5 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
7 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
8 of the Vaudenay padding oracle attack on CBC mode encryption
9 which enables an efficient plaintext recovery attack against
10 the OpenSSL implementation of DTLS. Their attack exploits timing
11 differences arising during decryption processing. A research
12 paper describing this attack can be found at:
13 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
14 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
15 Security Group at Royal Holloway, University of London
16 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
17 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
18 for preparing the fix. (CVE-2011-4108)
19 [Robin Seggelmann, Michael Tuexen]
21 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
22 [Ben Laurie, Kasper <ekasper@google.com>]
24 *) Clear bytes used for block padding of SSL 3.0 records.
26 [Adam Langley (Google)]
28 *) Only allow one SGC handshake restart for SSL/TLS. (CVE-2011-4619)
29 [Adam Langley (Google)]
31 *) Prevent malformed RFC3779 data triggering an assertion failure.
32 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
33 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
34 [Rob Austein <sra@hactrn.net>]
36 *) Fix ssl_ciph.c set-up race.
37 [Adam Langley (Google)]
39 *) Fix spurious failures in ecdsatest.c.
40 [Emilia Käsper (Google)]
42 *) Fix the BIO_f_buffer() implementation (which was mixing different
43 interpretations of the '..._len' fields).
44 [Adam Langley (Google)]
46 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
47 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
48 threads won't reuse the same blinding coefficients.
50 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
51 lock to call BN_BLINDING_invert_ex, and avoids one use of
52 BN_BLINDING_update for each BN_BLINDING structure (previously,
53 the last update always remained unused).
54 [Emilia Käsper (Google)]
56 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
57 for multi-threaded use of ECDH.
58 [Adam Langley (Google)]
60 *) Fix x509_name_ex_d2i memory leak on bad inputs.
63 *) Add protection against ECDSA timing attacks as mentioned in the paper
64 by Billy Bob Brumley and Nicola Tuveri, see:
66 http://eprint.iacr.org/2011/232.pdf
68 [Billy Bob Brumley and Nicola Tuveri]
70 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
72 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
73 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
75 *) Fix bug in string printing code: if *any* escaping is enabled we must
76 escape the escape character (backslash) or the resulting string is
80 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
82 *) Disable code workaround for ancient and obsolete Netscape browsers
83 and servers: an attacker can use it in a ciphersuite downgrade attack.
84 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
87 *) Fixed J-PAKE implementation error, originally discovered by
88 Sebastien Martini, further info and confirmation from Stefan
89 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
92 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
94 *) Fix extension code to avoid race conditions which can result in a buffer
95 overrun vulnerability: resumed sessions must not be modified as they can
96 be shared by multiple threads. CVE-2010-3864
99 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
102 *) Don't reencode certificate when calculating signature: cache and use
103 the original encoding instead. This makes signature verification of
104 some broken encodings work correctly.
107 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
108 is also one of the inputs.
109 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
111 *) Don't repeatedly append PBE algorithms to table if they already exist.
112 Sort table on each new add. This effectively makes the table read only
113 after all algorithms are added and subsequent calls to PKCS12_pbe_add
117 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
119 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
122 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
123 access or freeing data twice (CVE-2010-0742)
124 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
126 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
127 common in certificates and some applications which only call
128 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
132 Reduce copying into .apps and .test in makevms.com
133 Don't try to use blank CA certificate in CA.com
134 Allow use of C files from original directories in maketests.com
135 [Steven M. Schweda" <sms@antinode.info>]
137 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
139 *) When rejecting SSL/TLS records due to an incorrect version number, never
140 update s->server with a new major version number. As of
141 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
142 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
143 the previous behavior could result in a read attempt at NULL when
144 receiving specific incorrect SSL/TLS records once record payload
145 protection is active. (CVE-2010-0740)
146 [Bodo Moeller, Adam Langley <agl@chromium.org>]
148 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
149 could be crashed if the relevant tables were not present (e.g. chrooted).
150 [Tomas Hoger <thoger@redhat.com>]
152 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
154 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
155 [Martin Olsson, Neel Mehta]
157 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
158 accommodate for stack sorting, always a write lock!).
161 *) On some versions of WIN32 Heap32Next is very slow. This can cause
162 excessive delays in the RAND_poll(): over a minute. As a workaround
163 include a time check in the inner Heap32Next loop too.
166 *) The code that handled flushing of data in SSL/TLS originally used the
167 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
168 the problem outlined in PR#1949. The fix suggested there however can
169 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
170 of Apache). So instead simplify the code to flush unconditionally.
171 This should be fine since flushing with no data to flush is a no op.
174 *) Handle TLS versions 2.0 and later properly and correctly use the
175 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
176 off ancient servers have a habit of sticking around for a while...
179 *) Modify compression code so it frees up structures without using the
180 ex_data callbacks. This works around a problem where some applications
181 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
182 restarting) then use compression (e.g. SSL with compression) later.
183 This results in significant per-connection memory leaks and
184 has caused some security issues including CVE-2008-1678 and
188 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
189 change when encrypting or decrypting.
192 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
193 connect and renegotiate with servers which do not support RI.
194 Until RI is more widely deployed this option is enabled by default.
197 *) Add "missing" ssl ctrls to clear options and mode.
200 *) If client attempts to renegotiate and doesn't support RI respond with
201 a no_renegotiation alert as required by RFC5746. Some renegotiating
202 TLS clients will continue a connection gracefully when they receive
203 the alert. Unfortunately OpenSSL mishandled this alert and would hang
204 waiting for a server hello which it will never receive. Now we treat a
205 received no_renegotiation alert as a fatal error. This is because
206 applications requesting a renegotiation might well expect it to succeed
207 and would have no code in place to handle the server denying it so the
208 only safe thing to do is to terminate the connection.
211 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
212 peer supports secure renegotiation and 0 otherwise. Print out peer
213 renegotiation support in s_client/s_server.
216 *) Replace the highly broken and deprecated SPKAC certification method with
217 the updated NID creation version. This should correctly handle UTF8.
220 *) Implement RFC5746. Re-enable renegotiation but require the extension
221 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
222 turns out to be a bad idea. It has been replaced by
223 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
224 SSL_CTX_set_options(). This is really not recommended unless you
225 know what you are doing.
226 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
228 *) Fixes to stateless session resumption handling. Use initial_ctx when
229 issuing and attempting to decrypt tickets in case it has changed during
230 servername handling. Use a non-zero length session ID when attempting
231 stateless session resumption: this makes it possible to determine if
232 a resumption has occurred immediately after receiving server hello
233 (several places in OpenSSL subtly assume this) instead of later in
237 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
238 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
239 fixes for a few places where the return code is not checked
241 [Julia Lawall <julia@diku.dk>]
243 *) Add --strict-warnings option to Configure script to include devteam
244 warnings in other configurations.
247 *) Add support for --libdir option and LIBDIR variable in makefiles. This
248 makes it possible to install openssl libraries in locations which
249 have names other than "lib", for example "/usr/lib64" which some
251 [Steve Henson, based on patch from Jeremy Utley]
253 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
254 X690 8.9.12 and can produce some misleading textual output of OIDs.
255 [Steve Henson, reported by Dan Kaminsky]
257 *) Delete MD2 from algorithm tables. This follows the recommendation in
258 several standards that it is not used in new applications due to
259 several cryptographic weaknesses. For binary compatibility reasons
260 the MD2 API is still compiled in by default.
263 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
267 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
268 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
270 [Guenter <lists@gknw.net>]
272 *) Fix the server certificate chain building code to use X509_verify_cert(),
273 it used to have an ad-hoc builder which was unable to cope with anything
274 other than a simple chain.
275 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
277 *) Don't check self signed certificate signatures in X509_verify_cert()
278 by default (a flag can override this): it just wastes time without
279 adding any security. As a useful side effect self signed root CAs
280 with non-FIPS digests are now usable in FIPS mode.
283 *) In dtls1_process_out_of_seq_message() the check if the current message
284 is already buffered was missing. For every new message was memory
285 allocated, allowing an attacker to perform an denial of service attack
286 with sending out of seq handshake messages until there is no memory
287 left. Additionally every future messege was buffered, even if the
288 sequence number made no sense and would be part of another handshake.
289 So only messages with sequence numbers less than 10 in advance will be
290 buffered. (CVE-2009-1378)
291 [Robin Seggelmann, discovered by Daniel Mentz]
293 *) Records are buffered if they arrive with a future epoch to be
294 processed after finishing the corresponding handshake. There is
295 currently no limitation to this buffer allowing an attacker to perform
296 a DOS attack with sending records with future epochs until there is no
297 memory left. This patch adds the pqueue_size() function to detemine
298 the size of a buffer and limits the record buffer to 100 entries.
300 [Robin Seggelmann, discovered by Daniel Mentz]
302 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
303 parent structure is freed. (CVE-2009-1379)
306 *) Handle non-blocking I/O properly in SSL_shutdown() call.
307 [Darryl Miles <darryl-mailinglists@netbauds.net>]
310 [Ilya O. <vrghost@gmail.com>]
312 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
314 *) Disable renegotiation completely - this fixes a severe security
315 problem (CVE-2009-3555) at the cost of breaking all
316 renegotiation. Renegotiation can be re-enabled by setting
317 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
318 run-time. This is really not recommended unless you know what
322 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
324 *) Don't set val to NULL when freeing up structures, it is freed up by
325 underlying code. If sizeof(void *) > sizeof(long) this can result in
326 zeroing past the valid field. (CVE-2009-0789)
327 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
329 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
330 checked correctly. This would allow some invalid signed attributes to
331 appear to verify correctly. (CVE-2009-0591)
332 [Ivan Nestlerode <inestlerode@us.ibm.com>]
334 *) Reject UniversalString and BMPString types with invalid lengths. This
335 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
336 a legal length. (CVE-2009-0590)
339 *) Set S/MIME signing as the default purpose rather than setting it
340 unconditionally. This allows applications to override it at the store
344 *) Permit restricted recursion of ASN1 strings. This is needed in practice
345 to handle some structures.
348 *) Improve efficiency of mem_gets: don't search whole buffer each time
350 [Jeremy Shapiro <jnshapir@us.ibm.com>]
352 *) New -hex option for openssl rand.
355 *) Print out UTF8String and NumericString when parsing ASN1.
358 *) Support NumericString type for name components.
361 *) Allow CC in the environment to override the automatically chosen
362 compiler. Note that nothing is done to ensure flags work with the
366 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
368 *) Properly check EVP_VerifyFinal() and similar return values
370 [Ben Laurie, Bodo Moeller, Google Security Team]
372 *) Enable TLS extensions by default.
375 *) Allow the CHIL engine to be loaded, whether the application is
376 multithreaded or not. (This does not release the developer from the
377 obligation to set up the dynamic locking callbacks.)
378 [Sander Temme <sander@temme.net>]
380 *) Use correct exit code if there is an error in dgst command.
381 [Steve Henson; problem pointed out by Roland Dirlewanger]
383 *) Tweak Configure so that you need to say "experimental-jpake" to enable
384 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
387 *) Add experimental JPAKE support, including demo authentication in
388 s_client and s_server.
391 *) Set the comparison function in v3_addr_canonize().
392 [Rob Austein <sra@hactrn.net>]
394 *) Add support for XMPP STARTTLS in s_client.
395 [Philip Paeps <philip@freebsd.org>]
397 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
398 to ensure that even with this option, only ciphersuites in the
399 server's preference list will be accepted. (Note that the option
400 applies only when resuming a session, so the earlier behavior was
401 just about the algorithm choice for symmetric cryptography.)
404 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
406 *) Fix NULL pointer dereference if a DTLS server received
407 ChangeCipherSpec as first record (CVE-2009-1386).
410 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
411 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
414 *) The fix in 0.9.8c that supposedly got rid of unsafe
415 double-checked locking was incomplete for RSA blinding,
416 addressing just one layer of what turns out to have been
417 doubly unsafe triple-checked locking.
