 key files with invalid SEQUENCE encoding.
5650 *) Add a document (doc/standards.txt) that list all kinds of standards
5651 and so on that are implemented in OpenSSL.
5654 *) Enhance c_rehash script. Old version would mishandle certificates
5655 with the same subject name hash and wouldn't handle CRLs at all.
5656 Added -fingerprint option to crl utility, to support new c_rehash
5660 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
5663 *) Fix for SSL server purpose checking. Server checking was
5664 rejecting certificates which had extended key usage present
5665 but no ssl client purpose.
5666 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
5668 *) Make PKCS#12 code work with no password. The PKCS#12 spec
5669 is a little unclear about how a blank password is handled.
5670 Since the password in encoded as a BMPString with terminating
5671 double NULL a zero length password would end up as just the
5672 double NULL. However no password at all is different and is
5673 handled differently in the PKCS#12 key generation code. NS
5674 treats a blank password as zero length. MSIE treats it as no
5675 password on export: but it will try both on import. We now do
5676 the same: PKCS12_parse() tries zero length and no password if
5677 the password is set to "" or NULL (NULL is now a valid password:
5678 it wasn't before) as does the pkcs12 application.
5681 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
5682 perror when PEM_read_bio_X509_REQ fails, the error message must
5683 be obtained from the error queue.
5686 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
5687 it in ERR_remove_state if appropriate, and change ERR_get_state
5688 accordingly to avoid race conditions (this is necessary because
5689 thread_hash is no longer constant once set).
5692 *) Bugfix for linux-elf makefile.one.
5695 *) RSA_get_default_method() will now cause a default
5696 RSA_METHOD to be chosen if one doesn't exist already.
5697 Previously this was only set during a call to RSA_new()
5698 or RSA_new_method(NULL) meaning it was possible for
5699 RSA_get_default_method() to return NULL.
5702 *) Added native name translation to the existing DSO code
5703 that will convert (if the flag to do so is set) filenames
5704 that are sufficiently small and have no path information
5705 into a canonical native form. Eg. "blah" converted to
5706 "libblah.so" or "blah.dll" etc.
5709 *) New function ERR_error_string_n(e, buf, len) which is like
5710 ERR_error_string(e, buf), but writes at most 'len' bytes
5711 including the 0 terminator. For ERR_error_string_n, 'buf'
5713 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
5715 *) CONF library reworked to become more general. A new CONF
5716 configuration file reader "class" is implemented as well as a
5717 new functions (NCONF_*, for "New CONF") to handle it. The now
5718 old CONF_* functions are still there, but are reimplemented to
5719 work in terms of the new functions. Also, a set of functions
5720 to handle the internal storage of the configuration data is
5721 provided to make it easier to write new configuration file
5722 reader "classes" (I can definitely see something reading a
5723 configuration file in XML format, for example), called _CONF_*,
5724 or "the configuration storage API"...
5726 The new configuration file reading functions are:
5728 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
5729 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
5731 NCONF_default, NCONF_WIN32
5733 NCONF_dump_fp, NCONF_dump_bio
5735 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
5736 NCONF_new creates a new CONF object. This works in the same way
5737 as other interfaces in OpenSSL, like the BIO interface.
5738 NCONF_dump_* dump the internal storage of the configuration file,
5739 which is useful for debugging. All other functions take the same
5740 arguments as the old CONF_* functions wth the exception of the
5741 first that must be a `CONF *' instead of a `LHASH *'.
5743 To make it easer to use the new classes with the old CONF_* functions,
5744 the function CONF_set_default_method is provided.
5747 *) Add '-tls1' option to 'openssl ciphers', which was already
5748 mentioned in the documentation but had not been implemented.
5749 (This option is not yet really useful because even the additional
5750 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
5753 *) Initial DSO code added into libcrypto for letting OpenSSL (and
5754 OpenSSL-based applications) load shared libraries and bind to
5755 them in a portable way.
5756 [Geoff Thorpe, with contributions from Richard Levitte]
5758 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
5760 *) Make sure _lrotl and _lrotr are only used with MSVC.
5762 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
5763 (the default implementation of RAND_status).
5765 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
5766 to '-clrext' (= clear extensions), as intended and documented.
5767 [Bodo Moeller; inconsistency pointed out by Michael Attili
5770 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
5771 was larger than the MD block size.
5772 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
5774 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
5775 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
5776 using the passed key: if the passed key was a private key the result
5777 of X509_print(), for example, would be to print out all the private key
5781 *) des_quad_cksum() byte order bug fix.
5782 [Ulf Möller, using the problem description in krb4-0.9.7, where
5783 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
5785 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
5787 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
5789 *) For easily testing in shell scripts whether some command
5790 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
5791 returns with exit code 0 iff no command of the given name is available.
5792 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
5793 the output goes to stdout and nothing is printed to stderr.
5794 Additional arguments are always ignored.
5796 Since for each cipher there is a command of the same name,
5797 the 'no-cipher' compilation switches can be tested this way.
5799 ('openssl no-XXX' is not able to detect pseudo-commands such
5800 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
5803 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
5806 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
5807 is set; it will be thrown away anyway because each handshake creates
5809 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
5810 to parameters -- in previous versions (since OpenSSL 0.9.3) the
5811 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
5812 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
5815 *) New s_client option -ign_eof: EOF at stdin is ignored, and
5816 'Q' and 'R' lose their special meanings (quit/renegotiate).
5817 This is part of what -quiet does; unlike -quiet, -ign_eof
5818 does not suppress any output.
5821 *) Add compatibility options to the purpose and trust code. The
5822 purpose X509_PURPOSE_ANY is "any purpose" which automatically
5823 accepts a certificate or CA, this was the previous behaviour,
5824 with all the associated security issues.
5826 X509_TRUST_COMPAT is the old trust behaviour: only and
5827 automatically trust self signed roots in certificate store. A
5828 new trust setting X509_TRUST_DEFAULT is used to specify that
5829 a purpose has no associated trust setting and it should instead
5830 use the value in the default purpose.
5833 *) Fix the PKCS#8 DSA private key code so it decodes keys again
5834 and fix a memory leak.
5837 *) In util/mkerr.pl (which implements 'make errors'), preserve
5838 reason strings from the previous version of the .c file, as
5839 the default to have only downcase letters (and digits) in
5840 automatically generated reasons codes is not always appropriate.
5843 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
5844 using strerror. Previously, ERR_reason_error_string() returned
5845 library names as reason strings for SYSerr; but SYSerr is a special
5846 case where small numbers are errno values, not library numbers.
5849 *) Add '-dsaparam' option to 'openssl dhparam' application. This
5850 converts DSA parameters into DH parameters. (When creating parameters,
5851 DSA_generate_parameters is used.)
5854 *) Include 'length' (recommended exponent length) in C code generated
5855 by 'openssl dhparam -C'.
5858 *) The second argument to set_label in perlasm was already being used
5859 so couldn't be used as a "file scope" flag. Moved to third argument
5863 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
5864 instead of RAND_bytes for encryption IVs and salts.
5867 *) Include RAND_status() into RAND_METHOD instead of implementing
5868 it only for md_rand.c Otherwise replacing the PRNG by calling
5869 RAND_set_rand_method would be impossible.
5872 *) Don't let DSA_generate_key() enter an infinite loop if the random
5873 number generation fails.
5876 *) New 'rand' application for creating pseudo-random output.
5879 *) Added configuration support for Linux/IA64
5880 [Rolf Haberrecker <rolf@suse.de>]
5882 *) Assembler module support for Mingw32.
5885 *) Shared library support for HPUX (in shlib/).
5886 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
5888 *) Shared library support for Solaris gcc.
5889 [Lutz Behnke <behnke@trustcenter.de>]
5891 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
5893 *) PKCS7_encrypt() was adding text MIME headers twice because they
5894 were added manually and by SMIME_crlf_copy().
5897 *) In bntest.c don't call BN_rand with zero bits argument.
5898 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
5900 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
5901 case was implemented. This caused BN_div_recp() to fail occasionally.
5904 *) Add an optional second argument to the set_label() in the perl
5905 assembly language builder. If this argument exists and is set
5906 to 1 it signals that the assembler should use a symbol whose
5907 scope is the entire file, not just the current function. This
5908 is needed with MASM which uses the format label:: for this scope.
5909 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
5911 *) Change the ASN1 types so they are typedefs by default. Before
5912 almost all types were #define'd to ASN1_STRING which was causing
5913 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
5917 *) Change names of new functions to the new get1/get0 naming
5918 convention: After 'get1', the caller owns a reference count
5919 and has to call ..._free; 'get0' returns a pointer to some
5920 data structure without incrementing reference counters.
5921 (Some of the existing 'get' functions increment a reference
5922 counter, some don't.)
5923 Similarly, 'set1' and 'add1' functions increase reference
5924 counters or duplicate objects.
5927 *) Allow for the possibility of temp RSA key generation failure:
5928 the code used to assume it always worked and crashed on failure.
5931 *) Fix potential buffer overrun problem in BIO_printf().
5932 [Ulf Möller, using public domain code by Patrick Powell; problem
5933 pointed out by David Sacerdote <das33@cornell.edu>]
5935 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
5936 RAND_egd() and RAND_status(). In the command line application,
5937 the EGD socket can be specified like a seed file using RANDFILE
5941 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
5942 Some CAs (e.g. Verisign) distribute certificates in this form.
5945 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
5946 list to exclude them. This means that no special compilation option
5947 is needed to use anonymous DH: it just needs to be included in the
5951 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
5952 EVP_MD_type. The old functionality is available in a new macro called
5953 EVP_MD_md(). Change code that uses it and update docs.
5956 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
5957 where the 'void *' argument is replaced by a function pointer argument.
5958 Previously 'void *' was abused to point to functions, which works on
5959 many platforms, but is not correct. As these functions are usually
5960 called by macros defined in OpenSSL header files, most source code
5961 should work without changes.
5964 *) <openssl/opensslconf.h> (which is created by Configure) now contains
5965 sections with information on -D... compiler switches used for
5966 compiling the library so that applications can see them. To enable
5967 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
5968 must be defined. E.g.,
5969 #define OPENSSL_ALGORITHM_DEFINES
5970 #include <openssl/opensslconf.h>
5971 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
5972 [Richard Levitte, Ulf and Bodo Möller]
5974 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
5978 *) Change the 'other' type in certificate aux info to a STACK_OF
5979 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
5980 the required ASN1 format: arbitrary types determined by an OID.
5983 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
5984 argument to 'req'. This is not because the function is newer or
5985 better than others it just uses the work 'NEW' in the certificate
5986 request header lines. Some software needs this.
5989 *) Reorganise password command line arguments: now passwords can be
5990 obtained from various sources. Delete the PEM_cb function and make
5991 it the default behaviour: i.e. if the callback is NULL and the
5992 usrdata argument is not NULL interpret it as a null terminated pass
5993 phrase. If usrdata and the callback are NULL then the pass phrase
5994 is prompted for as usual.
5997 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
5998 the support is automatically enabled. The resulting binaries will
5999 autodetect the card and use it if present.