419 So now fix this for real by retiring the MONT_HELPER macro
420 in crypto/rsa/rsa_eay.c.
422 [Bodo Moeller; problem pointed out by Marius Schilder]
424 *) Various precautionary measures:
426 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
428 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
429 (NB: This would require knowledge of the secret session ticket key
430 to exploit, in which case you'd be SOL either way.)
432 - Change bn_nist.c so that it will properly handle input BIGNUMs
433 outside the expected range.
435 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
438 [Neel Mehta, Bodo Moeller]
440 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
441 the load fails. Useful for distros.
442 [Ben Laurie and the FreeBSD team]
444 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
447 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
450 *) Expand ENGINE to support engine supplied SSL client certificate functions.
452 This work was sponsored by Logica.
455 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
456 keystores. Support for SSL/TLS client authentication too.
457 Not compiled unless enable-capieng specified to Configure.
459 This work was sponsored by Logica.
462 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
463 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
464 attribute creation routines such as certifcate requests and PKCS#12
468 Changes between 0.9.8g and 0.9.8h [28 May 2008]
470 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
471 handshake which could lead to a cilent crash as found using the
472 Codenomicon TLS test suite (CVE-2008-1672)
473 [Steve Henson, Mark Cox]
475 *) Fix double free in TLS server name extensions which could lead to
476 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
479 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
481 Clear the error queue to ensure that error entries left from
482 older function calls do not interfere with the correct operation.
483 [Lutz Jaenicke, Erik de Castro Lopo]
485 *) Remove root CA certificates of commercial CAs:
487 The OpenSSL project does not recommend any specific CA and does not
488 have any policy with respect to including or excluding any CA.
489 Therefore it does not make any sense to ship an arbitrary selection
490 of root CA certificates with the OpenSSL software.
493 *) RSA OAEP patches to fix two separate invalid memory reads.
494 The first one involves inputs when 'lzero' is greater than
495 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
496 before the beginning of from). The second one involves inputs where
497 the 'db' section contains nothing but zeroes (there is a one-byte
498 invalid read after the end of 'db').
499 [Ivan Nestlerode <inestlerode@us.ibm.com>]
501 *) Partial backport from 0.9.9-dev:
503 Introduce bn_mul_mont (dedicated Montgomery multiplication
504 procedure) as a candidate for BIGNUM assembler implementation.
505 While 0.9.9-dev uses assembler for various architectures, only
506 x86_64 is available by default here in the 0.9.8 branch, and
507 32-bit x86 is available through a compile-time setting.
509 To try the 32-bit x86 assembler implementation, use Configure
510 option "enable-montasm" (which exists only for this backport).
512 As "enable-montasm" for 32-bit x86 disclaims code stability
513 anyway, in this constellation we activate additional code
514 backported from 0.9.9-dev for further performance improvements,
515 namely BN_from_montgomery_word. (To enable this otherwise,
516 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
518 [Andy Polyakov (backport partially by Bodo Moeller)]
520 *) Add TLS session ticket callback. This allows an application to set
521 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
522 values. This is useful for key rollover for example where several key
523 sets may exist with different names.
526 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
527 This was broken until now in 0.9.8 releases, such that the only way
528 a registered ENGINE could be used (assuming it initialises
529 successfully on the host) was to explicitly set it as the default
530 for the relevant algorithms. This is in contradiction with 0.9.7
531 behaviour and the documentation. With this fix, when an ENGINE is
532 registered into a given algorithm's table of implementations, the
533 'uptodate' flag is reset so that auto-discovery will be used next
534 time a new context for that algorithm attempts to select an
536 [Ian Lister (tweaked by Geoff Thorpe)]
538 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
539 implemention in the following ways:
541 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
544 Lack of BER streaming support means one pass streaming processing is
545 only supported if data is detached: setting the streaming flag is
546 ignored for embedded content.
548 CMS support is disabled by default and must be explicitly enabled
549 with the enable-cms configuration option.
552 *) Update the GMP engine glue to do direct copies between BIGNUM and
553 mpz_t when openssl and GMP use the same limb size. Otherwise the
554 existing "conversion via a text string export" trick is still used.
555 [Paul Sheer <paulsheer@gmail.com>]
557 *) Zlib compression BIO. This is a filter BIO which compressed and
558 uncompresses any data passed through it.
561 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
562 RFC3394 compatible AES key wrapping.
565 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
566 sets string data without copying. X509_ALGOR_set0() and
567 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
568 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
569 from an X509_ATTRIBUTE structure optionally checking it occurs only
570 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
574 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
575 to get the expected BN_FLG_CONSTTIME behavior.
576 [Bodo Moeller (Google)]
580 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
581 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
582 - added some more tests to do_tests.pl
583 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
584 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
585 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
586 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
587 - various changes to netware.pl to enable gcc-cross builds on Win32
589 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
590 - various changes to fix missing prototype warnings
591 - fixed x86nasm.pl to create correct asm files for NASM COFF output
592 - added AES, WHIRLPOOL and CPUID assembler code to build files
593 - added missing AES assembler make rules to mk1mf.pl
594 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
595 [Guenter Knauf <eflash@gmx.net>]
597 *) Implement certificate status request TLS extension defined in RFC3546.
598 A client can set the appropriate parameters and receive the encoded
599 OCSP response via a callback. A server can query the supplied parameters
600 and set the encoded OCSP response in the callback. Add simplified examples
601 to s_client and s_server.
604 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
607 + Binary incompatibility of ssl_ctx_st structure
608 + DTLS interoperation with non-compliant servers
609 + Don't call get_session_cb() without proposed session
610 + Fix ia64 assembler code
611 [Andy Polyakov, Steve Henson]
613 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
615 *) DTLS Handshake overhaul. There were longstanding issues with
616 OpenSSL DTLS implementation, which were making it impossible for
617 RFC 4347 compliant client to communicate with OpenSSL server.
618 Unfortunately just fixing these incompatibilities would "cut off"
619 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
620 server keeps tolerating non RFC compliant syntax. The opposite is
621 not true, 0.9.8f client can not communicate with earlier server.
622 This update even addresses CVE-2007-4995.
625 *) Changes to avoid need for function casts in OpenSSL: some compilers
626 (gcc 4.2 and later) reject their use.
627 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
630 *) Add RFC4507 support to OpenSSL. This includes the corrections in
631 RFC4507bis. The encrypted ticket format is an encrypted encoded
632 SSL_SESSION structure, that way new session features are automatically
635 If a client application caches session in an SSL_SESSION structure
636 support is transparent because tickets are now stored in the encoded
639 The SSL_CTX structure automatically generates keys for ticket
640 protection in servers so again support should be possible
641 with no application modification.
643 If a client or server wishes to disable RFC4507 support then the option
644 SSL_OP_NO_TICKET can be set.
646 Add a TLS extension debugging callback to allow the contents of any client
647 or server extensions to be examined.
649 This work was sponsored by Google.
652 *) Add initial support for TLS extensions, specifically for the server_name
653 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
654 have new members for a host name. The SSL data structure has an
655 additional member SSL_CTX *initial_ctx so that new sessions can be
656 stored in that context to allow for session resumption, even after the
657 SSL has been switched to a new SSL_CTX in reaction to a client's
658 server_name extension.
660 New functions (subject to change):
663 SSL_get_servername_type()
666 New CTRL codes and macros (subject to change):
668 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
669 - SSL_CTX_set_tlsext_servername_callback()
670 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
671 - SSL_CTX_set_tlsext_servername_arg()
672 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
674 openssl s_client has a new '-servername ...' option.
676 openssl s_server has new options '-servername_host ...', '-cert2 ...',
677 '-key2 ...', '-servername_fatal' (subject to change). This allows
678 testing the HostName extension for a specific single host name ('-cert'
679 and '-key' remain fallbacks for handshakes without HostName
680 negotiation). If the unrecogninzed_name alert has to be sent, this by
681 default is a warning; it becomes fatal with the '-servername_fatal'
684 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
686 *) Add AES and SSE2 assembly language support to VC++ build.
689 *) Mitigate attack on final subtraction in Montgomery reduction.
692 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
693 (which previously caused an internal error).
696 *) Squeeze another 10% out of IGE mode when in != out.
699 *) AES IGE mode speedup.
700 [Dean Gaudet (Google)]
702 *) Add the Korean symmetric 128-bit cipher SEED (see
703 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
704 add SEED ciphersuites from RFC 4162:
706 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
707 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
708 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
709 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
711 To minimize changes between patchlevels in the OpenSSL 0.9.8
712 series, SEED remains excluded from compilation unless OpenSSL
713 is configured with 'enable-seed'.
716 *) Mitigate branch prediction attacks, which can be practical if a
717 single processor is shared, allowing a spy process to extract
718 information. For detailed background information, see
719 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
720 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
721 and Necessary Software Countermeasures"). The core of the change
722 are new versions BN_div_no_branch() and
723 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
724 respectively, which are slower, but avoid the security-relevant
725 conditional branches. These are automatically called by BN_div()
726 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
727 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
728 remove a conditional branch.
730 BN_FLG_CONSTTIME is the new name for the previous
731 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
732 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
733 in the exponent causes BN_mod_exp_mont() to use the alternative
734 implementation in BN_mod_exp_mont_consttime().) The old name
735 remains as a deprecated alias.
737 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
738 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
739 constant-time implementations for more than just exponentiation.
740 Here too the old name is kept as a deprecated alias.
742 BN_BLINDING_new() will now use BN_dup() for the modulus so that
743 the BN_BLINDING structure gets an independent copy of the
744 modulus. This means that the previous "BIGNUM *m" argument to
745 BN_BLINDING_new() and to BN_BLINDING_create_param() now
746 essentially becomes "const BIGNUM *m", although we can't actually
747 change this in the header file before 0.9.9. It allows
748 RSA_setup_blinding() to use BN_with_flags() on the modulus to
749 enable BN_FLG_CONSTTIME.
751 [Matthew D Wood (Intel Corp)]
753 *) In the SSL/TLS server implementation, be strict about session ID
754 context matching (which matters if an application uses a single
755 external cache for different purposes). Previously,
756 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
757 set. This did ensure strict client verification, but meant that,
758 with applications using a single external cache for quite
759 different requirements, clients could circumvent ciphersuite
760 restrictions for a given session ID context by starting a session
761 in a different context.
764 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
765 a ciphersuite string such as "DEFAULT:RSA" cannot enable
766 authentication-only ciphersuites.
769 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
770 not complete and could lead to a possible single byte overflow
771 (CVE-2007-5135) [Ben Laurie]
773 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
775 *) Since AES128 and AES256 (and similarly Camellia128 and
776 Camellia256) share a single mask bit in the logic of
777 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
778 kludge to work properly if AES128 is available and AES256 isn't
779 (or if Camellia128 is available and Camellia256 isn't).
782 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
783 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
784 When a point or a seed is encoded in a BIT STRING, we need to
785 prevent the removal of trailing zero bits to get the proper DER
786 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
787 of a NamedBitList, for which trailing 0 bits need to be removed.)
790 *) Have SSL/TLS server implementation tolerate "mismatched" record
791 protocol version while receiving ClientHello even if the
792 ClientHello is fragmented. (The server can't insist on the
793 particular protocol version it has chosen before the ServerHello
794 message has informed the client about his choice.)
797 *) Add RFC 3779 support.
798 [Rob Austein for ARIN, Ben Laurie]
800 *) Load error codes if they are not already present instead of using a
801 static variable. This allows them to be cleanly unloaded and reloaded.
802 Improve header file function name parsing.
805 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
806 or CAPABILITY handshake as required by RFCs.