6000 [Ben Laurie and Compaq Inc.]
6002 *) Work around for Netscape hang bug. This sends certificate request
6003 and server done in one record. Since this is perfectly legal in the
6004 SSL/TLS protocol it isn't a "bug" option and is on by default. See
6005 the bugs/SSLv3 entry for more info.
6008 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
6011 *) Add -rand argument to smime and pkcs12 applications and read/write
6015 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
6018 *) Add command line password options to the remaining applications.
6021 *) Bug fix for BN_div_recp() for numerators with an even number of
6025 *) More tests in bntest.c, and changed test_bn output.
6028 *) ./config recognizes MacOS X now.
6031 *) Bug fix for BN_div() when the first words of num and divsor are
6032 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
6035 *) Add support for various broken PKCS#8 formats, and command line
6036 options to produce them.
6039 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
6040 get temporary BIGNUMs from a BN_CTX.
6043 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
6047 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
6048 include a #define from the old name to the new. The original intent
6049 was that statically linked binaries could for example just call
6050 SSLeay_add_all_ciphers() to just add ciphers to the table and not
6051 link with digests. This never worked becayse SSLeay_add_all_digests()
6052 and SSLeay_add_all_ciphers() were in the same source file so calling
6053 one would link with the other. They are now in separate source files.
6056 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
6059 *) Use a less unusual form of the Miller-Rabin primality test (it used
6060 a binary algorithm for exponentiation integrated into the Miller-Rabin
6061 loop, our standard modexp algorithms are faster).
6064 *) Support for the EBCDIC character set completed.
6065 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
6067 *) Source code cleanups: use const where appropriate, eliminate casts,
6068 use void * instead of char * in lhash.
6071 *) Bugfix: ssl3_send_server_key_exchange was not restartable
6072 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
6073 this the server could overwrite ephemeral keys that the client
6077 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
6078 using 50 iterations of the Rabin-Miller test.
6080 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
6081 iterations of the Rabin-Miller test as required by the appendix
6082 to FIPS PUB 186[-1]) instead of DSA_is_prime.
6083 As BN_is_prime_fasttest includes trial division, DSA parameter
6084 generation becomes much faster.
6086 This implies a change for the callback functions in DSA_is_prime
6087 and DSA_generate_parameters: The callback function is called once
6088 for each positive witness in the Rabin-Miller test, not just
6089 occasionally in the inner loop; and the parameters to the
6090 callback function now provide an iteration count for the outer
6091 loop rather than for the current invocation of the inner loop.
6092 DSA_generate_parameters additionally can call the callback
6093 function with an 'iteration count' of -1, meaning that a
6094 candidate has passed the trial division test (when q is generated
6095 from an application-provided seed, trial division is skipped).
6098 *) New function BN_is_prime_fasttest that optionally does trial
6099 division before starting the Rabin-Miller test and has
6100 an additional BN_CTX * argument (whereas BN_is_prime always
6101 has to allocate at least one BN_CTX).
6102 'callback(1, -1, cb_arg)' is called when a number has passed the
6103 trial division stage.
6106 *) Fix for bug in CRL encoding. The validity dates weren't being handled
6110 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
6113 *) New function BN_pseudo_rand().
6116 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
6117 bignum version of BN_from_montgomery() with the working code from
6118 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
6122 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
6123 made it impossible to use the same SSL_SESSION data structure in
6124 SSL2 clients in multiple threads.
6127 *) The return value of RAND_load_file() no longer counts bytes obtained
6128 by stat(). RAND_load_file(..., -1) is new and uses the complete file
6129 to seed the PRNG (previously an explicit byte count was required).
6130 [Ulf Möller, Bodo Möller]
6132 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
6133 used (char *) instead of (void *) and had casts all over the place.
6136 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
6139 *) Retain source code compatibility for BN_prime_checks macro:
6140 BN_is_prime(..., BN_prime_checks, ...) now uses
6141 BN_prime_checks_for_size to determine the appropriate number of
6142 Rabin-Miller iterations.
6145 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
6146 DH_CHECK_P_NOT_SAFE_PRIME.
6147 (Check if this is true? OpenPGP calls them "strong".)
6150 *) Merge the functionality of "dh" and "gendh" programs into a new program
6151 "dhparam". The old programs are retained for now but will handle DH keys
6152 (instead of parameters) in future.
6155 *) Make the ciphers, s_server and s_client programs check the return values
6156 when a new cipher list is set.
6159 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
6160 ciphers. Before when the 56bit ciphers were enabled the sorting was
6163 The syntax for the cipher sorting has been extended to support sorting by
6164 cipher-strength (using the strength_bits hard coded in the tables).
6165 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
6167 Fix a bug in the cipher-command parser: when supplying a cipher command
6168 string with an "undefined" symbol (neither command nor alphanumeric
6169 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
6170 an error is flagged.
6172 Due to the strength-sorting extension, the code of the
6173 ssl_create_cipher_list() function was completely rearranged. I hope that
6174 the readability was also increased :-)
6175 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6177 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
6178 for the first serial number and places 2 in the serial number file. This
6179 avoids problems when the root CA is created with serial number zero and
6180 the first user certificate has the same issuer name and serial number
6184 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
6185 the new code. Add documentation for this stuff.
6188 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
6189 X509_*() to X509at_*() on the grounds that they don't handle X509
6190 structures and behave in an analagous way to the X509v3 functions:
6191 they shouldn't be called directly but wrapper functions should be used
6194 So we also now have some wrapper functions that call the X509at functions
6195 when passed certificate requests. (TO DO: similar things can be done with
6196 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
6197 things. Some of these need some d2i or i2d and print functionality
6198 because they handle more complex structures.)
6201 *) Add missing #ifndefs that caused missing symbols when building libssl
6202 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
6203 NO_RSA in ssl/s2*.c.
6204 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
6206 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
6207 has a return value which indicates the quality of the random data
6208 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
6209 error queue. New function RAND_pseudo_bytes() generates output that is
6210 guaranteed to be unique but not unpredictable. RAND_add is like
6211 RAND_seed, but takes an extra argument for an entropy estimate
6212 (RAND_seed always assumes full entropy).
6215 *) Do more iterations of Rabin-Miller probable prime test (specifically,
6216 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
6217 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
6218 in crypto/bn/bn_prime.c for the complete table). This guarantees a
6219 false-positive rate of at most 2^-80 for random input.
6222 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
6225 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
6226 in the 0.9.5 release), this returns the chain
6227 from an X509_CTX structure with a dup of the stack and all
6228 the X509 reference counts upped: so the stack will exist
6229 after X509_CTX_cleanup() has been called. Modify pkcs12.c
6232 Also make SSL_SESSION_print() print out the verify return
6236 *) Add manpage for the pkcs12 command. Also change the default
6237 behaviour so MAC iteration counts are used unless the new
6238 -nomaciter option is used. This improves file security and
6239 only older versions of MSIE (4.0 for example) need it.
6242 *) Honor the no-xxx Configure options when creating .DEF files.
6245 *) Add PKCS#10 attributes to field table: challengePassword,
6246 unstructuredName and unstructuredAddress. These are taken from
6247 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
6248 international characters are used.
6250 More changes to X509_ATTRIBUTE code: allow the setting of types
6251 based on strings. Remove the 'loc' parameter when adding
6252 attributes because these will be a SET OF encoding which is sorted
6256 *) Initial changes to the 'req' utility to allow request generation
6257 automation. This will allow an application to just generate a template
6258 file containing all the field values and have req construct the
6261 Initial support for X509_ATTRIBUTE handling. Stacks of these are
6262 used all over the place including certificate requests and PKCS#7
6263 structures. They are currently handled manually where necessary with
6264 some primitive wrappers for PKCS#7. The new functions behave in a
6265 manner analogous to the X509 extension functions: they allow
6266 attributes to be looked up by NID and added.
6268 Later something similar to the X509V3 code would be desirable to
6269 automatically handle the encoding, decoding and printing of the
6270 more complex types. The string types like challengePassword can
6271 be handled by the string table functions.
6273 Also modified the multi byte string table handling. Now there is
6274 a 'global mask' which masks out certain types. The table itself
6275 can use the flag STABLE_NO_MASK to ignore the mask setting: this
6276 is useful when for example there is only one permissible type
6277 (as in countryName) and using the mask might result in no valid
6281 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
6282 SSL_get_peer_finished to allow applications to obtain the latest
6283 Finished messages sent to the peer or expected from the peer,
6284 respectively. (SSL_get_peer_finished is usually the Finished message
6285 actually received from the peer, otherwise the protocol will be aborted.)
6287 As the Finished message are message digests of the complete handshake
6288 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
6289 be used for external authentication procedures when the authentication
6290 provided by SSL/TLS is not desired or is not enough.
6293 *) Enhanced support for Alpha Linux is added. Now ./config checks if
6294 the host supports BWX extension and if Compaq C is present on the
6295 $PATH. Just exploiting of the BWX extension results in 20-30%
6296 performance kick for some algorithms, e.g. DES and RC4 to mention
6297 a couple. Compaq C in turn generates ~20% faster code for MD5 and
6301 *) Add support for MS "fast SGC". This is arguably a violation of the
6302 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
6303 weak crypto and after checking the certificate is SGC a second one
6304 with strong crypto. MS SGC stops the first handshake after receiving
6305 the server certificate message and sends a second client hello. Since
6306 a server will typically do all the time consuming operations before
6307 expecting any further messages from the client (server key exchange
6308 is the most expensive) there is little difference between the two.
6310 To get OpenSSL to support MS SGC we have to permit a second client
6311 hello message after we have sent server done. In addition we have to
6312 reset the MAC if we do get this second client hello.
6315 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
6316 if a DER encoded private key is RSA or DSA traditional format. Changed
6317 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
6318 format DER encoded private key. Newer code should use PKCS#8 format which
6319 has the key type encoded in the ASN1 structure. Added DER private key
6320 support to pkcs8 application.
6323 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
6324 ciphersuites has been selected (as required by the SSL 3/TLS 1
6325 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
6326 is set, we interpret this as a request to violate the specification
6327 (the worst that can happen is a handshake failure, and 'correct'
6328 behaviour would result in a handshake failure anyway).
6331 *) In SSL_CTX_add_session, take into account that there might be multiple
6332 SSL_SESSION structures with the same session ID (e.g. when two threads
6333 concurrently obtain them from an external cache).
6334 The internal cache can handle only one SSL_SESSION with a given ID,
6335 so if there's a conflict, we now throw out the old one to achieve
6339 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
6340 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
6341 some routines that use cipher OIDs: some ciphers do not have OIDs
6342 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
6346 *) Simplify the trust setting structure and code. Now we just have
6347 two sequences of OIDs for trusted and rejected settings. These will
6348 typically have values the same as the extended key usage extension
6349 and any application specific purposes.
6351 The trust checking code now has a default behaviour: it will just
6352 check for an object with the same NID as the passed id. Functions can
6353 be provided to override either the default behaviour or the behaviour
6354 for a given id. SSL client, server and email already have functions
6355 in place for compatibility: they check the NID and also return "trusted"
6356 if the certificate is self signed.