809 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
811 *) Introduce limits to prevent malicious keys being able to
812 cause a denial of service. (CVE-2006-2940)
813 [Steve Henson, Bodo Moeller]
815 *) Fix ASN.1 parsing of certain invalid structures that can result
816 in a denial of service. (CVE-2006-2937) [Steve Henson]
818 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
819 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
821 *) Fix SSL client code which could crash if connecting to a
822 malicious SSLv2 server. (CVE-2006-4343)
823 [Tavis Ormandy and Will Drewry, Google Security Team]
825 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
826 match only those. Before that, "AES256-SHA" would be interpreted
827 as a pattern and match "AES128-SHA" too (since AES128-SHA got
828 the same strength classification in 0.9.7h) as we currently only
829 have a single AES bit in the ciphersuite description bitmap.
830 That change, however, also applied to ciphersuite strings such as
831 "RC4-MD5" that intentionally matched multiple ciphersuites --
832 namely, SSL 2.0 ciphersuites in addition to the more common ones
833 from SSL 3.0/TLS 1.0.
835 So we change the selection algorithm again: Naming an explicit
836 ciphersuite selects this one ciphersuite, and any other similar
837 ciphersuite (same bitmap) from *other* protocol versions.
838 Thus, "RC4-MD5" again will properly select both the SSL 2.0
839 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
841 Since SSL 2.0 does not have any ciphersuites for which the
842 128/256 bit distinction would be relevant, this works for now.
843 The proper fix will be to use different bits for AES128 and
844 AES256, which would have avoided the problems from the beginning;
845 however, bits are scarce, so we can only do this in a new release
846 (not just a patchlevel) when we can change the SSL_CIPHER
847 definition to split the single 'unsigned long mask' bitmap into
848 multiple values to extend the available space.
852 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
854 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
855 (CVE-2006-4339) [Ben Laurie and Google Security Team]
857 *) Add AES IGE and biIGE modes.
860 *) Change the Unix randomness entropy gathering to use poll() when
861 possible instead of select(), since the latter has some
862 undesirable limitations.
863 [Darryl Miles via Richard Levitte and Bodo Moeller]
865 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
866 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
867 cannot be implicitly activated as part of, e.g., the "AES" alias.
868 However, please upgrade to OpenSSL 0.9.9[-dev] for
869 non-experimental use of the ECC ciphersuites to get TLS extension
870 support, which is required for curve and point format negotiation
871 to avoid potential handshake problems.
874 *) Disable rogue ciphersuites:
876 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
877 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
878 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
880 The latter two were purportedly from
881 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
884 Also deactivate the remaining ciphersuites from
885 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
886 unofficial, and the ID has long expired.
889 *) Fix RSA blinding Heisenbug (problems sometimes occured on
890 dual-core machines) and other potential thread-safety issues.
893 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
894 versions), which is now available for royalty-free use
895 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
896 Also, add Camellia TLS ciphersuites from RFC 4132.
898 To minimize changes between patchlevels in the OpenSSL 0.9.8
899 series, Camellia remains excluded from compilation unless OpenSSL
900 is configured with 'enable-camellia'.
903 *) Disable the padding bug check when compression is in use. The padding
904 bug check assumes the first packet is of even length, this is not
905 necessarily true if compresssion is enabled and can result in false
906 positives causing handshake failure. The actual bug test is ancient
907 code so it is hoped that implementations will either have fixed it by
908 now or any which still have the bug do not support compression.
911 Changes between 0.9.8a and 0.9.8b [04 May 2006]
913 *) When applying a cipher rule check to see if string match is an explicit
914 cipher suite and only match that one cipher suite if it is.
917 *) Link in manifests for VC++ if needed.
918 [Austin Ziegler <halostatue@gmail.com>]
920 *) Update support for ECC-based TLS ciphersuites according to
921 draft-ietf-tls-ecc-12.txt with proposed changes (but without
922 TLS extensions, which are supported starting with the 0.9.9
923 branch, not in the OpenSSL 0.9.8 branch).
926 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
927 opaque EVP_CIPHER_CTX handling.
930 *) Fixes and enhancements to zlib compression code. We now only use
931 "zlib1.dll" and use the default __cdecl calling convention on Win32
932 to conform with the standards mentioned here:
933 http://www.zlib.net/DLL_FAQ.txt
934 Static zlib linking now works on Windows and the new --with-zlib-include
935 --with-zlib-lib options to Configure can be used to supply the location
936 of the headers and library. Gracefully handle case where zlib library
940 *) Several fixes and enhancements to the OID generation code. The old code
941 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
942 handle numbers larger than ULONG_MAX, truncated printing and had a
943 non standard OBJ_obj2txt() behaviour.
946 *) Add support for building of engines under engine/ as shared libraries
947 under VC++ build system.
950 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
951 Hopefully, we will not see any false combination of paths any more.
954 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
956 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
957 (part of SSL_OP_ALL). This option used to disable the
958 countermeasure against man-in-the-middle protocol-version
959 rollback in the SSL 2.0 server implementation, which is a bad
960 idea. (CVE-2005-2969)
962 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
963 for Information Security, National Institute of Advanced Industrial
964 Science and Technology [AIST], Japan)]
966 *) Add two function to clear and return the verify parameter flags.
969 *) Keep cipherlists sorted in the source instead of sorting them at
970 runtime, thus removing the need for a lock.
973 *) Avoid some small subgroup attacks in Diffie-Hellman.
974 [Nick Mathewson and Ben Laurie]
976 *) Add functions for well-known primes.
979 *) Extended Windows CE support.
980 [Satoshi Nakamura and Andy Polyakov]
982 *) Initialize SSL_METHOD structures at compile time instead of during
983 runtime, thus removing the need for a lock.
986 *) Make PKCS7_decrypt() work even if no certificate is supplied by
987 attempting to decrypt each encrypted key in turn. Add support to
991 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
993 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
996 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
999 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
1000 key into the same file any more.
1003 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
1006 *) Add -utf8 command line and config file option to 'ca'.
1007 [Stefan <stf@udoma.org]
1009 *) Removed the macro des_crypt(), as it seems to conflict with some
1010 libraries. Use DES_crypt().
1013 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
1014 involves renaming the source and generated shared-libs for
1015 both. The engines will accept the corrected or legacy ids
1016 ('ncipher' and '4758_cca' respectively) when binding. NB,
1017 this only applies when building 'shared'.
1018 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
1020 *) Add attribute functions to EVP_PKEY structure. Modify
1021 PKCS12_create() to recognize a CSP name attribute and
1022 use it. Make -CSP option work again in pkcs12 utility.
1025 *) Add new functionality to the bn blinding code:
1026 - automatic re-creation of the BN_BLINDING parameters after
1027 a fixed number of uses (currently 32)
1028 - add new function for parameter creation
1029 - introduce flags to control the update behaviour of the
1030 BN_BLINDING parameters
1031 - hide BN_BLINDING structure
1032 Add a second BN_BLINDING slot to the RSA structure to improve
1033 performance when a single RSA object is shared among several
1037 *) Add support for DTLS.
1038 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
1040 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
1041 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
1044 *) Remove buggy and incompletet DH cert support from
1045 ssl/ssl_rsa.c and ssl/s3_both.c
1048 *) Use SHA-1 instead of MD5 as the default digest algorithm for
1049 the apps/openssl applications.
1052 *) Compile clean with "-Wall -Wmissing-prototypes
1053 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
1054 DEBUG_SAFESTACK must also be set.
1057 *) Change ./Configure so that certain algorithms can be disabled by default.
1058 The new counterpiece to "no-xxx" is "enable-xxx".
1060 The patented RC5 and MDC2 algorithms will now be disabled unless
1061 "enable-rc5" and "enable-mdc2", respectively, are specified.
1063 (IDEA remains enabled despite being patented. This is because IDEA
1064 is frequently required for interoperability, and there is no license
1065 fee for non-commercial use. As before, "no-idea" can be used to
1066 avoid this algorithm.)
1070 *) Add processing of proxy certificates (see RFC 3820). This work was
1071 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
1072 EGEE (Enabling Grids for E-science in Europe).
1075 *) RC4 performance overhaul on modern architectures/implementations, such
1076 as Intel P4, IA-64 and AMD64.
1079 *) New utility extract-section.pl. This can be used specify an alternative
1080 section number in a pod file instead of having to treat each file as
1081 a separate case in Makefile. This can be done by adding two lines to the
1084 =for comment openssl_section:XXX
1086 The blank line is mandatory.
1090 *) New arguments -certform, -keyform and -pass for s_client and s_server
1091 to allow alternative format key and certificate files and passphrase
1095 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
1096 update associated structures and add various utility functions.
1098 Add new policy related verify parameters, include policy checking in
1099 standard verify code. Enhance 'smime' application with extra parameters
1100 to support policy checking and print out.
1103 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
1104 Nehemiah processors. These extensions support AES encryption in hardware
1105 as well as RNG (though RNG support is currently disabled).
1106 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
1108 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
1111 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
1112 [Andy Polyakov and a number of other people]
1114 *) Improved PowerPC platform support. Most notably BIGNUM assembler
1115 implementation contributed by IBM.
1116 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
1118 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
1119 exponent rather than 'unsigned long'. There is a corresponding change to
1120 the new 'rsa_keygen' element of the RSA_METHOD structure.
1121 [Jelte Jansen, Geoff Thorpe]
1123 *) Functionality for creating the initial serial number file is now
1124 moved from CA.pl to the 'ca' utility with a new option -create_serial.
1126 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
1127 number file to 1, which is bound to cause problems. To avoid
1128 the problems while respecting compatibility between different 0.9.7
1129 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
1130 CA.pl for serial number initialization. With the new release 0.9.8,
1131 we can fix the problem directly in the 'ca' utility.)
1134 *) Reduced header interdepencies by declaring more opaque objects in
1135 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
1136 give fewer recursive includes, which could break lazy source code - so
1137 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
1138 developers should define this symbol when building and using openssl to
1139 ensure they track the recommended behaviour, interfaces, [etc], but
1140 backwards-compatible behaviour prevails when this isn't defined.
1143 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
1146 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
1147 This will generate a random key of the appropriate length based on the
1148 cipher context. The EVP_CIPHER can provide its own random key generation
1149 routine to support keys of a specific form. This is used in the des and
1150 3des routines to generate a key of the correct parity. Update S/MIME
1151 code to use new functions and hence generate correct parity DES keys.
1152 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
1153 valid (weak or incorrect parity).
1156 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
1157 as looking them up. This is useful when the verified structure may contain
1158 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
1159 present unless the new PKCS7_NO_CRL flag is asserted.
1162 *) Extend ASN1 oid configuration module. It now additionally accepts the
1165 shortName = some long name, 1.2.3.4
1168 *) Reimplemented the BN_CTX implementation. There is now no more static
1169 limitation on the number of variables it can handle nor the depth of the
1170 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
1171 information can now expand as required, and rather than having a single
1172 static array of bignums, BN_CTX now uses a linked-list of such arrays
1173 allowing it to expand on demand whilst maintaining the usefulness of
1174 BN_CTX's "bundling".
1177 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
1178 to allow all RSA operations to function using a single BN_CTX.
1181 *) Preliminary support for certificate policy evaluation and checking. This
1182 is initially intended to pass the tests outlined in "Conformance Testing
1183 of Relying Party Client Certificate Path Processing Logic" v1.07.
1186 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
1187 remained unused and not that useful. A variety of other little bignum
1188 tweaks and fixes have also been made continuing on from the audit (see
1192 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
1193 associated ASN1, EVP and SSL functions and old ASN1 macros.
1196 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
1197 and this should never fail. So the return value from the use of
1198 BN_set_word() (which can fail due to needless expansion) is now deprecated;
1199 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
1202 *) BN_CTX_get() should return zero-valued bignums, providing the same
1203 initialised value as BN_new().
1204 [Geoff Thorpe, suggested by Ulf Möller]
1206 *) Support for inhibitAnyPolicy certificate extension.