6359 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
6360 traditional format into an EVP_PKEY structure.
6363 *) Add a password callback function PEM_cb() which either prompts for
6364 a password if usr_data is NULL or otherwise assumes it is a null
6365 terminated password. Allow passwords to be passed on command line
6366 environment or config files in a few more utilities.
6369 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
6370 keys. Add some short names for PKCS#8 PBE algorithms and allow them
6371 to be specified on the command line for the pkcs8 and pkcs12 utilities.
6372 Update documentation.
6375 *) Support for ASN1 "NULL" type. This could be handled before by using
6376 ASN1_TYPE but there wasn't any function that would try to read a NULL
6377 and produce an error if it couldn't. For compatibility we also have
6378 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
6379 don't allocate anything because they don't need to.
6382 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
6384 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
6386 *) Rebuild of the memory allocation routines used by OpenSSL code and
6387 possibly others as well. The purpose is to make an interface that
6388 provide hooks so anyone can build a separate set of allocation and
6389 deallocation routines to be used by OpenSSL, for example memory
6390 pool implementations, or something else, which was previously hard
6391 since Malloc(), Realloc() and Free() were defined as macros having
6392 the values malloc, realloc and free, respectively (except for Win32
6393 compilations). The same is provided for memory debugging code.
6394 OpenSSL already comes with functionality to find memory leaks, but
6395 this gives people a chance to debug other memory problems.
6397 With these changes, a new set of functions and macros have appeared:
6399 CRYPTO_set_mem_debug_functions() [F]
6400 CRYPTO_get_mem_debug_functions() [F]
6401 CRYPTO_dbg_set_options() [F]
6402 CRYPTO_dbg_get_options() [F]
6403 CRYPTO_malloc_debug_init() [M]
6405 The memory debug functions are NULL by default, unless the library
6406 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
6407 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
6408 gives the standard debugging functions that come with OpenSSL) or
6409 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
6410 provided by the library user) must be used. When the standard
6411 debugging functions are used, CRYPTO_dbg_set_options can be used to
6412 request additional information:
6413 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
6414 the CRYPTO_MDEBUG_xxx macro when compiling the library.
6416 Also, things like CRYPTO_set_mem_functions will always give the
6417 expected result (the new set of functions is used for allocation
6418 and deallocation) at all times, regardless of platform and compiler
6421 To finish it up, some functions that were never use in any other
6422 way than through macros have a new API and new semantic:
6425 CRYPTO_dbg_realloc()
6428 All macros of value have retained their old syntax.
6429 [Richard Levitte and Bodo Moeller]
6431 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
6432 ordering of SMIMECapabilities wasn't in "strength order" and there
6433 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
6437 *) Some ASN1 types with illegal zero length encoding (INTEGER,
6438 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
6439 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
6441 *) Merge in my S/MIME library for OpenSSL. This provides a simple
6442 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
6443 functionality to handle multipart/signed properly) and a utility
6444 called 'smime' to call all this stuff. This is based on code I
6445 originally wrote for Celo who have kindly allowed it to be
6446 included in OpenSSL.
6449 *) Add variants des_set_key_checked and des_set_key_unchecked of
6450 des_set_key (aka des_key_sched). Global variable des_check_key
6451 decides which of these is called by des_set_key; this way
6452 des_check_key behaves as it always did, but applications and
6453 the library itself, which was buggy for des_check_key == 1,
6454 have a cleaner way to pick the version they need.
6457 *) New function PKCS12_newpass() which changes the password of a
6461 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
6462 dynamic mix. In both cases the ids can be used as an index into the
6463 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
6464 functions so they accept a list of the field values and the
6465 application doesn't need to directly manipulate the X509_TRUST
6469 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
6473 *) Modify the way the V3 extension code looks up extensions. This now
6474 works in a similar way to the object code: we have some "standard"
6475 extensions in a static table which is searched with OBJ_bsearch()
6476 and the application can add dynamic ones if needed. The file
6477 crypto/x509v3/ext_dat.h now has the info: this file needs to be
6478 updated whenever a new extension is added to the core code and kept
6479 in ext_nid order. There is a simple program 'tabtest.c' which checks
6480 this. New extensions are not added too often so this file can readily
6481 be maintained manually.
6483 There are two big advantages in doing things this way. The extensions
6484 can be looked up immediately and no longer need to be "added" using
6485 X509V3_add_standard_extensions(): this function now does nothing.
6486 [Side note: I get *lots* of email saying the extension code doesn't
6487 work because people forget to call this function]
6488 Also no dynamic allocation is done unless new extensions are added:
6489 so if we don't add custom extensions there is no need to call
6490 X509V3_EXT_cleanup().
6493 *) Modify enc utility's salting as follows: make salting the default. Add a
6494 magic header, so unsalted files fail gracefully instead of just decrypting
6495 to garbage. This is because not salting is a big security hole, so people
6496 should be discouraged from doing it.
6499 *) Fixes and enhancements to the 'x509' utility. It allowed a message
6500 digest to be passed on the command line but it only used this
6501 parameter when signing a certificate. Modified so all relevant
6502 operations are affected by the digest parameter including the
6503 -fingerprint and -x509toreq options. Also -x509toreq choked if a
6504 DSA key was used because it didn't fix the digest.
6507 *) Initial certificate chain verify code. Currently tests the untrusted
6508 certificates for consistency with the verify purpose (which is set
6509 when the X509_STORE_CTX structure is set up) and checks the pathlength.
6511 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
6512 this is because it will reject chains with invalid extensions whereas
6513 every previous version of OpenSSL and SSLeay made no checks at all.
6515 Trust code: checks the root CA for the relevant trust settings. Trust
6516 settings have an initial value consistent with the verify purpose: e.g.
6517 if the verify purpose is for SSL client use it expects the CA to be
6518 trusted for SSL client use. However the default value can be changed to
6519 permit custom trust settings: one example of this would be to only trust
6520 certificates from a specific "secure" set of CAs.
6522 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
6523 which should be used for version portability: especially since the
6524 verify structure is likely to change more often now.
6526 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
6527 to set them. If not set then assume SSL clients will verify SSL servers
6530 Two new options to the verify program: -untrusted allows a set of
6531 untrusted certificates to be passed in and -purpose which sets the
6532 intended purpose of the certificate. If a purpose is set then the
6533 new chain verify code is used to check extension consistency.
6536 *) Support for the authority information access extension.
6539 *) Modify RSA and DSA PEM read routines to transparently handle
6540 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
6541 public keys in a format compatible with certificate
6542 SubjectPublicKeyInfo structures. Unfortunately there were already
6543 functions called *_PublicKey_* which used various odd formats so
6544 these are retained for compatibility: however the DSA variants were
6545 never in a public release so they have been deleted. Changed dsa/rsa
6546 utilities to handle the new format: note no releases ever handled public
6547 keys so we should be OK.
6549 The primary motivation for this change is to avoid the same fiasco
6550 that dogs private keys: there are several incompatible private key
6551 formats some of which are standard and some OpenSSL specific and
6552 require various evil hacks to allow partial transparent handling and
6553 even then it doesn't work with DER formats. Given the option anything
6554 other than PKCS#8 should be dumped: but the other formats have to
6555 stay in the name of compatibility.
6557 With public keys and the benefit of hindsight one standard format
6558 is used which works with EVP_PKEY, RSA or DSA structures: though
6559 it clearly returns an error if you try to read the wrong kind of key.
6561 Added a -pubkey option to the 'x509' utility to output the public key.
6562 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
6563 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
6564 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
6565 that do the same as the EVP_PKEY_assign_*() except they up the
6566 reference count of the added key (they don't "swallow" the
6570 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
6571 CRLs would fail if the file contained no certificates or no CRLs:
6572 added a new function to read in both types and return the number
6573 read: this means that if none are read it will be an error. The
6574 DER versions of the certificate and CRL reader would always fail
6575 because it isn't possible to mix certificates and CRLs in DER format
6576 without choking one or the other routine. Changed this to just read
6577 a certificate: this is the best we can do. Also modified the code
6578 in apps/verify.c to take notice of return codes: it was previously
6579 attempting to read in certificates from NULL pointers and ignoring
6580 any errors: this is one reason why the cert and CRL reader seemed
6581 to work. It doesn't check return codes from the default certificate
6582 routines: these may well fail if the certificates aren't installed.
6585 *) Code to support otherName option in GeneralName.
6588 *) First update to verify code. Change the verify utility
6589 so it warns if it is passed a self signed certificate:
6590 for consistency with the normal behaviour. X509_verify
6591 has been modified to it will now verify a self signed
6592 certificate if *exactly* the same certificate appears
6593 in the store: it was previously impossible to trust a
6594 single self signed certificate. This means that:
6595 openssl verify ss.pem
6596 now gives a warning about a self signed certificate but
6597 openssl verify -CAfile ss.pem ss.pem
6601 *) For servers, store verify_result in SSL_SESSION data structure
6602 (and add it to external session representation).
6603 This is needed when client certificate verifications fails,
6604 but an application-provided verification callback (set by
6605 SSL_CTX_set_cert_verify_callback) allows accepting the session
6606 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
6607 but returns 1): When the session is reused, we have to set
6608 ssl->verify_result to the appropriate error code to avoid
6610 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
6612 *) Fix a bug in the new PKCS#7 code: it didn't consider the
6613 case in PKCS7_dataInit() where the signed PKCS7 structure
6614 didn't contain any existing data because it was being created.
6615 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
6617 *) Add a salt to the key derivation routines in enc.c. This
6618 forms the first 8 bytes of the encrypted file. Also add a
6619 -S option to allow a salt to be input on the command line.
6622 *) New function X509_cmp(). Oddly enough there wasn't a function
6623 to compare two certificates. We do this by working out the SHA1
6624 hash and comparing that. X509_cmp() will be needed by the trust
6628 *) SSL_get1_session() is like SSL_get_session(), but increments
6629 the reference count in the SSL_SESSION returned.
6630 [Geoff Thorpe <geoff@eu.c2.net>]
6632 *) Fix for 'req': it was adding a null to request attributes.
6633 Also change the X509_LOOKUP and X509_INFO code to handle
6634 certificate auxiliary information.
6637 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
6641 *) Add the possibility to add extra information to the memory leak
6642 detecting output, to form tracebacks, showing from where each
6643 allocation was originated: CRYPTO_push_info("constant string") adds
6644 the string plus current file name and line number to a per-thread
6645 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
6646 is like calling CYRPTO_pop_info() until the stack is empty.
6647 Also updated memory leak detection code to be multi-thread-safe.
6650 *) Add options -text and -noout to pkcs7 utility and delete the
6651 encryption options which never did anything. Update docs.
6654 *) Add options to some of the utilities to allow the pass phrase
6655 to be included on either the command line (not recommended on
6656 OSes like Unix) or read from the environment. Update the
6657 manpages and fix a few bugs.
6660 *) Add a few manpages for some of the openssl commands.
6663 *) Fix the -revoke option in ca. It was freeing up memory twice,
6664 leaking and not finding already revoked certificates.