1209 *) An audit of the BIGNUM code is underway, for which debugging code is
1210 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
1211 is considered valid when processing BIGNUMs, and causes execution to
1212 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
1213 further steps are taken to deliberately pollute unused data in BIGNUM
1214 structures to try and expose faulty code further on. For now, openssl will
1215 (in its default mode of operation) continue to tolerate the inconsistent
1216 forms that it has tolerated in the past, but authors and packagers should
1217 consider trying openssl and their own applications when compiled with
1218 these debugging symbols defined. It will help highlight potential bugs in
1219 their own code, and will improve the test coverage for OpenSSL itself. At
1220 some point, these tighter rules will become openssl's default to improve
1221 maintainability, though the assert()s and other overheads will remain only
1222 in debugging configurations. See bn.h for more details.
1223 [Geoff Thorpe, Nils Larsch, Ulf Möller]
1225 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
1226 that can only be obtained through BN_CTX_new() (which implicitly
1227 initialises it). The presence of this function only made it possible
1228 to overwrite an existing structure (and cause memory leaks).
1231 *) Because of the callback-based approach for implementing LHASH as a
1232 template type, lh_insert() adds opaque objects to hash-tables and
1233 lh_doall() or lh_doall_arg() are typically used with a destructor callback
1234 to clean up those corresponding objects before destroying the hash table
1235 (and losing the object pointers). So some over-zealous constifications in
1236 LHASH have been relaxed so that lh_insert() does not take (nor store) the
1237 objects as "const" and the lh_doall[_arg] callback wrappers are not
1238 prototyped to have "const" restrictions on the object pointers they are
1239 given (and so aren't required to cast them away any more).
1242 *) The tmdiff.h API was so ugly and minimal that our own timing utility
1243 (speed) prefers to use its own implementation. The two implementations
1244 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
1245 its object type properly exposed (MS_TM) instead of casting to/from "char
1246 *". This may still change yet if someone realises MS_TM and "ms_time_***"
1247 aren't necessarily the greatest nomenclatures - but this is what was used
1248 internally to the implementation so I've used that for now.
1251 *) Ensure that deprecated functions do not get compiled when
1252 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
1253 the self-tests were still using deprecated key-generation functions so
1254 these have been updated also.
1257 *) Reorganise PKCS#7 code to separate the digest location functionality
1258 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
1259 New function PKCS7_set_digest() to set the digest type for PKCS#7
1260 digestedData type. Add additional code to correctly generate the
1261 digestedData type and add support for this type in PKCS7 initialization
1265 *) New function PKCS7_set0_type_other() this initializes a PKCS7
1266 structure of type "other".
1269 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
1270 sure the loop does correctly stop and breaking ("division by zero")
1271 modulus operations are not performed. The (pre-generated) prime
1272 table crypto/bn/bn_prime.h was already correct, but it could not be
1273 re-generated on some platforms because of the "division by zero"
1274 situation in the script.
1275 [Ralf S. Engelschall]
1277 *) Update support for ECC-based TLS ciphersuites according to
1278 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
1279 SHA-1 now is only used for "small" curves (where the
1280 representation of a field element takes up to 24 bytes); for
1281 larger curves, the field element resulting from ECDH is directly
1282 used as premaster secret.
1283 [Douglas Stebila (Sun Microsystems Laboratories)]
1285 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
1286 curve secp160r1 to the tests.
1287 [Douglas Stebila (Sun Microsystems Laboratories)]
1289 *) Add the possibility to load symbols globally with DSO.
1290 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
1292 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
1293 control of the error stack.
1296 *) Add support for STORE in ENGINE.
1299 *) Add the STORE type. The intention is to provide a common interface
1300 to certificate and key stores, be they simple file-based stores, or
1301 HSM-type store, or LDAP stores, or...
1302 NOTE: The code is currently UNTESTED and isn't really used anywhere.
1305 *) Add a generic structure called OPENSSL_ITEM. This can be used to
1306 pass a list of arguments to any function as well as provide a way
1307 for a function to pass data back to the caller.
1310 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
1311 works like BUF_strdup() but can be used to duplicate a portion of
1312 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
1316 *) Add the function sk_find_ex() which works like sk_find(), but will
1317 return an index to an element even if an exact match couldn't be
1318 found. The index is guaranteed to point at the element where the
1319 searched-for key would be inserted to preserve sorting order.
1322 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
1323 takes an extra flags argument for optional functionality. Currently,
1324 the following flags are defined:
1326 OBJ_BSEARCH_VALUE_ON_NOMATCH
1327 This one gets OBJ_bsearch_ex() to return a pointer to the first
1328 element where the comparing function returns a negative or zero
1331 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
1332 This one gets OBJ_bsearch_ex() to return a pointer to the first
1333 element where the comparing function returns zero. This is useful
1334 if there are more than one element where the comparing function
1338 *) Make it possible to create self-signed certificates with 'openssl ca'
1339 in such a way that the self-signed certificate becomes part of the
1340 CA database and uses the same mechanisms for serial number generation
1341 as all other certificate signing. The new flag '-selfsign' enables
1342 this functionality. Adapt CA.sh and CA.pl.in.
1345 *) Add functionality to check the public key of a certificate request
1346 against a given private. This is useful to check that a certificate
1347 request can be signed by that key (self-signing).
1350 *) Make it possible to have multiple active certificates with the same
1351 subject in the CA index file. This is done only if the keyword
1352 'unique_subject' is set to 'no' in the main CA section (default
1353 if 'CA_default') of the configuration file. The value is saved
1354 with the database itself in a separate index attribute file,
1355 named like the index file with '.attr' appended to the name.
1358 *) Generate muti valued AVAs using '+' notation in config files for
1362 *) Support for nameConstraints certificate extension.
1365 *) Support for policyConstraints certificate extension.
1368 *) Support for policyMappings certificate extension.
1371 *) Make sure the default DSA_METHOD implementation only uses its
1372 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
1373 and change its own handlers to be NULL so as to remove unnecessary
1374 indirection. This lets alternative implementations fallback to the
1375 default implementation more easily.
1378 *) Support for directoryName in GeneralName related extensions
1382 *) Make it possible to link applications using Makefile.shared.
1383 Make that possible even when linking against static libraries!
1386 *) Support for single pass processing for S/MIME signing. This now
1387 means that S/MIME signing can be done from a pipe, in addition
1388 cleartext signing (multipart/signed type) is effectively streaming
1389 and the signed data does not need to be all held in memory.
1391 This is done with a new flag PKCS7_STREAM. When this flag is set
1392 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
1393 is done after the data is output (and digests calculated) in
1394 SMIME_write_PKCS7().
1397 *) Add full support for -rpath/-R, both in shared libraries and
1398 applications, at least on the platforms where it's known how
1402 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
1403 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
1404 will now compute a table of multiples of the generator that
1405 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
1406 faster (notably in the case of a single point multiplication,
1407 scalar * generator).
1408 [Nils Larsch, Bodo Moeller]
1410 *) IPv6 support for certificate extensions. The various extensions
1411 which use the IP:a.b.c.d can now take IPv6 addresses using the
1412 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
1416 *) Added an ENGINE that implements RSA by performing private key
1417 exponentiations with the GMP library. The conversions to and from
1418 GMP's mpz_t format aren't optimised nor are any montgomery forms
1419 cached, and on x86 it appears OpenSSL's own performance has caught up.
1420 However there are likely to be other architectures where GMP could
1421 provide a boost. This ENGINE is not built in by default, but it can be
1422 specified at Configure time and should be accompanied by the necessary
1423 linker additions, eg;
1424 ./config -DOPENSSL_USE_GMP -lgmp
1427 *) "openssl engine" will not display ENGINE/DSO load failure errors when
1428 testing availability of engines with "-t" - the old behaviour is
1429 produced by increasing the feature's verbosity with "-tt".
1432 *) ECDSA routines: under certain error conditions uninitialized BN objects
1433 could be freed. Solution: make sure initialization is performed early
1434 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
1438 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
1439 and DH_METHOD (eg. by ENGINE implementations) to override the normal
1440 software implementations. For DSA and DH, parameter generation can
1441 also be overriden by providing the appropriate method callbacks.
1444 *) Change the "progress" mechanism used in key-generation and
1445 primality testing to functions that take a new BN_GENCB pointer in
1446 place of callback/argument pairs. The new API functions have "_ex"
1447 postfixes and the older functions are reimplemented as wrappers for
1448 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
1449 declarations of the old functions to help (graceful) attempts to
1450 migrate to the new functions. Also, the new key-generation API
1451 functions operate on a caller-supplied key-structure and return
1452 success/failure rather than returning a key or NULL - this is to
1453 help make "keygen" another member function of RSA_METHOD etc.
1455 Example for using the new callback interface:
1457 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
1461 BN_GENCB_set(&my_cb, my_callback, my_arg);
1463 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
1464 /* For the meaning of a, b in calls to my_callback(), see the
1465 * documentation of the function that calls the callback.
1466 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
1467 * my_callback should return 1 if it wants BN_is_prime_ex()
1468 * to continue, or 0 to stop.
1473 *) Change the ZLIB compression method to be stateful, and make it
1474 available to TLS with the number defined in
1475 draft-ietf-tls-compression-04.txt.
1478 *) Add the ASN.1 structures and functions for CertificatePair, which
1479 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
1481 CertificatePair ::= SEQUENCE {
1482 forward [0] Certificate OPTIONAL,
1483 reverse [1] Certificate OPTIONAL,
1484 -- at least one of the pair shall be present -- }
1486 Also implement the PEM functions to read and write certificate
1487 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
1489 This needed to be defined, mostly for the sake of the LDAP
1490 attribute crossCertificatePair, but may prove useful elsewhere as
1494 *) Make it possible to inhibit symlinking of shared libraries in
1495 Makefile.shared, for Cygwin's sake.
1498 *) Extend the BIGNUM API by creating a function
1499 void BN_set_negative(BIGNUM *a, int neg);
1500 and a macro that behave like
1501 int BN_is_negative(const BIGNUM *a);
1503 to avoid the need to access 'a->neg' directly in applications.
1506 *) Implement fast modular reduction for pseudo-Mersenne primes
1507 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
1508 EC_GROUP_new_curve_GFp() will now automatically use this
1510 [Nils Larsch <nla@trustcenter.de>]
1512 *) Add new lock type (CRYPTO_LOCK_BN).
1515 *) Change the ENGINE framework to automatically load engines
1516 dynamically from specific directories unless they could be
1517 found to already be built in or loaded. Move all the
1518 current engines except for the cryptodev one to a new
1520 The engines in engines/ are built as shared libraries if
1521 the "shared" options was given to ./Configure or ./config.
1522 Otherwise, they are inserted in libcrypto.a.
1523 /usr/local/ssl/engines is the default directory for dynamic
1524 engines, but that can be overriden at configure time through
1525 the usual use of --prefix and/or --openssldir, and at run
1526 time with the environment variable OPENSSL_ENGINES.
1527 [Geoff Thorpe and Richard Levitte]
1529 *) Add Makefile.shared, a helper makefile to build shared
1530 libraries. Addapt Makefile.org.
1533 *) Add version info to Win32 DLLs.
1534 [Peter 'Luna' Runestig" <peter@runestig.com>]
1536 *) Add new 'medium level' PKCS#12 API. Certificates and keys
1537 can be added using this API to created arbitrary PKCS#12
1538 files while avoiding the low level API.
1540 New options to PKCS12_create(), key or cert can be NULL and
1541 will then be omitted from the output file. The encryption
1542 algorithm NIDs can be set to -1 for no encryption, the mac
1543 iteration count can be set to 0 to omit the mac.
1545 Enhance pkcs12 utility by making the -nokeys and -nocerts
1546 options work when creating a PKCS#12 file. New option -nomac
1547 to omit the mac, NONE can be set for an encryption algorithm.
1548 New code is modified to use the enhanced PKCS12_create()
1549 instead of the low level API.
1552 *) Extend ASN1 encoder to support indefinite length constructed
1553 encoding. This can output sequences tags and octet strings in
1554 this form. Modify pk7_asn1.c to support indefinite length
1555 encoding. This is experimental and needs additional code to
1556 be useful, such as an ASN1 bio and some enhanced streaming
1559 Extend template encode functionality so that tagging is passed
1560 down to the template encoder.