6667 *) Extensive changes to support certificate auxiliary information.
6668 This involves the use of X509_CERT_AUX structure and X509_AUX
6669 functions. An X509_AUX function such as PEM_read_X509_AUX()
6670 can still read in a certificate file in the usual way but it
6671 will also read in any additional "auxiliary information". By
6672 doing things this way a fair degree of compatibility can be
6673 retained: existing certificates can have this information added
6674 using the new 'x509' options.
6676 Current auxiliary information includes an "alias" and some trust
6677 settings. The trust settings will ultimately be used in enhanced
6678 certificate chain verification routines: currently a certificate
6679 can only be trusted if it is self signed and then it is trusted
6683 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
6684 The problem was that one of the replacement routines had not been working
6685 since SSLeay releases. For now the offending routine has been replaced
6686 with non-optimised assembler. Even so, this now gives around 95%
6687 performance improvement for 1024 bit RSA signs.
6690 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
6691 handling. Most clients have the effective key size in bits equal to
6692 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
6693 A few however don't do this and instead use the size of the decrypted key
6694 to determine the RC2 key length and the AlgorithmIdentifier to determine
6695 the effective key length. In this case the effective key length can still
6696 be 40 bits but the key length can be 168 bits for example. This is fixed
6697 by manually forcing an RC2 key into the EVP_PKEY structure because the
6698 EVP code can't currently handle unusual RC2 key sizes: it always assumes
6699 the key length and effective key length are equal.
6702 *) Add a bunch of functions that should simplify the creation of
6703 X509_NAME structures. Now you should be able to do:
6704 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
6705 and have it automatically work out the correct field type and fill in
6706 the structures. The more adventurous can try:
6707 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
6708 and it will (hopefully) work out the correct multibyte encoding.
6711 *) Change the 'req' utility to use the new field handling and multibyte
6712 copy routines. Before the DN field creation was handled in an ad hoc
6713 way in req, ca, and x509 which was rather broken and didn't support
6714 BMPStrings or UTF8Strings. Since some software doesn't implement
6715 BMPStrings or UTF8Strings yet, they can be enabled using the config file
6716 using the dirstring_type option. See the new comment in the default
6717 openssl.cnf for more info.
6720 *) Make crypto/rand/md_rand.c more robust:
6721 - Assure unique random numbers after fork().
6722 - Make sure that concurrent threads access the global counter and
6723 md serializably so that we never lose entropy in them
6724 or use exactly the same state in multiple threads.
6725 Access to the large state is not always serializable because
6726 the additional locking could be a performance killer, and
6727 md should be large enough anyway.
6730 *) New file apps/app_rand.c with commonly needed functionality
6731 for handling the random seed file.
6733 Use the random seed file in some applications that previously did not:
6735 dsaparam -genkey (which also ignored its '-rand' option),
6738 x509 (when signing).
6739 Except on systems with /dev/urandom, it is crucial to have a random
6740 seed file at least for key creation, DSA signing, and for DH exchanges;
6741 for RSA signatures we could do without one.
6743 gendh and gendsa (unlike genrsa) used to read only the first byte
6744 of each file listed in the '-rand' option. The function as previously
6745 found in genrsa is now in app_rand.c and is used by all programs
6746 that support '-rand'.
6749 *) In RAND_write_file, use mode 0600 for creating files;
6750 don't just chmod when it may be too late.
6753 *) Report an error from X509_STORE_load_locations
6754 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
6757 *) New function ASN1_mbstring_copy() this copies a string in either
6758 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
6759 into an ASN1_STRING type. A mask of permissible types is passed
6760 and it chooses the "minimal" type to use or an error if not type
6764 *) Add function equivalents to the various macros in asn1.h. The old
6765 macros are retained with an M_ prefix. Code inside the library can
6766 use the M_ macros. External code (including the openssl utility)
6767 should *NOT* in order to be "shared library friendly".
6770 *) Add various functions that can check a certificate's extensions
6771 to see if it usable for various purposes such as SSL client,
6772 server or S/MIME and CAs of these types. This is currently
6773 VERY EXPERIMENTAL but will ultimately be used for certificate chain
6774 verification. Also added a -purpose flag to x509 utility to
6775 print out all the purposes.
6778 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
6782 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
6783 for, obtain and decode and extension and obtain its critical flag.
6784 This allows all the necessary extension code to be handled in a
6785 single function call.
6788 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
6789 platforms. See crypto/rc4/rc4_enc.c for further details.
6792 *) New -noout option to asn1parse. This causes no output to be produced
6793 its main use is when combined with -strparse and -out to extract data
6794 from a file (which may not be in ASN.1 format).
6797 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
6798 when producing the local key id.
6799 [Richard Levitte <levitte@stacken.kth.se>]
6801 *) New option -dhparam in s_server. This allows a DH parameter file to be
6802 stated explicitly. If it is not stated then it tries the first server
6803 certificate file. The previous behaviour hard coded the filename
6807 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
6808 a public key to be input or output. For example:
6809 openssl rsa -in key.pem -pubout -out pubkey.pem
6810 Also added necessary DSA public key functions to handle this.
6813 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
6814 in the message. This was handled by allowing
6815 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
6816 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
6818 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
6819 to the end of the strings whereas this didn't. This would cause problems
6820 if strings read with d2i_ASN1_bytes() were later modified.
6821 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
6823 *) Fix for base64 decode bug. When a base64 bio reads only one line of
6824 data and it contains EOF it will end up returning an error. This is
6825 caused by input 46 bytes long. The cause is due to the way base64
6826 BIOs find the start of base64 encoded data. They do this by trying a
6827 trial decode on each line until they find one that works. When they
6828 do a flag is set and it starts again knowing it can pass all the
6829 data directly through the decoder. Unfortunately it doesn't reset
6830 the context it uses. This means that if EOF is reached an attempt
6831 is made to pass two EOFs through the context and this causes the
6832 resulting error. This can also cause other problems as well. As is
6833 usual with these problems it takes *ages* to find and the fix is
6834 trivial: move one line.
6835 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
6837 *) Ugly workaround to get s_client and s_server working under Windows. The
6838 old code wouldn't work because it needed to select() on sockets and the
6839 tty (for keypresses and to see if data could be written). Win32 only
6840 supports select() on sockets so we select() with a 1s timeout on the
6841 sockets and then see if any characters are waiting to be read, if none
6842 are present then we retry, we also assume we can always write data to
6843 the tty. This isn't nice because the code then blocks until we've
6844 received a complete line of data and it is effectively polling the
6845 keyboard at 1s intervals: however it's quite a bit better than not
6846 working at all :-) A dedicated Windows application might handle this
6847 with an event loop for example.
6850 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
6851 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
6852 will be called when RSA_sign() and RSA_verify() are used. This is useful
6853 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
6854 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
6855 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
6856 This necessitated the support of an extra signature type NID_md5_sha1
6857 for SSL signatures and modifications to the SSL library to use it instead
6858 of calling RSA_public_decrypt() and RSA_private_encrypt().
6861 *) Add new -verify -CAfile and -CApath options to the crl program, these
6862 will lookup a CRL issuers certificate and verify the signature in a
6863 similar way to the verify program. Tidy up the crl program so it
6864 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
6865 less strict. It will now permit CRL extensions even if it is not
6866 a V2 CRL: this will allow it to tolerate some broken CRLs.
6869 *) Initialize all non-automatic variables each time one of the openssl
6870 sub-programs is started (this is necessary as they may be started
6871 multiple times from the "OpenSSL>" prompt).
6872 [Lennart Bang, Bodo Moeller]
6874 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
6875 removing all other RSA functionality (this is what NO_RSA does). This
6876 is so (for example) those in the US can disable those operations covered
6877 by the RSA patent while allowing storage and parsing of RSA keys and RSA
6881 *) Non-copying interface to BIO pairs.
6882 (still largely untested)
6885 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
6886 ASCII string. This was handled independently in various places before.
6889 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
6890 UTF8 strings a character at a time.
6893 *) Use client_version from client hello to select the protocol
6894 (s23_srvr.c) and for RSA client key exchange verification
6895 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
6898 *) Add various utility functions to handle SPKACs, these were previously
6899 handled by poking round in the structure internals. Added new function
6900 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
6901 print, verify and generate SPKACs. Based on an original idea from
6902 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
6905 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
6908 *) Allow the config file extension section to be overwritten on the
6909 command line. Based on an original idea from Massimiliano Pala
6910 <madwolf@comune.modena.it>. The new option is called -extensions
6911 and can be applied to ca, req and x509. Also -reqexts to override
6912 the request extensions in req and -crlexts to override the crl extensions
6916 *) Add new feature to the SPKAC handling in ca. Now you can include
6917 the same field multiple times by preceding it by "XXXX." for example:
6920 this is the same syntax as used in the req config file.
6923 *) Allow certificate extensions to be added to certificate requests. These
6924 are specified in a 'req_extensions' option of the req section of the
6925 config file. They can be printed out with the -text option to req but
6926 are otherwise ignored at present.
6929 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
6930 data read consists of only the final block it would not decrypted because
6931 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
6932 A misplaced 'break' also meant the decrypted final block might not be
6933 copied until the next read.
6936 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
6937 a few extra parameters to the DH structure: these will be useful if
6938 for example we want the value of 'q' or implement X9.42 DH.
6941 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
6942 provides hooks that allow the default DSA functions or functions on a
6943 "per key" basis to be replaced. This allows hardware acceleration and
6944 hardware key storage to be handled without major modification to the
6945 library. Also added low level modexp hooks and CRYPTO_EX structure and
6946 associated functions.
6949 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
6950 as "read only": it can't be written to and the buffer it points to will
6951 not be freed. Reading from a read only BIO is much more efficient than
6952 a normal memory BIO. This was added because there are several times when
6953 an area of memory needs to be read from a BIO. The previous method was
6954 to create a memory BIO and write the data to it, this results in two
6955 copies of the data and an O(n^2) reading algorithm. There is a new
6956 function BIO_new_mem_buf() which creates a read only memory BIO from
6957 an area of memory. Also modified the PKCS#7 routines to use read only
6961 *) Bugfix: ssl23_get_client_hello did not work properly when called in
6962 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
6963 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
6964 but a retry condition occured while trying to read the rest.
6967 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
6968 NID_pkcs7_encrypted by default: this was wrong since this should almost
6969 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
6970 the encrypted data type: this is a more sensible place to put it and it
6971 allows the PKCS#12 code to be tidied up that duplicated this
6975 *) Changed obj_dat.pl script so it takes its input and output files on
6976 the command line. This should avoid shell escape redirection problems
6980 *) Initial support for certificate extension requests, these are included
6981 in things like Xenroll certificate requests. Included functions to allow
6982 extensions to be obtained and added.
6985 *) -crlf option to s_client and s_server for sending newlines as
6986 CRLF (as required by many protocols).
6989 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
6991 *) Install libRSAglue.a when OpenSSL is built with RSAref.