1563 *) Let 'openssl req' fail if an argument to '-newkey' is not
1564 recognized instead of using RSA as a default.
1567 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
1568 As these are not official, they are not included in "ALL";
1569 the "ECCdraft" ciphersuite group alias can be used to select them.
1570 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
1572 *) Add ECDH engine support.
1573 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
1575 *) Add ECDH in new directory crypto/ecdh/.
1576 [Douglas Stebila (Sun Microsystems Laboratories)]
1578 *) Let BN_rand_range() abort with an error after 100 iterations
1579 without success (which indicates a broken PRNG).
1582 *) Change BN_mod_sqrt() so that it verifies that the input value
1583 is really the square of the return value. (Previously,
1584 BN_mod_sqrt would show GIGO behaviour.)
1587 *) Add named elliptic curves over binary fields from X9.62, SECG,
1588 and WAP/WTLS; add OIDs that were still missing.
1590 [Sheueling Chang Shantz and Douglas Stebila
1591 (Sun Microsystems Laboratories)]
1593 *) Extend the EC library for elliptic curves over binary fields
1594 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
1597 EC_GF2m_simple_method
1601 EC_GROUP_new_curve_GF2m
1602 EC_GROUP_set_curve_GF2m
1603 EC_GROUP_get_curve_GF2m
1604 EC_POINT_set_affine_coordinates_GF2m
1605 EC_POINT_get_affine_coordinates_GF2m
1606 EC_POINT_set_compressed_coordinates_GF2m
1608 Point compression for binary fields is disabled by default for
1609 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
1612 As binary polynomials are represented as BIGNUMs, various members
1613 of the EC_GROUP and EC_POINT data structures can be shared
1614 between the implementations for prime fields and binary fields;
1615 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
1616 are essentially identical to their ..._GFp counterparts.
1617 (For simplicity, the '..._GFp' prefix has been dropped from
1618 various internal method names.)
1620 An internal 'field_div' method (similar to 'field_mul' and
1621 'field_sqr') has been added; this is used only for binary fields.
1623 [Sheueling Chang Shantz and Douglas Stebila
1624 (Sun Microsystems Laboratories)]
1626 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
1627 through methods ('mul', 'precompute_mult').
1629 The generic implementations (now internally called 'ec_wNAF_mul'
1630 and 'ec_wNAF_precomputed_mult') remain the default if these
1631 methods are undefined.
1633 [Sheueling Chang Shantz and Douglas Stebila
1634 (Sun Microsystems Laboratories)]
1636 *) New function EC_GROUP_get_degree, which is defined through
1637 EC_METHOD. For curves over prime fields, this returns the bit
1638 length of the modulus.
1640 [Sheueling Chang Shantz and Douglas Stebila
1641 (Sun Microsystems Laboratories)]
1643 *) New functions EC_GROUP_dup, EC_POINT_dup.
1644 (These simply call ..._new and ..._copy).
1646 [Sheueling Chang Shantz and Douglas Stebila
1647 (Sun Microsystems Laboratories)]
1649 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
1650 Polynomials are represented as BIGNUMs (where the sign bit is not
1651 used) in the following functions [macros]:
1654 BN_GF2m_sub [= BN_GF2m_add]
1655 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
1656 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
1657 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
1659 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
1660 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
1661 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
1662 BN_GF2m_cmp [= BN_ucmp]
1664 (Note that only the 'mod' functions are actually for fields GF(2^m).
1665 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
1667 For some functions, an the irreducible polynomial defining a
1668 field can be given as an 'unsigned int[]' with strictly
1669 decreasing elements giving the indices of those bits that are set;
1670 i.e., p[] represents the polynomial
1671 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
1673 p[0] > p[1] > ... > p[k] = 0.
1674 This applies to the following functions:
1679 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
1680 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
1682 BN_GF2m_mod_sqrt_arr
1683 BN_GF2m_mod_solve_quad_arr
1687 Conversion can be performed by the following functions:
1692 bntest.c has additional tests for binary polynomial arithmetic.
1694 Two implementations for BN_GF2m_mod_div() are available.
1695 The default algorithm simply uses BN_GF2m_mod_inv() and
1696 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
1697 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
1698 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
1700 [Sheueling Chang Shantz and Douglas Stebila
1701 (Sun Microsystems Laboratories)]
1703 *) Add new error code 'ERR_R_DISABLED' that can be used when some
1704 functionality is disabled at compile-time.
1705 [Douglas Stebila <douglas.stebila@sun.com>]
1707 *) Change default behaviour of 'openssl asn1parse' so that more
1708 information is visible when viewing, e.g., a certificate:
1710 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
1711 mode the content of non-printable OCTET STRINGs is output in a
1712 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
1713 avoid the appearance of a printable string.
1714 [Nils Larsch <nla@trustcenter.de>]
1716 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
1718 EC_GROUP_set_asn1_flag()
1719 EC_GROUP_get_asn1_flag()
1720 EC_GROUP_set_point_conversion_form()
1721 EC_GROUP_get_point_conversion_form()
1722 These control ASN1 encoding details:
1723 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
1724 has been set to OPENSSL_EC_NAMED_CURVE.
1725 - Points are encoded in uncompressed form by default; options for
1726 asn1_for are as for point2oct, namely
1727 POINT_CONVERSION_COMPRESSED
1728 POINT_CONVERSION_UNCOMPRESSED
1729 POINT_CONVERSION_HYBRID
1731 Also add 'seed' and 'seed_len' members to EC_GROUP with access
1734 EC_GROUP_get0_seed()
1735 EC_GROUP_get_seed_len()
1736 This is used only for ASN1 purposes (so far).
1737 [Nils Larsch <nla@trustcenter.de>]
1739 *) Add 'field_type' member to EC_METHOD, which holds the NID
1740 of the appropriate field type OID. The new function
1741 EC_METHOD_get_field_type() returns this value.
1742 [Nils Larsch <nla@trustcenter.de>]
1747 EC_POINT_point2hex()
1748 EC_POINT_hex2point()
1749 providing useful interfaces to EC_POINT_point2oct() and
1750 EC_POINT_oct2point().
1751 [Nils Larsch <nla@trustcenter.de>]
1753 *) Change internals of the EC library so that the functions
1754 EC_GROUP_set_generator()
1755 EC_GROUP_get_generator()
1756 EC_GROUP_get_order()
1757 EC_GROUP_get_cofactor()
1758 are implemented directly in crypto/ec/ec_lib.c and not dispatched
1759 to methods, which would lead to unnecessary code duplication when
1760 adding different types of curves.
1761 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
1763 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
1764 arithmetic, and such that modified wNAFs are generated
1765 (which avoid length expansion in many cases).
1768 *) Add a function EC_GROUP_check_discriminant() (defined via
1769 EC_METHOD) that verifies that the curve discriminant is non-zero.
1771 Add a function EC_GROUP_check() that makes some sanity tests
1772 on a EC_GROUP, its generator and order. This includes
1773 EC_GROUP_check_discriminant().
1774 [Nils Larsch <nla@trustcenter.de>]
1776 *) Add ECDSA in new directory crypto/ecdsa/.
1778 Add applications 'openssl ecparam' and 'openssl ecdsa'
1779 (these are based on 'openssl dsaparam' and 'openssl dsa').
1781 ECDSA support is also included in various other files across the
1782 library. Most notably,
1783 - 'openssl req' now has a '-newkey ecdsa:file' option;
1784 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
1785 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
1786 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
1787 them suitable for ECDSA where domain parameters must be
1788 extracted before the specific public key;
1789 - ECDSA engine support has been added.
1790 [Nils Larsch <nla@trustcenter.de>]
1792 *) Include some named elliptic curves, and add OIDs from X9.62,
1793 SECG, and WAP/WTLS. Each curve can be obtained from the new
1795 EC_GROUP_new_by_curve_name(),
1796 and the list of available named curves can be obtained with
1797 EC_get_builtin_curves().
1798 Also add a 'curve_name' member to EC_GROUP objects, which can be
1800 EC_GROUP_set_curve_name()
1801 EC_GROUP_get_curve_name()
1802 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
1804 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1805 was actually never needed) and in BN_mul(). The removal in BN_mul()
1806 required a small change in bn_mul_part_recursive() and the addition
1807 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1808 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1809 bn_sub_words() and bn_add_words() except they take arrays with
1813 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
1815 *) Cleanse PEM buffers before freeing them since they may contain
1817 [Benjamin Bennett <ben@psc.edu>]
1819 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1820 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1821 authentication-only ciphersuites.
1824 *) Since AES128 and AES256 share a single mask bit in the logic of
1825 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1826 kludge to work properly if AES128 is available and AES256 isn't.
1829 *) Expand security boundary to match 1.1.1 module.
1832 *) Remove redundant features: hash file source, editing of test vectors
1833 modify fipsld to use external fips_premain.c signature.
1836 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
1837 run algorithm test programs.
1840 *) Make algorithm test programs more tolerant of whitespace.
1843 *) Have SSL/TLS server implementation tolerate "mismatched" record
1844 protocol version while receiving ClientHello even if the
1845 ClientHello is fragmented. (The server can't insist on the
1846 particular protocol version it has chosen before the ServerHello
1847 message has informed the client about his choice.)
1850 *) Load error codes if they are not already present instead of using a
1851 static variable. This allows them to be cleanly unloaded and reloaded.
1854 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
1856 *) Introduce limits to prevent malicious keys being able to
1857 cause a denial of service. (CVE-2006-2940)
1858 [Steve Henson, Bodo Moeller]
1860 *) Fix ASN.1 parsing of certain invalid structures that can result
1861 in a denial of service. (CVE-2006-2937) [Steve Henson]
1863 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1864 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1866 *) Fix SSL client code which could crash if connecting to a
1867 malicious SSLv2 server. (CVE-2006-4343)
1868 [Tavis Ormandy and Will Drewry, Google Security Team]
1870 *) Change ciphersuite string processing so that an explicit
1871 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
1872 will no longer include "AES128-SHA"), and any other similar
1873 ciphersuite (same bitmap) from *other* protocol versions (so that
1874 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
1875 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
1876 changes from 0.9.8b and 0.9.8d.
1879 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
1881 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1882 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1884 *) Change the Unix randomness entropy gathering to use poll() when
1885 possible instead of select(), since the latter has some
1886 undesirable limitations.
1887 [Darryl Miles via Richard Levitte and Bodo Moeller]
1889 *) Disable rogue ciphersuites:
1891 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1892 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1893 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1895 The latter two were purportedly from
1896 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1899 Also deactive the remaining ciphersuites from
1900 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1901 unofficial, and the ID has long expired.
1904 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1905 dual-core machines) and other potential thread-safety issues.
1908 Changes between 0.9.7i and 0.9.7j [04 May 2006]
1910 *) Adapt fipsld and the build system to link against the validated FIPS
1911 module in FIPS mode.
1914 *) Fixes for VC++ 2005 build under Windows.
1917 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
1918 from a Windows bash shell such as MSYS. It is autodetected from the
1919 "config" script when run from a VC++ environment. Modify standard VC++
1920 build to use fipscanister.o from the GNU make build.
1923 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
1925 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
1926 The value now differs depending on if you build for FIPS or not.
1927 BEWARE! A program linked with a shared FIPSed libcrypto can't be
1928 safely run with a non-FIPSed libcrypto, as it may crash because of
1929 the difference induced by this change.
1932 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
1934 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1935 (part of SSL_OP_ALL). This option used to disable the
1936 countermeasure against man-in-the-middle protocol-version
1937 rollback in the SSL 2.0 server implementation, which is a bad
1938 idea. (CVE-2005-2969)
1940 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1941 for Information Security, National Institute of Advanced Industrial
1942 Science and Technology [AIST], Japan)]
1944 *) Minimal support for X9.31 signatures and PSS padding modes. This is
1945 mainly for FIPS compliance and not fully integrated at this stage.