6992 [Ralf S. Engelschall]
6994 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
6995 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
6997 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
7001 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
7002 DH parameters/keys (q is lost during that conversion, but the resulting
7003 DH parameters contain its length).
7005 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
7006 much faster than DH_generate_parameters (which creates parameters
7007 where p = 2*q + 1), and also the smaller q makes DH computations
7008 much more efficient (160-bit exponentiation instead of 1024-bit
7009 exponentiation); so this provides a convenient way to support DHE
7010 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
7011 utter importance to use
7012 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7014 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7015 when such DH parameters are used, because otherwise small subgroup
7016 attacks may become possible!
7019 *) Avoid memory leak in i2d_DHparams.
7022 *) Allow the -k option to be used more than once in the enc program:
7023 this allows the same encrypted message to be read by multiple recipients.
7026 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
7027 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
7028 it will always use the numerical form of the OID, even if it has a short
7032 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
7033 method only got called if p,q,dmp1,dmq1,iqmp components were present,
7034 otherwise bn_mod_exp was called. In the case of hardware keys for example
7035 no private key components need be present and it might store extra data
7036 in the RSA structure, which cannot be accessed from bn_mod_exp.
7037 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
7038 private key operations.
7041 *) Added support for SPARC Linux.
7044 *) pem_password_cb function type incompatibly changed from
7045 typedef int pem_password_cb(char *buf, int size, int rwflag);
7047 ....(char *buf, int size, int rwflag, void *userdata);
7048 so that applications can pass data to their callbacks:
7049 The PEM[_ASN1]_{read,write}... functions and macros now take an
7050 additional void * argument, which is just handed through whenever
7051 the password callback is called.
7052 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
7054 New function SSL_CTX_set_default_passwd_cb_userdata.
7056 Compatibility note: As many C implementations push function arguments
7057 onto the stack in reverse order, the new library version is likely to
7058 interoperate with programs that have been compiled with the old
7059 pem_password_cb definition (PEM_whatever takes some data that
7060 happens to be on the stack as its last argument, and the callback
7061 just ignores this garbage); but there is no guarantee whatsoever that
7064 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
7065 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
7066 problems not only on Windows, but also on some Unix platforms.
7067 To avoid problematic command lines, these definitions are now in an
7068 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
7069 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
7072 *) MIPS III/IV assembler module is reimplemented.
7075 *) More DES library cleanups: remove references to srand/rand and
7076 delete an unused file.
7079 *) Add support for the the free Netwide assembler (NASM) under Win32,
7080 since not many people have MASM (ml) and it can be hard to obtain.
7081 This is currently experimental but it seems to work OK and pass all
7082 the tests. Check out INSTALL.W32 for info.
7085 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
7086 without temporary keys kept an extra copy of the server key,
7087 and connections with temporary keys did not free everything in case
7091 *) New function RSA_check_key and new openssl rsa option -check
7092 for verifying the consistency of RSA keys.
7093 [Ulf Moeller, Bodo Moeller]
7095 *) Various changes to make Win32 compile work:
7096 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
7097 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
7098 comparison" warnings.
7099 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
7102 *) Add a debugging option to PKCS#5 v2 key generation function: when
7103 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
7104 derived keys are printed to stderr.
7107 *) Copy the flags in ASN1_STRING_dup().
7108 [Roman E. Pavlov <pre@mo.msk.ru>]
7110 *) The x509 application mishandled signing requests containing DSA
7111 keys when the signing key was also DSA and the parameters didn't match.
7113 It was supposed to omit the parameters when they matched the signing key:
7114 the verifying software was then supposed to automatically use the CA's
7115 parameters if they were absent from the end user certificate.
7117 Omitting parameters is no longer recommended. The test was also
7118 the wrong way round! This was probably due to unusual behaviour in
7119 EVP_cmp_parameters() which returns 1 if the parameters match.
7120 This meant that parameters were omitted when they *didn't* match and
7121 the certificate was useless. Certificates signed with 'ca' didn't have
7123 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
7125 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
7126 The interface is as follows:
7127 Applications can use
7128 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
7129 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
7130 "off" is now the default.
7131 The library internally uses
7132 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
7133 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
7134 to disable memory-checking temporarily.
7136 Some inconsistent states that previously were possible (and were
7137 even the default) are now avoided.
7139 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
7140 with each memory chunk allocated; this is occasionally more helpful
7141 than just having a counter.
7143 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
7145 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
7149 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
7150 which largely parallels "options", but is for changing API behaviour,
7151 whereas "options" are about protocol behaviour.
7152 Initial "mode" flags are:
7154 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
7155 a single record has been written.
7156 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
7157 retries use the same buffer location.
7158 (But all of the contents must be
7162 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
7165 *) Fix problems with no-hmac etc.
7166 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
7168 *) New functions RSA_get_default_method(), RSA_set_method() and
7169 RSA_get_method(). These allows replacement of RSA_METHODs without having
7170 to mess around with the internals of an RSA structure.
7173 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
7174 Also really enable memory leak checks in openssl.c and in some
7176 [Chad C. Mulligan, Bodo Moeller]
7178 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
7179 up the length of negative integers. This has now been simplified to just
7180 store the length when it is first determined and use it later, rather
7181 than trying to keep track of where data is copied and updating it to
7183 [Steve Henson, reported by Brien Wheeler
7184 <bwheeler@authentica-security.com>]
7186 *) Add a new function PKCS7_signatureVerify. This allows the verification
7187 of a PKCS#7 signature but with the signing certificate passed to the
7188 function itself. This contrasts with PKCS7_dataVerify which assumes the
7189 certificate is present in the PKCS#7 structure. This isn't always the
7190 case: certificates can be omitted from a PKCS#7 structure and be
7191 distributed by "out of band" means (such as a certificate database).
7194 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
7195 function prototypes in pem.h, also change util/mkdef.pl to add the
7196 necessary function names.
7199 *) mk1mf.pl (used by Windows builds) did not properly read the
7200 options set by Configure in the top level Makefile, and Configure
7201 was not even able to write more than one option correctly.
7202 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
7205 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
7206 file to be loaded from a BIO or FILE pointer. The BIO version will
7207 for example allow memory BIOs to contain config info.
7210 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
7211 Whoever hopes to achieve shared-library compatibility across versions
7212 must use this, not the compile-time macro.
7213 (Exercise 0.9.4: Which is the minimum library version required by
7215 Note: All this applies only to multi-threaded programs, others don't
7219 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
7220 through a BIO pair triggered the default case, i.e.
7221 SSLerr(...,SSL_R_UNKNOWN_STATE).
7224 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
7225 can use the SSL library even if none of the specific BIOs is
7229 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
7230 for the encoded length.
7231 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
7233 *) Add initial documentation of the X509V3 functions.
7236 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
7237 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
7238 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
7239 secure PKCS#8 private key format with a high iteration count.
7242 *) Fix determination of Perl interpreter: A perl or perl5
7243 _directory_ in $PATH was also accepted as the interpreter.
7244 [Ralf S. Engelschall]
7246 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
7247 wrong with it but it was very old and did things like calling
7248 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
7252 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
7253 to use the new extension code.
7256 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
7257 with macros. This should make it easier to change their form, add extra
7258 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
7262 *) Add to configuration table a new entry that can specify an alternative
7263 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
7264 according to Mark Crispin <MRC@Panda.COM>.
7268 *) DES CBC did not update the IV. Weird.
7271 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
7272 Changing the behaviour of the former might break existing programs --
7273 where IV updating is needed, des_ncbc_encrypt can be used.
7276 *) When bntest is run from "make test" it drives bc to check its
7277 calculations, as well as internally checking them. If an internal check
7278 fails, it needs to cause bc to give a non-zero result or make test carries
7279 on without noticing the failure. Fixed.
7282 *) DES library cleanups.
7285 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
7286 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
7287 ciphers. NOTE: although the key derivation function has been verified
7288 against some published test vectors it has not been extensively tested
7289 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
7293 *) Instead of "mkdir -p", which is not fully portable, use new
7294 Perl script "util/mkdir-p.pl".
7297 *) Rewrite the way password based encryption (PBE) is handled. It used to
7298 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
7299 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
7300 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
7301 the 'parameter' field of the AlgorithmIdentifier is passed to the
7302 underlying key generation function so it must do its own ASN1 parsing.
7303 This has also changed the EVP_PBE_CipherInit() function which now has a
7304 'parameter' argument instead of literal salt and iteration count values
7305 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
7308 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
7309 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
7310 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
7311 KEY" because this clashed with PKCS#8 unencrypted string. Since this
7312 value was just used as a "magic string" and not used directly its
7313 value doesn't matter.
7316 *) Introduce some semblance of const correctness to BN. Shame C doesn't
7320 *) "linux-sparc64" configuration (ultrapenguin).
7321 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
7322 "linux-sparc" configuration.
7323 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
7325 *) config now generates no-xxx options for missing ciphers.
7328 *) Support the EBCDIC character set (work in progress).
7329 File ebcdic.c not yet included because it has a different license.
7330 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7332 *) Support BS2000/OSD-POSIX.
7333 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7335 *) Make callbacks for key generation use void * instead of char *.
7338 *) Make S/MIME samples compile (not yet tested).
7341 *) Additional typesafe stacks.
7344 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
7348 Changes between 0.9.3 and 0.9.3a [29 May 1999]
7350 *) New configuration variant "sco5-gcc".
7352 *) Updated some demos.
7353 [Sean O Riordain, Wade Scholine]
7355 *) Add missing BIO_free at exit of pkcs12 application.
7358 *) Fix memory leak in conf.c.
7361 *) Updates for Win32 to assembler version of MD5.
7364 *) Set #! path to perl in apps/der_chop to where we found it
7365 instead of using a fixed path.
7368 *) SHA library changes for irix64-mips4-cc.
7371 *) Improvements for VMS support.
7375 Changes between 0.9.2b and 0.9.3 [24 May 1999]
7377 *) Bignum library bug fix. IRIX 6 passes "make test" now!
7378 This also avoids the problems with SC4.2 and unpatched SC5.
7379 [Andy Polyakov <appro@fy.chalmers.se>]
7381 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
7382 These are required because of the typesafe stack would otherwise break
7383 existing code. If old code used a structure member which used to be STACK
7384 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
7385 sk_num or sk_value it would produce an error because the num, data members
7386 are not present in STACK_OF. Now it just produces a warning. sk_set
7387 replaces the old method of assigning a value to sk_value
7388 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
7389 that does this will no longer work (and should use sk_set instead) but
7390 this could be regarded as a "questionable" behaviour anyway.
7393 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
7394 correctly handle encrypted S/MIME data.
7397 *) Change type of various DES function arguments from des_cblock
7398 (which means, in function argument declarations, pointer to char)
7399 to des_cblock * (meaning pointer to array with 8 char elements),
7400 which allows the compiler to do more typechecking; it was like
7401 that back in SSLeay, but with lots of ugly casts.
7403 Introduce new type const_des_cblock.
7406 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
7407 problems: find RecipientInfo structure that matches recipient certificate
7408 and initialise the ASN1 structures properly based on passed cipher.
7411 *) Belatedly make the BN tests actually check the results.
7414 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
7415 to and from BNs: it was completely broken. New compilation option
7416 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
7417 key elements as negative integers.
7420 *) Reorganize and speed up MD5.
7421 [Andy Polyakov <appro@fy.chalmers.se>]
7424 [Richard Levitte <richard@levitte.org>]
7426 *) New option -out to asn1parse to allow the parsed structure to be
7427 output to a file. This is most useful when combined with the -strparse
7428 option to examine the output of things like OCTET STRINGS.
7431 *) Make SSL library a little more fool-proof by not requiring any longer
7432 that SSL_set_{accept,connect}_state be called before
7433 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
7434 in many applications because usually everything *appeared* to work as
7435 intended anyway -- now it really works as intended).
7438 *) Move openssl.cnf out of lib/.
7441 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
7442 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
7443 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
7444 [Ralf S. Engelschall]
7446 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
7447 handle PKCS#7 enveloped data properly.
7448 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
7450 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
7451 copying pointers. The cert_st handling is changed by this in
7452 various ways (and thus what used to be known as ctx->default_cert
7453 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
7454 any longer when s->cert does not give us what we need).
7455 ssl_cert_instantiate becomes obsolete by this change.
7456 As soon as we've got the new code right (possibly it already is?),
7457 we have solved a couple of bugs of the earlier code where s->cert
7458 was used as if it could not have been shared with other SSL structures.
7460 Note that using the SSL API in certain dirty ways now will result
7461 in different behaviour than observed with earlier library versions:
7462 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
7463 does not influence s as it used to.
7465 In order to clean up things more thoroughly, inside SSL_SESSION
7466 we don't use CERT any longer, but a new structure SESS_CERT
7467 that holds per-session data (if available); currently, this is
7468 the peer's certificate chain and, for clients, the server's certificate
7469 and temporary key. CERT holds only those values that can have
7470 meaningful defaults in an SSL_CTX.
7473 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
7474 from the internal representation. Various PKCS#7 fixes: remove some
7475 evil casts and set the enc_dig_alg field properly based on the signing
7479 *) Allow PKCS#12 password to be set from the command line or the
7480 environment. Let 'ca' get its config file name from the environment
7481 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
7485 *) Allow certificate policies extension to use an IA5STRING for the
7486 organization field. This is contrary to the PKIX definition but
7487 VeriSign uses it and IE5 only recognises this form. Document 'x509'
7491 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
7492 without disallowing inline assembler and the like for non-pedantic builds.
7495 *) Support Borland C++ builder.
7496 [Janez Jere <jj@void.si>, modified by Ulf Möller]
7501 *) SHA-1 cleanups and performance enhancements.
7502 [Andy Polyakov <appro@fy.chalmers.se>]
7504 *) Sparc v8plus assembler for the bignum library.
7505 [Andy Polyakov <appro@fy.chalmers.se>]
7507 *) Accept any -xxx and +xxx compiler options in Configure.
7510 *) Update HPUX configuration.
7513 *) Add missing sk_<type>_unshift() function to safestack.h
7514 [Ralf S. Engelschall]
7516 *) New function SSL_CTX_use_certificate_chain_file that sets the
7517 "extra_cert"s in addition to the certificate. (This makes sense
7518 only for "PEM" format files, as chains as a whole are not
7522 *) Support verify_depth from the SSL API.
7523 x509_vfy.c had what can be considered an off-by-one-error:
7524 Its depth (which was not part of the external interface)
7525 was actually counting the number of certificates in a chain;
7526 now it really counts the depth.
7529 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
7530 instead of X509err, which often resulted in confusing error
7531 messages since the error codes are not globally unique
7532 (e.g. an alleged error in ssl3_accept when a certificate
7533 didn't match the private key).
7535 *) New function SSL_CTX_set_session_id_context that allows to set a default
7536 value (so that you don't need SSL_set_session_id_context for each
7537 connection using the SSL_CTX).
7540 *) OAEP decoding bug fix.
7543 *) Support INSTALL_PREFIX for package builders, as proposed by
7547 *) New Configure options "threads" and "no-threads". For systems
7548 where the proper compiler options are known (currently Solaris
7549 and Linux), "threads" is the default.
7552 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
7555 *) Install various scripts to $(OPENSSLDIR)/misc, not to
7556 $(INSTALLTOP)/bin -- they shouldn't clutter directories
7557 such as /usr/local/bin.
7560 *) "make linux-shared" to build shared libraries.
7561 [Niels Poppe <niels@netbox.org>]
7563 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
7566 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
7567 extension adding in x509 utility.
7570 *) Remove NOPROTO sections and error code comments.
7573 *) Partial rewrite of the DEF file generator to now parse the ANSI
7577 *) New Configure options --prefix=DIR and --openssldir=DIR.
7580 *) Complete rewrite of the error code script(s). It is all now handled
7581 by one script at the top level which handles error code gathering,
7582 header rewriting and C source file generation. It should be much better
7583 than the old method: it now uses a modified version of Ulf's parser to
7584 read the ANSI prototypes in all header files (thus the old K&R definitions
7585 aren't needed for error creation any more) and do a better job of
7586 translating function codes into names. The old 'ASN1 error code imbedded
7587 in a comment' is no longer necessary and it doesn't use .err files which
7588 have now been deleted. Also the error code call doesn't have to appear all
7589 on one line (which resulted in some large lines...).
7592 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
7595 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
7596 0 (which usually indicates a closed connection), but continue reading.
7599 *) Fix some race conditions.
7602 *) Add support for CRL distribution points extension. Add Certificate
7603 Policies and CRL distribution points documentation.
7606 *) Move the autogenerated header file parts to crypto/opensslconf.h.
7609 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
7610 8 of keying material. Merlin has also confirmed interop with this fix
7611 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
7612 [Merlin Hughes <merlin@baltimore.ie>]
7614 *) Fix lots of warnings.
7615 [Richard Levitte <levitte@stacken.kth.se>]
7617 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
7618 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
7619 [Richard Levitte <levitte@stacken.kth.se>]
7621 *) Fix problems with sizeof(long) == 8.
7622 [Andy Polyakov <appro@fy.chalmers.se>]
7624 *) Change functions to ANSI C.
7627 *) Fix typos in error codes.
7628 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
7630 *) Remove defunct assembler files from Configure.
7633 *) SPARC v8 assembler BIGNUM implementation.
7634 [Andy Polyakov <appro@fy.chalmers.se>]
7636 *) Support for Certificate Policies extension: both print and set.
7637 Various additions to support the r2i method this uses.
7640 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
7641 return a const string when you are expecting an allocated buffer.
7644 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
7645 types DirectoryString and DisplayText.
7648 *) Add code to allow r2i extensions to access the configuration database,
7649 add an LHASH database driver and add several ctx helper functions.
7652 *) Fix an evil bug in bn_expand2() which caused various BN functions to
7653 fail when they extended the size of a BIGNUM.
7656 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
7657 support typesafe stack.
7660 *) Fix typo in SSL_[gs]et_options().
7661 [Nils Frostberg <nils@medcom.se>]
7663 *) Delete various functions and files that belonged to the (now obsolete)
7664 old X509V3 handling code.
7667 *) New Configure option "rsaref".
7670 *) Don't auto-generate pem.h.
7673 *) Introduce type-safe ASN.1 SETs.
7676 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
7677 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
7679 *) Introduce type-safe STACKs. This will almost certainly break lots of code
7680 that links with OpenSSL (well at least cause lots of warnings), but fear
7681 not: the conversion is trivial, and it eliminates loads of evil casts. A
7682 few STACKed things have been converted already. Feel free to convert more.
7683 In the fullness of time, I'll do away with the STACK type altogether.
7686 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
7687 specified in <certfile> by updating the entry in the index.txt file.
7688 This way one no longer has to edit the index.txt file manually for
7689 revoking a certificate. The -revoke option does the gory details now.
7690 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
7692 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
7693 `-text' option at all and this way the `-noout -text' combination was
7694 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
7695 [Ralf S. Engelschall]
7697 *) Make sure a corresponding plain text error message exists for the
7698 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
7699 verify callback function determined that a certificate was revoked.
7700 [Ralf S. Engelschall]
7702 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
7703 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
7704 all available cipers including rc5, which was forgotten until now.
7705 In order to let the testing shell script know which algorithms
7706 are available, a new (up to now undocumented) command
7707 "openssl list-cipher-commands" is used.
7710 *) Bugfix: s_client occasionally would sleep in select() when
7711 it should have checked SSL_pending() first.
7714 *) New functions DSA_do_sign and DSA_do_verify to provide access to
7715 the raw DSA values prior to ASN.1 encoding.
7718 *) Tweaks to Configure
7719 [Niels Poppe <niels@netbox.org>]
7721 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
7725 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
7728 *) New config option to avoid instructions that are illegal on the 80386.
7729 The default code is faster, but requires at least a 486.
7732 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
7733 SSL2_SERVER_VERSION (not used at all) macros, which are now the
7734 same as SSL2_VERSION anyway.
7737 *) New "-showcerts" option for s_client.
7740 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
7741 application. Various cleanups and fixes.
7744 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
7745 modify error routines to work internally. Add error codes and PBE init
7746 to library startup routines.
7749 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
7750 packing functions to asn1 and evp. Changed function names and error
7751 codes along the way.
7754 *) PKCS12 integration: and so it begins... First of several patches to
7755 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
7756 objects to objects.h
7759 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
7760 and display support for Thawte strong extranet extension.
7763 *) Add LinuxPPC support.
7764 [Jeff Dubrule <igor@pobox.org>]
7766 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
7767 bn_div_words in alpha.s.
7768 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
7770 *) Make sure the RSA OAEP test is skipped under -DRSAref because
7771 OAEP isn't supported when OpenSSL is built with RSAref.
7772 [Ulf Moeller <ulf@fitug.de>]
7774 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
7775 so they no longer are missing under -DNOPROTO.
7776 [Soren S. Jorvang <soren@t.dk>]
7779 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
7781 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
7782 doesn't work when the session is reused. Coming soon!
7785 *) Fix a security hole, that allows sessions to be reused in the wrong
7786 context thus bypassing client cert protection! All software that uses
7787 client certs and session caches in multiple contexts NEEDS PATCHING to
7788 allow session reuse! A fuller solution is in the works.
7789 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
7791 *) Some more source tree cleanups (removed obsolete files
7792 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
7793 permission on "config" script to be executable) and a fix for the INSTALL
7795 [Ulf Moeller <ulf@fitug.de>]
7797 *) Remove some legacy and erroneous uses of malloc, free instead of
7799 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
7801 *) Make rsa_oaep_test return non-zero on error.
7802 [Ulf Moeller <ulf@fitug.de>]
7804 *) Add support for native Solaris shared libraries. Configure
7805 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
7806 if someone would make that last step automatic.
7807 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
7809 *) ctx_size was not built with the right compiler during "make links". Fixed.
7812 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
7813 except NULL ciphers". This means the default cipher list will no longer
7814 enable NULL ciphers. They need to be specifically enabled e.g. with
7815 the string "DEFAULT:eNULL".