1948 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
1949 the exponentiation using a fixed-length exponent. (Otherwise,
1950 the information leaked through timing could expose the secret key
1951 after many signatures; cf. Bleichenbacher's attack on DSA with
1955 *) Make a new fixed-window mod_exp implementation the default for
1956 RSA, DSA, and DH private-key operations so that the sequence of
1957 squares and multiplies and the memory access pattern are
1958 independent of the particular secret key. This will mitigate
1959 cache-timing and potential related attacks.
1961 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
1962 and this is automatically used by BN_mod_exp_mont() if the new flag
1963 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
1964 will use this BN flag for private exponents unless the flag
1965 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
1966 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
1968 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
1970 *) Change the client implementation for SSLv23_method() and
1971 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
1972 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
1973 (Previously, the SSL 2.0 backwards compatible Client Hello
1974 message format would be used even with SSL_OP_NO_SSLv2.)
1977 *) Add support for smime-type MIME parameter in S/MIME messages which some
1981 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
1982 a threadsafe manner. Modify rsa code to use new function and add calls
1983 to dsa and dh code (which had race conditions before).
1986 *) Include the fixed error library code in the C error file definitions
1987 instead of fixing them up at runtime. This keeps the error code
1988 structures constant.
1991 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
1993 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
1996 *) Fixes for newer kerberos headers. NB: the casts are needed because
1997 the 'length' field is signed on one version and unsigned on another
1998 with no (?) obvious way to tell the difference, without these VC++
1999 complains. Also the "definition" of FAR (blank) is no longer included
2000 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
2001 some needed definitions.
2004 *) Undo Cygwin change.
2007 *) Added support for proxy certificates according to RFC 3820.
2008 Because they may be a security thread to unaware applications,
2009 they must be explicitely allowed in run-time. See
2010 docs/HOWTO/proxy_certificates.txt for further information.
2013 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
2015 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
2016 server and client random values. Previously
2017 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
2018 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
2020 This change has negligible security impact because:
2022 1. Server and client random values still have 24 bytes of pseudo random
2025 2. Server and client random values are sent in the clear in the initial
2028 3. The master secret is derived using the premaster secret (48 bytes in
2029 size for static RSA ciphersuites) as well as client server and random
2032 The OpenSSL team would like to thank the UK NISCC for bringing this issue
2035 [Stephen Henson, reported by UK NISCC]
2037 *) Use Windows randomness collection on Cygwin.
2040 *) Fix hang in EGD/PRNGD query when communication socket is closed
2041 prematurely by EGD/PRNGD.
2042 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
2044 *) Prompt for pass phrases when appropriate for PKCS12 input format.
2047 *) Back-port of selected performance improvements from development
2048 branch, as well as improved support for PowerPC platforms.
2051 *) Add lots of checks for memory allocation failure, error codes to indicate
2052 failure and freeing up memory if a failure occurs.
2053 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
2055 *) Add new -passin argument to dgst.
2058 *) Perform some character comparisons of different types in X509_NAME_cmp:
2059 this is needed for some certificates that reencode DNs into UTF8Strings
2060 (in violation of RFC3280) and can't or wont issue name rollover
2064 *) Make an explicit check during certificate validation to see that
2065 the CA setting in each certificate on the chain is correct. As a
2066 side effect always do the following basic checks on extensions,
2067 not just when there's an associated purpose to the check:
2069 - if there is an unhandled critical extension (unless the user
2070 has chosen to ignore this fault)
2071 - if the path length has been exceeded (if one is set at all)
2072 - that certain extensions fit the associated purpose (if one has
2076 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
2078 *) Avoid a race condition when CRLs are checked in a multi threaded
2079 environment. This would happen due to the reordering of the revoked
2080 entries during signature checking and serial number lookup. Now the
2081 encoding is cached and the serial number sort performed under a lock.
2082 Add new STACK function sk_is_sorted().
2085 *) Add Delta CRL to the extension code.
2088 *) Various fixes to s3_pkt.c so alerts are sent properly.
2089 [David Holmes <d.holmes@f5.com>]
2091 *) Reduce the chances of duplicate issuer name and serial numbers (in
2092 violation of RFC3280) using the OpenSSL certificate creation utilities.
2093 This is done by creating a random 64 bit value for the initial serial
2094 number when a serial number file is created or when a self signed
2095 certificate is created using 'openssl req -x509'. The initial serial
2096 number file is created using 'openssl x509 -next_serial' in CA.pl
2097 rather than being initialized to 1.
2100 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
2102 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
2103 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
2104 [Joe Orton, Steve Henson]
2106 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
2108 [Joe Orton, Steve Henson]
2110 *) Make it possible to have multiple active certificates with the same
2111 subject in the CA index file. This is done only if the keyword
2112 'unique_subject' is set to 'no' in the main CA section (default
2113 if 'CA_default') of the configuration file. The value is saved
2114 with the database itself in a separate index attribute file,
2115 named like the index file with '.attr' appended to the name.
2118 *) X509 verify fixes. Disable broken certificate workarounds when
2119 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
2120 keyUsage extension present. Don't accept CRLs with unhandled critical
2121 extensions: since verify currently doesn't process CRL extensions this
2122 rejects a CRL with *any* critical extensions. Add new verify error codes
2126 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
2127 A clarification of RFC2560 will require the use of OCTET STRINGs and
2128 some implementations cannot handle the current raw format. Since OpenSSL
2129 copies and compares OCSP nonces as opaque blobs without any attempt at
2130 parsing them this should not create any compatibility issues.
2133 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
2134 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
2135 this HMAC (and other) operations are several times slower than OpenSSL
2139 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
2140 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2142 *) Use the correct content when signing type "other".
2145 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
2147 *) Fix various bugs revealed by running the NISCC test suite:
2149 Stop out of bounds reads in the ASN1 code when presented with
2150 invalid tags (CVE-2003-0543 and CVE-2003-0544).
2152 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
2154 If verify callback ignores invalid public key errors don't try to check
2155 certificate signature with the NULL public key.
2159 *) New -ignore_err option in ocsp application to stop the server
2160 exiting on the first error in a request.
2163 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
2164 if the server requested one: as stated in TLS 1.0 and SSL 3.0
2168 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
2169 extra data after the compression methods not only for TLS 1.0
2170 but also for SSL 3.0 (as required by the specification).
2171 [Bodo Moeller; problem pointed out by Matthias Loepfe]
2173 *) Change X509_certificate_type() to mark the key as exported/exportable
2174 when it's 512 *bits* long, not 512 bytes.
2177 *) Change AES_cbc_encrypt() so it outputs exact multiple of
2178 blocks during encryption.
2181 *) Various fixes to base64 BIO and non blocking I/O. On write
2182 flushes were not handled properly if the BIO retried. On read
2183 data was not being buffered properly and had various logic bugs.
2184 This also affects blocking I/O when the data being decoded is a
2188 *) Various S/MIME bugfixes and compatibility changes:
2189 output correct application/pkcs7 MIME type if
2190 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
2191 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
2192 of files as .eml work). Correctly handle very long lines in MIME
2196 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
2198 *) Countermeasure against the Klima-Pokorny-Rosa extension of
2199 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
2200 a protocol version number mismatch like a decryption error
2201 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
2204 *) Turn on RSA blinding by default in the default implementation
2205 to avoid a timing attack. Applications that don't want it can call
2206 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
2207 They would be ill-advised to do so in most cases.
2208 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
2210 *) Change RSA blinding code so that it works when the PRNG is not
2211 seeded (in this case, the secret RSA exponent is abused as
2212 an unpredictable seed -- if it is not unpredictable, there
2213 is no point in blinding anyway). Make RSA blinding thread-safe
2214 by remembering the creator's thread ID in rsa->blinding and
2215 having all other threads use local one-time blinding factors
2216 (this requires more computation than sharing rsa->blinding, but
2217 avoids excessive locking; and if an RSA object is not shared
2218 between threads, blinding will still be very fast).
2221 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
2222 ENGINE as defaults for all supported algorithms irrespective of
2223 the 'flags' parameter. 'flags' is now honoured, so applications
2224 should make sure they are passing it correctly.
2227 *) Target "mingw" now allows native Windows code to be generated in
2228 the Cygwin environment as well as with the MinGW compiler.
2231 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
2233 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
2234 via timing by performing a MAC computation even if incorrrect
2235 block cipher padding has been found. This is a countermeasure
2236 against active attacks where the attacker has to distinguish
2237 between bad padding and a MAC verification error. (CVE-2003-0078)
2239 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
2240 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
2241 Martin Vuagnoux (EPFL, Ilion)]
2243 *) Make the no-err option work as intended. The intention with no-err
2244 is not to have the whole error stack handling routines removed from
2245 libcrypto, it's only intended to remove all the function name and
2246 reason texts, thereby removing some of the footprint that may not
2247 be interesting if those errors aren't displayed anyway.
2249 NOTE: it's still possible for any application or module to have it's
2250 own set of error texts inserted. The routines are there, just not
2251 used by default when no-err is given.
2254 *) Add support for FreeBSD on IA64.
2255 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
2257 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
2258 Kerberos function mit_des_cbc_cksum(). Before this change,
2259 the value returned by DES_cbc_cksum() was like the one from
2260 mit_des_cbc_cksum(), except the bytes were swapped.
2261 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
2263 *) Allow an application to disable the automatic SSL chain building.
2264 Before this a rather primitive chain build was always performed in
2265 ssl3_output_cert_chain(): an application had no way to send the
2266 correct chain if the automatic operation produced an incorrect result.
2268 Now the chain builder is disabled if either:
2270 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
2272 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
2274 The reasoning behind this is that an application would not want the
2275 auto chain building to take place if extra chain certificates are
2276 present and it might also want a means of sending no additional
2277 certificates (for example the chain has two certificates and the
2281 *) Add the possibility to build without the ENGINE framework.
2282 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2284 *) Under Win32 gmtime() can return NULL: check return value in
2285 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
2288 *) DSA routines: under certain error conditions uninitialized BN objects
2289 could be freed. Solution: make sure initialization is performed early
2290 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
2291 Nils Larsch <nla@trustcenter.de> via PR#459)
2294 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
2295 checked on reconnect on the client side, therefore session resumption
2296 could still fail with a "ssl session id is different" error. This
2297 behaviour is masked when SSL_OP_ALL is used due to
2298 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
2299 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2300 followup to PR #377.
2303 *) IA-32 assembler support enhancements: unified ELF targets, support
2304 for SCO/Caldera platforms, fix for Cygwin shared build.
2307 *) Add support for FreeBSD on sparc64. As a consequence, support for
2308 FreeBSD on non-x86 processors is separate from x86 processors on
2309 the config script, much like the NetBSD support.
2310 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
2312 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
2314 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
2317 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
2318 code (06) was taken as the first octet of the session ID and the last
2319 octet was ignored consequently. As a result SSLv2 client side session
2320 caching could not have worked due to the session ID mismatch between
2322 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2326 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
2327 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
2331 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
2332 seems that in spite of existing for more than a year, many application
2333 author have done nothing to provide the necessary callbacks, which
2334 means that this particular engine will not work properly anywhere.
2335 This is a very unfortunate situation which forces us, in the name
2336 of usability, to give the hw_ncipher.c a static lock, which is part
2338 NOTE: This is for the 0.9.7 series ONLY. This hack will never
2339 appear in 0.9.8 or later. We EXPECT application authors to have
2340 dealt properly with this when 0.9.8 is released (unless we actually
2341 make such changes in the libcrypto locking code that changes will
2342 have to be made anyway).
2345 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
2346 octets have been read, EOF or an error occurs. Without this change
2347 some truncated ASN1 structures will not produce an error.
2350 *) Disable Heimdal support, since it hasn't been fully implemented.
2351 Still give the possibility to force the use of Heimdal, but with
2352 warnings and a request that patches get sent to openssl-dev.