7818 *) Fix to RSA private encryption routines: if p < q then it would
7819 occasionally produce an invalid result. This will only happen with
7820 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
7823 *) Be less restrictive and allow also `perl util/perlpath.pl
7824 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
7825 because this way one can also use an interpreter named `perl5' (which is
7826 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
7827 installed as `perl').
7828 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7830 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
7831 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7833 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
7834 advapi32.lib to Win32 build and change the pem test comparision
7835 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
7836 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
7837 and crypto/des/ede_cbcm_enc.c.
7840 *) DES quad checksum was broken on big-endian architectures. Fixed.
7843 *) Comment out two functions in bio.h that aren't implemented. Fix up the
7844 Win32 test batch file so it (might) work again. The Win32 test batch file
7845 is horrible: I feel ill....
7848 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
7849 in e_os.h. Audit of header files to check ANSI and non ANSI
7850 sections: 10 functions were absent from non ANSI section and not exported
7851 from Windows DLLs. Fixed up libeay.num for new functions.
7854 *) Make `openssl version' output lines consistent.
7855 [Ralf S. Engelschall]
7857 *) Fix Win32 symbol export lists for BIO functions: Added
7858 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
7859 to ms/libeay{16,32}.def.
7860 [Ralf S. Engelschall]
7862 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
7863 fine under Unix and passes some trivial tests I've now added. But the
7864 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
7865 added to make sure no one expects that this stuff really works in the
7866 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
7867 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
7869 [Ralf S. Engelschall]
7871 *) Fix the generation of two part addresses in perl.
7872 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
7874 *) Add config entry for Linux on MIPS.
7875 [John Tobey <jtobey@channel1.com>]
7877 *) Make links whenever Configure is run, unless we are on Windoze.
7880 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
7881 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
7885 *) Add a useful kludge to allow package maintainers to specify compiler and
7886 other platforms details on the command line without having to patch the
7887 Configure script everytime: One now can use ``perl Configure
7888 <id>:<details>'', i.e. platform ids are allowed to have details appended
7889 to them (seperated by colons). This is treated as there would be a static
7890 pre-configured entry in Configure's %table under key <id> with value
7891 <details> and ``perl Configure <id>'' is called. So, when you want to
7892 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
7893 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
7894 now, which overrides the FreeBSD-elf entry on-the-fly.
7895 [Ralf S. Engelschall]
7897 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
7900 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
7901 on the `perl Configure ...' command line. This way one can compile
7902 OpenSSL libraries with Position Independent Code (PIC) which is needed
7903 for linking it into DSOs.
7904 [Ralf S. Engelschall]
7906 *) Remarkably, export ciphers were totally broken and no-one had noticed!
7910 *) Cleaned up the LICENSE document: The official contact for any license
7911 questions now is the OpenSSL core team under openssl-core@openssl.org.
7912 And add a paragraph about the dual-license situation to make sure people
7913 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
7914 to the OpenSSL toolkit.
7915 [Ralf S. Engelschall]
7917 *) General source tree makefile cleanups: Made `making xxx in yyy...'
7918 display consistent in the source tree and replaced `/bin/rm' by `rm'.
7919 Additonally cleaned up the `make links' target: Remove unnecessary
7920 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
7921 to speed processing and no longer clutter the display with confusing
7922 stuff. Instead only the actually done links are displayed.
7923 [Ralf S. Engelschall]
7925 *) Permit null encryption ciphersuites, used for authentication only. It used
7926 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
7927 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
7931 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
7932 signed attributes when verifying signatures (this would break them),
7933 the detached data encoding was wrong and public keys obtained using
7934 X509_get_pubkey() weren't freed.
7937 *) Add text documentation for the BUFFER functions. Also added a work around
7938 to a Win95 console bug. This was triggered by the password read stuff: the
7939 last character typed gets carried over to the next fread(). If you were
7940 generating a new cert request using 'req' for example then the last
7941 character of the passphrase would be CR which would then enter the first
7945 *) Added the new `Includes OpenSSL Cryptography Software' button as
7946 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
7947 button and can be used by applications based on OpenSSL to show the
7948 relationship to the OpenSSL project.
7949 [Ralf S. Engelschall]
7951 *) Remove confusing variables in function signatures in files
7952 ssl/ssl_lib.c and ssl/ssl.h.
7953 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7955 *) Don't install bss_file.c under PREFIX/include/
7956 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7958 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
7959 functions that return function pointers and has support for NT specific
7960 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
7961 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
7962 unsigned to signed types: this was killing the Win32 compile.
7965 *) Add new certificate file to stack functions,
7966 SSL_add_dir_cert_subjects_to_stack() and
7967 SSL_add_file_cert_subjects_to_stack(). These largely supplant
7968 SSL_load_client_CA_file(), and can be used to add multiple certs easily
7969 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
7970 This means that Apache-SSL and similar packages don't have to mess around
7971 to add as many CAs as they want to the preferred list.
7974 *) Experiment with doxygen documentation. Currently only partially applied to
7976 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
7977 openssl.doxy as the configuration file.
7980 *) Get rid of remaining C++-style comments which strict C compilers hate.
7981 [Ralf S. Engelschall, pointed out by Carlos Amengual]
7983 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
7984 compiled in by default: it has problems with large keys.
7987 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
7988 DH private keys and/or callback functions which directly correspond to
7989 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
7990 is needed for applications which have to configure certificates on a
7991 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
7993 For the RSA certificate situation is makes no difference, but
7994 for the DSA certificate situation this fixes the "no shared cipher"
7995 problem where the OpenSSL cipher selection procedure failed because the
7996 temporary keys were not overtaken from the context and the API provided
7997 no way to reconfigure them.
7998 The new functions now let applications reconfigure the stuff and they
7999 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
8000 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
8001 non-public-API function ssl_cert_instantiate() is used as a helper
8002 function and also to reduce code redundancy inside ssl_rsa.c.
8003 [Ralf S. Engelschall]
8005 *) Move s_server -dcert and -dkey options out of the undocumented feature
8006 area because they are useful for the DSA situation and should be
8007 recognized by the users.
8008 [Ralf S. Engelschall]
8010 *) Fix the cipher decision scheme for export ciphers: the export bits are
8011 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
8012 SSL_EXP_MASK. So, the original variable has to be used instead of the
8013 already masked variable.
8014 [Richard Levitte <levitte@stacken.kth.se>]
8016 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
8017 [Richard Levitte <levitte@stacken.kth.se>]
8019 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
8020 from `int' to `unsigned int' because it's a length and initialized by
8021 EVP_DigestFinal() which expects an `unsigned int *'.
8022 [Richard Levitte <levitte@stacken.kth.se>]
8024 *) Don't hard-code path to Perl interpreter on shebang line of Configure
8025 script. Instead use the usual Shell->Perl transition trick.
8026 [Ralf S. Engelschall]
8028 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
8029 (in addition to RSA certificates) to match the behaviour of `openssl dsa
8030 -noout -modulus' as it's already the case for `openssl rsa -noout
8031 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
8032 currently the public key is printed (a decision which was already done by
8033 `openssl dsa -modulus' in the past) which serves a similar purpose.
8034 Additionally the NO_RSA no longer completely removes the whole -modulus
8035 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
8037 [Ralf S. Engelschall]
8039 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
8040 BIO. See the source (crypto/evp/bio_ok.c) for more info.
8041 [Arne Ansper <arne@ats.cyber.ee>]
8043 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
8044 to be added. Now both 'req' and 'ca' can use new objects defined in the
8048 *) Add cool BIO that does syslog (or event log on NT).
8049 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
8051 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
8052 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
8053 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
8054 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
8057 *) Add preliminary config info for new extension code.
8060 *) Make RSA_NO_PADDING really use no padding.
8061 [Ulf Moeller <ulf@fitug.de>]
8063 *) Generate errors when private/public key check is done.
8066 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
8067 for some CRL extensions and new objects added.
8070 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
8071 key usage extension and fuller support for authority key id.
8074 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
8075 padding method for RSA, which is recommended for new applications in PKCS
8076 #1 v2.0 (RFC 2437, October 1998).
8077 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
8078 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
8079 against Bleichbacher's attack on RSA.
8080 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
8083 *) Updates to the new SSL compression code
8084 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8086 *) Fix so that the version number in the master secret, when passed
8087 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
8088 (because the server will not accept higher), that the version number
8089 is 0x03,0x01, not 0x03,0x00
8090 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8092 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
8093 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
8094 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
8097 *) Support for RAW extensions where an arbitrary extension can be
8098 created by including its DER encoding. See apps/openssl.cnf for
8102 *) Make sure latest Perl versions don't interpret some generated C array
8103 code as Perl array code in the crypto/err/err_genc.pl script.
8104 [Lars Weber <3weber@informatik.uni-hamburg.de>]
8106 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
8107 not many people have the assembler. Various Win32 compilation fixes and
8108 update to the INSTALL.W32 file with (hopefully) more accurate Win32
8112 *) Modify configure script 'Configure' to automatically create crypto/date.h
8113 file under Win32 and also build pem.h from pem.org. New script
8114 util/mkfiles.pl to create the MINFO file on environments that can't do a
8115 'make files': perl util/mkfiles.pl >MINFO should work.
8118 *) Major rework of DES function declarations, in the pursuit of correctness
8119 and purity. As a result, many evil casts evaporated, and some weirdness,
8120 too. You may find this causes warnings in your code. Zapping your evil
8121 casts will probably fix them. Mostly.
8124 *) Fix for a typo in asn1.h. Bug fix to object creation script
8125 obj_dat.pl. It considered a zero in an object definition to mean
8126 "end of object": none of the objects in objects.h have any zeros
8127 so it wasn't spotted.
8128 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
8130 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
8131 Masking (CBCM). In the absence of test vectors, the best I have been able
8132 to do is check that the decrypt undoes the encrypt, so far. Send me test
8133 vectors if you have them.
8136 *) Correct calculation of key length for export ciphers (too much space was
8137 allocated for null ciphers). This has not been tested!
8140 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
8141 message is now correct (it understands "crypto" and "ssl" on its
8142 command line). There is also now an "update" option. This will update
8143 the util/ssleay.num and util/libeay.num files with any new functions.
8145 perl util/mkdef.pl crypto ssl update
8146 it will update them.
8149 *) Overhauled the Perl interface (perl/*):
8150 - ported BN stuff to OpenSSL's different BN library
8151 - made the perl/ source tree CVS-aware
8152 - renamed the package from SSLeay to OpenSSL (the files still contain
8153 their history because I've copied them in the repository)
8154 - removed obsolete files (the test scripts will be replaced
8155 by better Test::Harness variants in the future)
8156 [Ralf S. Engelschall]
8158 *) First cut for a very conservative source tree cleanup:
8159 1. merge various obsolete readme texts into doc/ssleay.txt
8160 where we collect the old documents and readme texts.
8161 2. remove the first part of files where I'm already sure that we no
8162 longer need them because of three reasons: either they are just temporary
8163 files which were left by Eric or they are preserved original files where
8164 I've verified that the diff is also available in the CVS via "cvs diff
8165 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
8166 the crypto/md/ stuff).
8167 [Ralf S. Engelschall]
8169 *) More extension code. Incomplete support for subject and issuer alt
8170 name, issuer and authority key id. Change the i2v function parameters
8171 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
8172 what that's for :-) Fix to ASN1 macro which messed up
8173 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
8176 *) Preliminary support for ENUMERATED type. This is largely copied from the
8180 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
8181 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8183 *) Make sure `make rehash' target really finds the `openssl' program.
8184 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8186 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
8187 like to hear about it if this slows down other processors.
8190 *) Add CygWin32 platform information to Configure script.
8191 [Alan Batie <batie@aahz.jf.intel.com>]
8193 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
8194 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
8196 *) New program nseq to manipulate netscape certificate sequences
8199 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
8203 *) Fixes to BN code. Previously the default was to define BN_RECURSION
8204 but the BN code had some problems that would cause failures when
8205 doing certificate verification and some other functions.
8206 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8208 *) Add ASN1 and PEM code to support netscape certificate sequences.
8211 *) Add ASN1 and PEM code to support netscape certificate sequences.
8214 *) Add several PKIX and private extended key usage OIDs.
8217 *) Modify the 'ca' program to handle the new extension code. Modify
8218 openssl.cnf for new extension format, add comments.
8221 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
8222 and add a sample to openssl.cnf so req -x509 now adds appropriate
8226 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
8227 error code, add initial support to X509_print() and x509 application.
8230 *) Takes a deep breath and start addding X509 V3 extension support code. Add
8231 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
8232 stuff is currently isolated and isn't even compiled yet.
8235 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
8236 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
8237 Removed the versions check from X509 routines when loading extensions:
8238 this allows certain broken certificates that don't set the version
8239 properly to be processed.
8242 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
8243 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
8244 can still be regenerated with "make depend".
8247 *) Spelling mistake in C version of CAST-128.
8248 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
8250 *) Changes to the error generation code. The perl script err-code.pl
8251 now reads in the old error codes and retains the old numbers, only
8252 adding new ones if necessary. It also only changes the .err files if new
8253 codes are added. The makefiles have been modified to only insert errors
8254 when needed (to avoid needlessly modifying header files). This is done
8255 by only inserting errors if the .err file is newer than the auto generated
8256 C file. To rebuild all the error codes from scratch (the old behaviour)
8257 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
8258 or delete all the .err files.
8261 *) CAST-128 was incorrectly implemented for short keys. The C version has
8262 been fixed, but is untested. The assembler versions are also fixed, but
8263 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
8264 to regenerate it if needed.
8265 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
8266 Hagino <itojun@kame.net>]
8268 *) File was opened incorrectly in randfile.c.
8269 [Ulf Möller <ulf@fitug.de>]
8271 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
8272 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
8273 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
8274 al: it's just almost always a UTCTime. Note this patch adds new error
8275 codes so do a "make errors" if there are problems.
8278 *) Correct Linux 1 recognition in config.
8279 [Ulf Möller <ulf@fitug.de>]
8281 *) Remove pointless MD5 hash when using DSA keys in ca.
8282 [Anonymous <nobody@replay.com>]
8284 *) Generate an error if given an empty string as a cert directory. Also
8285 generate an error if handed NULL (previously returned 0 to indicate an
8286 error, but didn't set one).
8287 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
8289 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
8292 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
8293 parameters. This was causing a warning which killed off the Win32 compile.
8296 *) Remove C++ style comments from crypto/bn/bn_local.h.
8297 [Neil Costigan <neil.costigan@celocom.com>]
8299 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
8300 based on a text string, looking up short and long names and finally
8301 "dot" format. The "dot" format stuff didn't work. Added new function
8302 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
8303 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
8304 OID is not part of the table.
8307 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
8308 X509_LOOKUP_by_alias().
8311 *) Sort openssl functions by name.
8314 *) Get the gendsa program working (hopefully) and add it to app list. Remove
8315 encryption from sample DSA keys (in case anyone is interested the password
8319 *) Make _all_ *_free functions accept a NULL pointer.
8320 [Frans Heymans <fheymans@isaserver.be>]
8322 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
8324 [Anonymous <nobody@replay.com>]
8326 *) s_server should send the CAfile as acceptable CAs, not its own cert.
8327 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8329 *) Don't blow it for numeric -newkey arguments to apps/req.
8330 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8332 *) Temp key "for export" tests were wrong in s3_srvr.c.
8333 [Anonymous <nobody@replay.com>]
8335 *) Add prototype for temp key callback functions
8336 SSL_CTX_set_tmp_{rsa,dh}_callback().
8339 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
8340 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
8343 *) X509_name_add_entry() freed the wrong thing after an error.
8344 [Arne Ansper <arne@ats.cyber.ee>]
8346 *) rsa_eay.c would attempt to free a NULL context.
8347 [Arne Ansper <arne@ats.cyber.ee>]
8349 *) BIO_s_socket() had a broken should_retry() on Windoze.
8350 [Arne Ansper <arne@ats.cyber.ee>]
8352 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
8353 [Arne Ansper <arne@ats.cyber.ee>]
8355 *) Make sure the already existing X509_STORE->depth variable is initialized
8356 in X509_STORE_new(), but document the fact that this variable is still
8357 unused in the certificate verification process.
8358 [Ralf S. Engelschall]
8360 *) Fix the various library and apps files to free up pkeys obtained from
8361 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
8364 *) Fix reference counting in X509_PUBKEY_get(). This makes
8365 demos/maurice/example2.c work, amongst others, probably.
8366 [Steve Henson and Ben Laurie]
8368 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
8369 `openssl' and second, the shortcut symlinks for the `openssl <command>'
8370 are no longer created. This way we have a single and consistent command
8371 line interface `openssl <command>', similar to `cvs <command>'.
8372 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
8374 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
8375 BIT STRING wrapper always have zero unused bits.
8378 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
8381 *) Make the top-level INSTALL documentation easier to understand.
8384 *) Makefiles updated to exit if an error occurs in a sub-directory
8385 make (including if user presses ^C) [Paul Sutton]
8387 *) Make Montgomery context stuff explicit in RSA data structure.
8390 *) Fix build order of pem and err to allow for generated pem.h.
8393 *) Fix renumbering bug in X509_NAME_delete_entry().
8396 *) Enhanced the err-ins.pl script so it makes the error library number
8397 global and can add a library name. This is needed for external ASN1 and
8398 other error libraries.
8401 *) Fixed sk_insert which never worked properly.
8404 *) Fix ASN1 macros so they can handle indefinite length construted
8405 EXPLICIT tags. Some non standard certificates use these: they can now
8409 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
8410 into a single doc/ssleay.txt bundle. This way the information is still
8411 preserved but no longer messes up this directory. Now it's new room for
8412 the new set of documenation files.
8413 [Ralf S. Engelschall]
8415 *) SETs were incorrectly DER encoded. This was a major pain, because they
8416 shared code with SEQUENCEs, which aren't coded the same. This means that
8417 almost everything to do with SETs or SEQUENCEs has either changed name or
8418 number of arguments.
8419 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
8421 *) Fix test data to work with the above.
8424 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
8425 was already fixed by Eric for 0.9.1 it seems.
8426 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
8428 *) Autodetect FreeBSD3.
8431 *) Fix various bugs in Configure. This affects the following platforms:
8435 unixware-2.0-pentium
8439 *) Eliminate generated files from CVS. Reorder tests to regenerate files
8440 before they are needed.
8443 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
8447 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
8449 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
8450 changed SSLeay to OpenSSL in version strings.
8451 [Ralf S. Engelschall]
8453 *) Some fixups to the top-level documents.
8456 *) Fixed the nasty bug where rsaref.h was not found under compile-time
8457 because the symlink to include/ was missing.
8458 [Ralf S. Engelschall]
8460 *) Incorporated the popular no-RSA/DSA-only patches
8461 which allow to compile a RSA-free SSLeay.
8462 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
8464 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
8465 when "ssleay" is still not found.
8466 [Ralf S. Engelschall]
8468 *) Added more platforms to Configure: Cray T3E, HPUX 11,
8469 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
8471 *) Updated the README file.
8472 [Ralf S. Engelschall]
8474 *) Added various .cvsignore files in the CVS repository subdirs
8475 to make a "cvs update" really silent.
8476 [Ralf S. Engelschall]
8478 *) Recompiled the error-definition header files and added
8479 missing symbols to the Win32 linker tables.
8480 [Ralf S. Engelschall]
8482 *) Cleaned up the top-level documents;
8483 o new files: CHANGES and LICENSE
8484 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
8485 o merged COPYRIGHT into LICENSE
8486 o removed obsolete TODO file
8487 o renamed MICROSOFT to INSTALL.W32
8488 [Ralf S. Engelschall]
8490 *) Removed dummy files from the 0.9.1b source tree:
8491 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
8492 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
8493 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
8494 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
8495 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
8496 [Ralf S. Engelschall]
8498 *) Added various platform portability fixes.
8501 *) The Genesis of the OpenSSL rpject:
8502 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
8503 Young and Tim J. Hudson created while they were working for C2Net until
8505 [The OpenSSL Project]
8508 Changes between 0.9.0b and 0.9.1b [not released]
8510 *) Updated a few CA certificates under certs/
8513 *) Changed some BIGNUM api stuff.
8516 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
8517 DGUX x86, Linux Alpha, etc.
8520 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
8521 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
8525 *) Add -strparse option to asn1pars program which parses nested
8527 [Dr Stephen Henson <shenson@bigfoot.com>]
8529 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
8532 *) DSA fix for "ca" program.
8535 *) Added "-genkey" option to "dsaparam" program.
8538 *) Added RIPE MD160 (rmd160) message digest.
8541 *) Added -a (all) option to "ssleay version" command.
8544 *) Added PLATFORM define which is the id given to Configure.
8547 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
8550 *) Extended the ASN.1 parser routines.
8553 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
8556 *) Added a BN_CTX to the BN library.
8559 *) Fixed the weak key values in DES library
8562 *) Changed API in EVP library for cipher aliases.
8565 *) Added support for RC2/64bit cipher.
8568 *) Converted the lhash library to the crypto/mem.c functions.
8571 *) Added more recognized ASN.1 object ids.
8574 *) Added more RSA padding checks for SSL/TLS.
8577 *) Added BIO proxy/filter functionality.
8580 *) Added extra_certs to SSL_CTX which can be used
8581 send extra CA certificates to the client in the CA cert chain sending
8582 process. It can be configured with SSL_CTX_add_extra_chain_cert().
8585 *) Now Fortezza is denied in the authentication phase because
8586 this is key exchange mechanism is not supported by SSLeay at all.
8589 *) Additional PKCS1 checks.
8592 *) Support the string "TLSv1" for all TLS v1 ciphers.
8595 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
8596 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
8599 *) Fixed a few memory leaks.
8602 *) Fixed various code and comment typos.
8605 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
8606 bytes sent in the client random.
8607 [Edward Bishop <ebishop@spyglass.com>]
projects / openssl.git / blob