2355 *) Add the VC-CE target, introduce the WINCE sysname, and add
2356 INSTALL.WCE and appropriate conditionals to make it build.
2357 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2359 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
2360 cygssl-x.y.z.dll, where x, y and z are the major, minor and
2361 edit numbers of the version.
2362 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2364 *) Introduce safe string copy and catenation functions
2365 (BUF_strlcpy() and BUF_strlcat()).
2366 [Ben Laurie (CHATS) and Richard Levitte]
2368 *) Avoid using fixed-size buffers for one-line DNs.
2369 [Ben Laurie (CHATS)]
2371 *) Add BUF_MEM_grow_clean() to avoid information leakage when
2372 resizing buffers containing secrets, and use where appropriate.
2373 [Ben Laurie (CHATS)]
2375 *) Avoid using fixed size buffers for configuration file location.
2376 [Ben Laurie (CHATS)]
2378 *) Avoid filename truncation for various CA files.
2379 [Ben Laurie (CHATS)]
2381 *) Use sizeof in preference to magic numbers.
2382 [Ben Laurie (CHATS)]
2384 *) Avoid filename truncation in cert requests.
2385 [Ben Laurie (CHATS)]
2387 *) Add assertions to check for (supposedly impossible) buffer
2389 [Ben Laurie (CHATS)]
2391 *) Don't cache truncated DNS entries in the local cache (this could
2392 potentially lead to a spoofing attack).
2393 [Ben Laurie (CHATS)]
2395 *) Fix various buffers to be large enough for hex/decimal
2396 representations in a platform independent manner.
2397 [Ben Laurie (CHATS)]
2399 *) Add CRYPTO_realloc_clean() to avoid information leakage when
2400 resizing buffers containing secrets, and use where appropriate.
2401 [Ben Laurie (CHATS)]
2403 *) Add BIO_indent() to avoid much slightly worrying code to do
2405 [Ben Laurie (CHATS)]
2407 *) Convert sprintf()/BIO_puts() to BIO_printf().
2408 [Ben Laurie (CHATS)]
2410 *) buffer_gets() could terminate with the buffer only half
2412 [Ben Laurie (CHATS)]
2414 *) Add assertions to prevent user-supplied crypto functions from
2415 overflowing internal buffers by having large block sizes, etc.
2416 [Ben Laurie (CHATS)]
2418 *) New OPENSSL_assert() macro (similar to assert(), but enabled
2420 [Ben Laurie (CHATS)]
2422 *) Eliminate unused copy of key in RC4.
2423 [Ben Laurie (CHATS)]
2425 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
2426 [Ben Laurie (CHATS)]
2428 *) Fix off-by-one error in EGD path.
2429 [Ben Laurie (CHATS)]
2431 *) If RANDFILE path is too long, ignore instead of truncating.
2432 [Ben Laurie (CHATS)]
2434 *) Eliminate unused and incorrectly sized X.509 structure
2436 [Ben Laurie (CHATS)]
2438 *) Eliminate unused and dangerous function knumber().
2439 [Ben Laurie (CHATS)]
2441 *) Eliminate unused and dangerous structure, KSSL_ERR.
2442 [Ben Laurie (CHATS)]
2444 *) Protect against overlong session ID context length in an encoded
2445 session object. Since these are local, this does not appear to be
2447 [Ben Laurie (CHATS)]
2449 *) Change from security patch (see 0.9.6e below) that did not affect
2450 the 0.9.6 release series:
2452 Remote buffer overflow in SSL3 protocol - an attacker could
2453 supply an oversized master key in Kerberos-enabled versions.
2455 [Ben Laurie (CHATS)]
2457 *) Change the SSL kerb5 codes to match RFC 2712.
2460 *) Make -nameopt work fully for req and add -reqopt switch.
2461 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
2463 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
2464 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
2466 *) Make sure tests can be performed even if the corresponding algorithms
2467 have been removed entirely. This was also the last step to make
2468 OpenSSL compilable with DJGPP under all reasonable conditions.
2469 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
2471 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
2472 to allow version independent disabling of normally unselected ciphers,
2473 which may be activated as a side-effect of selecting a single cipher.
2475 (E.g., cipher list string "RSA" enables ciphersuites that are left
2476 out of "ALL" because they do not provide symmetric encryption.
2477 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
2478 [Lutz Jaenicke, Bodo Moeller]
2480 *) Add appropriate support for separate platform-dependent build
2481 directories. The recommended way to make a platform-dependent
2482 build directory is the following (tested on Linux), maybe with
2485 # Place yourself outside of the OpenSSL source tree. In
2486 # this example, the environment variable OPENSSL_SOURCE
2487 # is assumed to contain the absolute OpenSSL source directory.
2488 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
2489 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
2490 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
2491 mkdir -p `dirname $F`
2492 ln -s $OPENSSL_SOURCE/$F $F
2495 To be absolutely sure not to disturb the source tree, a "make clean"
2496 is a good thing. If it isn't successfull, don't worry about it,
2497 it probably means the source directory is very clean.
2500 *) Make sure any ENGINE control commands make local copies of string
2501 pointers passed to them whenever necessary. Otherwise it is possible
2502 the caller may have overwritten (or deallocated) the original string
2503 data when a later ENGINE operation tries to use the stored values.
2504 [Götz Babin-Ebell <babinebell@trustcenter.de>]
2506 *) Improve diagnostics in file reading and command-line digests.
2507 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
2509 *) Add AES modes CFB and OFB to the object database. Correct an
2510 error in AES-CFB decryption.
2513 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
2514 allows existing EVP_CIPHER_CTX structures to be reused after
2515 calling EVP_*Final(). This behaviour is used by encryption
2516 BIOs and some applications. This has the side effect that
2517 applications must explicitly clean up cipher contexts with
2518 EVP_CIPHER_CTX_cleanup() or they will leak memory.
2521 *) Check the values of dna and dnb in bn_mul_recursive before calling
2522 bn_mul_comba (a non zero value means the a or b arrays do not contain
2523 n2 elements) and fallback to bn_mul_normal if either is not zero.
2526 *) Fix escaping of non-ASCII characters when using the -subj option
2527 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
2530 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
2531 form for "surname", serialNumber has no short form.
2532 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
2533 therefore remove "mail" short name for "internet 7".
2534 The OID for unique identifiers in X509 certificates is
2535 x500UniqueIdentifier, not uniqueIdentifier.
2536 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
2539 *) Add an "init" command to the ENGINE config module and auto initialize
2540 ENGINEs. Without any "init" command the ENGINE will be initialized
2541 after all ctrl commands have been executed on it. If init=1 the
2542 ENGINE is initailized at that point (ctrls before that point are run
2543 on the uninitialized ENGINE and after on the initialized one). If
2544 init=0 then the ENGINE will not be iniatialized at all.
2547 *) Fix the 'app_verify_callback' interface so that the user-defined
2548 argument is actually passed to the callback: In the
2549 SSL_CTX_set_cert_verify_callback() prototype, the callback
2550 declaration has been changed from
2553 int (*cb)(X509_STORE_CTX *,void *);
2554 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
2555 i=s->ctx->app_verify_callback(&ctx)
2556 has been changed into
2557 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
2559 To update applications using SSL_CTX_set_cert_verify_callback(),
2560 a dummy argument can be added to their callback functions.
2561 [D. K. Smetters <smetters@parc.xerox.com>]
2563 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
2564 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
2566 *) Add and OPENSSL_LOAD_CONF define which will cause
2567 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
2568 This allows older applications to transparently support certain
2569 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
2570 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
2571 load the config file and OPENSSL_add_all_algorithms_conf() which will
2572 always load it have also been added.
2575 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
2576 Adjust NIDs and EVP layer.
2577 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2579 *) Config modules support in openssl utility.
2581 Most commands now load modules from the config file,
2582 though in a few (such as version) this isn't done
2583 because it couldn't be used for anything.
2585 In the case of ca and req the config file used is
2586 the same as the utility itself: that is the -config
2587 command line option can be used to specify an
2591 *) Move default behaviour from OPENSSL_config(). If appname is NULL
2592 use "openssl_conf" if filename is NULL use default openssl config file.
2595 *) Add an argument to OPENSSL_config() to allow the use of an alternative
2596 config section name. Add a new flag to tolerate a missing config file
2597 and move code to CONF_modules_load_file().
2600 *) Support for crypto accelerator cards from Accelerated Encryption
2601 Processing, www.aep.ie. (Use engine 'aep')
2602 The support was copied from 0.9.6c [engine] and adapted/corrected
2603 to work with the new engine framework.
2604 [AEP Inc. and Richard Levitte]
2606 *) Support for SureWare crypto accelerator cards from Baltimore
2607 Technologies. (Use engine 'sureware')
2608 The support was copied from 0.9.6c [engine] and adapted
2609 to work with the new engine framework.
2612 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
2613 make the newer ENGINE framework commands for the CHIL engine work.
2614 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
2616 *) Make it possible to produce shared libraries on ReliantUNIX.
2617 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
2619 *) Add the configuration target debug-linux-ppro.
2620 Make 'openssl rsa' use the general key loading routines
2621 implemented in apps.c, and make those routines able to
2622 handle the key format FORMAT_NETSCAPE and the variant
2624 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2626 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2627 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2629 *) Add -keyform to rsautl, and document -engine.
2630 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
2632 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
2633 BIO_R_NO_SUCH_FILE error code rather than the generic
2634 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
2637 *) Add new functions
2639 ERR_peek_last_error_line
2640 ERR_peek_last_error_line_data.
2641 These are similar to
2644 ERR_peek_error_line_data,
2645 but report on the latest error recorded rather than the first one
2646 still in the error queue.
2647 [Ben Laurie, Bodo Moeller]
2649 *) default_algorithms option in ENGINE config module. This allows things
2651 default_algorithms = ALL
2652 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
2655 *) Prelminary ENGINE config module.
2658 *) New experimental application configuration code.
2661 *) Change the AES code to follow the same name structure as all other
2662 symmetric ciphers, and behave the same way. Move everything to
2663 the directory crypto/aes, thereby obsoleting crypto/rijndael.
2664 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2666 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
2667 [Ben Laurie and Theo de Raadt]
2669 *) Add option to output public keys in req command.
2670 [Massimiliano Pala madwolf@openca.org]
2672 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
2673 (up to about 10% better than before for P-192 and P-224).
2676 *) New functions/macros
2678 SSL_CTX_set_msg_callback(ctx, cb)
2679 SSL_CTX_set_msg_callback_arg(ctx, arg)
2680 SSL_set_msg_callback(ssl, cb)
2681 SSL_set_msg_callback_arg(ssl, arg)
2683 to request calling a callback function
2685 void cb(int write_p, int version, int content_type,
2686 const void *buf, size_t len, SSL *ssl, void *arg)
2688 whenever a protocol message has been completely received
2689 (write_p == 0) or sent (write_p == 1). Here 'version' is the
2690 protocol version according to which the SSL library interprets
2691 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
2692 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
2693 the content type as defined in the SSL 3.0/TLS 1.0 protocol
2694 specification (change_cipher_spec(20), alert(21), handshake(22)).
2695 'buf' and 'len' point to the actual message, 'ssl' to the
2696 SSL object, and 'arg' is the application-defined value set by
2697 SSL[_CTX]_set_msg_callback_arg().
2699 'openssl s_client' and 'openssl s_server' have new '-msg' options
2700 to enable a callback that displays all protocol messages.
2703 *) Change the shared library support so shared libraries are built as
2704 soon as the corresponding static library is finished, and thereby get
2705 openssl and the test programs linked against the shared library.
2706 This still only happens when the keyword "shard" has been given to
2707 the configuration scripts.
2709 NOTE: shared library support is still an experimental thing, and
2710 backward binary compatibility is still not guaranteed.
2711 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
2713 *) Add support for Subject Information Access extension.
2714 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2716 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
2717 additional bytes when new memory had to be allocated, not just
2718 when reusing an existing buffer.
2721 *) New command line and configuration option 'utf8' for the req command.
2722 This allows field values to be specified as UTF8 strings.
2725 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
2726 runs for the former and machine-readable output for the latter.
2729 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
2730 of the e-mail address in the DN (i.e., it will go into a certificate
2731 extension only). The new configuration file option 'email_in_dn = no'
2732 has the same effect.
2733 [Massimiliano Pala madwolf@openca.org]
2735 *) Change all functions with names starting with des_ to be starting
2736 with DES_ instead. Add wrappers that are compatible with libdes,
2737 but are named _ossl_old_des_*. Finally, add macros that map the
2738 des_* symbols to the corresponding _ossl_old_des_* if libdes
2739 compatibility is desired. If OpenSSL 0.9.6c compatibility is
2740 desired, the des_* symbols will be mapped to DES_*, with one
2743 Since we provide two compatibility mappings, the user needs to
2744 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
2745 compatibility is desired. The default (i.e., when that macro
2746 isn't defined) is OpenSSL 0.9.6c compatibility.
2748 There are also macros that enable and disable the support of old
2749 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
2750 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
2751 are defined, the default will apply: to support the old des routines.
2753 In either case, one must include openssl/des.h to get the correct
2754 definitions. Do not try to just include openssl/des_old.h, that
2757 NOTE: This is a major break of an old API into a new one. Software
2758 authors are encouraged to switch to the DES_ style functions. Some
2759 time in the future, des_old.h and the libdes compatibility functions
2760 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
2761 default), and then completely removed.
2764 *) Test for certificates which contain unsupported critical extensions.
2765 If such a certificate is found during a verify operation it is
2766 rejected by default: this behaviour can be overridden by either
2767 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
2768 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
2769 X509_supported_extension() has also been added which returns 1 if a
2770 particular extension is supported.
2773 *) Modify the behaviour of EVP cipher functions in similar way to digests
2774 to retain compatibility with existing code.
2777 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
2778 compatibility with existing code. In particular the 'ctx' parameter does
2779 not have to be to be initialized before the call to EVP_DigestInit() and
2780 it is tidied up after a call to EVP_DigestFinal(). New function
2781 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
2782 EVP_MD_CTX_copy() changed to not require the destination to be
2783 initialized valid and new function EVP_MD_CTX_copy_ex() added which
2784 requires the destination to be valid.
2786 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
2787 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
2790 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
2791 so that complete 'Handshake' protocol structures are kept in memory
2792 instead of overwriting 'msg_type' and 'length' with 'body' data.
2795 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
2796 [Massimo Santin via Richard Levitte]
2798 *) Major restructuring to the underlying ENGINE code. This includes
2799 reduction of linker bloat, separation of pure "ENGINE" manipulation
2800 (initialisation, etc) from functionality dealing with implementations
2801 of specific crypto iterfaces. This change also introduces integrated
2802 support for symmetric ciphers and digest implementations - so ENGINEs
2803 can now accelerate these by providing EVP_CIPHER and EVP_MD
2804 implementations of their own. This is detailed in crypto/engine/README
2805 as it couldn't be adequately described here. However, there are a few
2806 API changes worth noting - some RSA, DSA, DH, and RAND functions that
2807 were changed in the original introduction of ENGINE code have now
2808 reverted back - the hooking from this code to ENGINE is now a good
2809 deal more passive and at run-time, operations deal directly with
2810 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
2811 dereferencing through an ENGINE pointer any more. Also, the ENGINE
2812 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
2813 they were not being used by the framework as there is no concept of a
2814 BIGNUM_METHOD and they could not be generalised to the new
2815 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
2816 ENGINE_cpy() has been removed as it cannot be consistently defined in
2820 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
2823 *) Change mkdef.pl to sort symbols that get the same entry number,
2824 and make sure the automatically generated functions ERR_load_*
2825 become part of libeay.num as well.
2828 *) New function SSL_renegotiate_pending(). This returns true once
2829 renegotiation has been requested (either SSL_renegotiate() call
2830 or HelloRequest/ClientHello receveived from the peer) and becomes
2831 false once a handshake has been completed.
2832 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
2833 sends a HelloRequest, but does not ensure that a handshake takes
2834 place. SSL_renegotiate_pending() is useful for checking if the
2835 client has followed the request.)
2838 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
2839 By default, clients may request session resumption even during
2840 renegotiation (if session ID contexts permit); with this option,
2841 session resumption is possible only in the first handshake.
2843 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
2844 more bits available for options that should not be part of
2845 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
2848 *) Add some demos for certificate and certificate request creation.
2851 *) Make maximum certificate chain size accepted from the peer application
2852 settable (SSL*_get/set_max_cert_list()), as proposed by
2853 "Douglas E. Engert" <deengert@anl.gov>.
2856 *) Add support for shared libraries for Unixware-7
2857 (Boyd Lynn Gerber <gerberb@zenez.com>).
2860 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
2861 be done prior to destruction. Use this to unload error strings from
2862 ENGINEs that load their own error strings. NB: This adds two new API
2863 functions to "get" and "set" this destroy handler in an ENGINE.
2866 *) Alter all existing ENGINE implementations (except "openssl" and
2867 "openbsd") to dynamically instantiate their own error strings. This
2868 makes them more flexible to be built both as statically-linked ENGINEs
2869 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
2870 Also, add stub code to each that makes building them as self-contained
2871 shared-libraries easier (see README.ENGINE).
2874 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
2875 implementations into applications that are completely implemented in
2876 self-contained shared-libraries. The "dynamic" ENGINE exposes control
2877 commands that can be used to configure what shared-library to load and
2878 to control aspects of the way it is handled. Also, made an update to
2879 the README.ENGINE file that brings its information up-to-date and
2880 provides some information and instructions on the "dynamic" ENGINE
2881 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
2884 *) Make it possible to unload ranges of ERR strings with a new
2885 "ERR_unload_strings" function.
2888 *) Add a copy() function to EVP_MD.
2891 *) Make EVP_MD routines take a context pointer instead of just the
2892 md_data void pointer.
2895 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
2896 that the digest can only process a single chunk of data
2897 (typically because it is provided by a piece of
2898 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
2899 is only going to provide a single chunk of data, and hence the
2900 framework needn't accumulate the data for oneshot drivers.
2903 *) As with "ERR", make it possible to replace the underlying "ex_data"
2904 functions. This change also alters the storage and management of global
2905 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
2906 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
2907 index counters. The API functions that use this state have been changed
2908 to take a "class_index" rather than pointers to the class's local STACK
2909 and counter, and there is now an API function to dynamically create new
2910 classes. This centralisation allows us to (a) plug a lot of the
2911 thread-safety problems that existed, and (b) makes it possible to clean
2912 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
2913 such data would previously have always leaked in application code and
2914 workarounds were in place to make the memory debugging turn a blind eye
2915 to it. Application code that doesn't use this new function will still
2916 leak as before, but their memory debugging output will announce it now
2917 rather than letting it slide.
2919 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
2920 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
2921 has a return value to indicate success or failure.
2924 *) Make it possible to replace the underlying "ERR" functions such that the
2925 global state (2 LHASH tables and 2 locks) is only used by the "default"
2926 implementation. This change also adds two functions to "get" and "set"
2927 the implementation prior to it being automatically set the first time
2928 any other ERR function takes place. Ie. an application can call "get",
2929 pass the return value to a module it has just loaded, and that module
2930 can call its own "set" function using that value. This means the
2931 module's "ERR" operations will use (and modify) the error state in the
2932 application and not in its own statically linked copy of OpenSSL code.
2935 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
2936 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
2937 the operation, and provides a more encapsulated way for external code
2938 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
2939 to use these functions rather than manually incrementing the counts.
2941 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
2944 *) Add EVP test program.
2947 *) Add symmetric cipher support to ENGINE. Expect the API to change!
2950 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
2951 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
2952 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
2953 These allow a CRL to be built without having to access X509_CRL fields
2954 directly. Modify 'ca' application to use new functions.
2957 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
2958 bug workarounds. Rollback attack detection is a security feature.
2959 The problem will only arise on OpenSSL servers when TLSv1 is not
2960 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
2961 Software authors not wanting to support TLSv1 will have special reasons
2962 for their choice and can explicitly enable this option.
2963 [Bodo Moeller, Lutz Jaenicke]
2965 *) Rationalise EVP so it can be extended: don't include a union of
2966 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
2967 (similar to those existing for EVP_CIPHER_CTX).
2972 EVP_MD_CTX_init(&md); /* new function call */
2973 EVP_DigestInit(&md, EVP_sha1());
2974 EVP_DigestUpdate(&md, in, len);
2975 EVP_DigestFinal(&md, out, NULL);
2976 EVP_MD_CTX_cleanup(&md); /* new function call */
2980 *) Make DES key schedule conform to the usual scheme, as well as
2981 correcting its structure. This means that calls to DES functions
2982 now have to pass a pointer to a des_key_schedule instead of a
2983 plain des_key_schedule (which was actually always a pointer
2986 des_key_schedule ks;
2988 des_set_key_checked(..., &ks);
2989 des_ncbc_encrypt(..., &ks, ...);
2991 (Note that a later change renames 'des_...' into 'DES_...'.)
2994 *) Initial reduction of linker bloat: the use of some functions, such as
2995 PEM causes large amounts of unused functions to be linked in due to
2996 poor organisation. For example pem_all.c contains every PEM function
2997 which has a knock on effect of linking in large amounts of (unused)
2998 ASN1 code. Grouping together similar functions and splitting unrelated
2999 functions prevents this.
3002 *) Cleanup of EVP macros.
3005 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
3006 correct _ecb suffix.
3009 *) Add initial OCSP responder support to ocsp application. The
3010 revocation information is handled using the text based index
3011 use by the ca application. The responder can either handle
3012 requests generated internally, supplied in files (for example
3013 via a CGI script) or using an internal minimal server.
3016 *) Add configuration choices to get zlib compression for TLS.
3019 *) Changes to Kerberos SSL for RFC 2712 compliance:
3020 1. Implemented real KerberosWrapper, instead of just using
3021 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
3022 2. Implemented optional authenticator field of KerberosWrapper.
3024 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
3025 and authenticator structs; see crypto/krb5/.
3027 Generalized Kerberos calls to support multiple Kerberos libraries.
3028 [Vern Staats <staatsvr@asc.hpc.mil>,
3029 Jeffrey Altman <jaltman@columbia.edu>
3030 via Richard Levitte]
3032 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
3033 already does with RSA. testdsa.h now has 'priv_key/pub_key'
3034 values for each of the key sizes rather than having just
3035 parameters (and 'speed' generating keys each time).
3038 *) Speed up EVP routines.
3041 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
3042 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
3043 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
3044 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
3046 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
3047 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
3048 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
3051 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
3053 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
3056 *) Added the OS2-EMX target.
3057 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
3059 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
3060 to support NCONF routines in extension code. New function CONF_set_nconf()
3061 to allow functions which take an NCONF to also handle the old LHASH
3062 structure: this means that the old CONF compatible routines can be
3063 retained (in particular wrt extensions) without having to duplicate the
3064 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
3067 *) Enhance the general user interface with mechanisms for inner control
3068 and with possibilities to have yes/no kind of prompts.
3071 *) Change all calls to low level digest routines in the library and
3072 applications to use EVP. Add missing calls to HMAC_cleanup() and
3073 don't assume HMAC_CTX can be copied using memcpy().
3074 [Verdon Walker <VWalker@novell.com>, Steve Henson]
3076 *) Add the possibility to control engines through control names but with
3077 arbitrary arguments instead of just a string.
3078 Change the key loaders to take a UI_METHOD instead of a callback
3079 function pointer. NOTE: this breaks binary compatibility with earlier
3080 versions of OpenSSL [engine].
3081 Adapt the nCipher code for these new conditions and add a card insertion
3085 *) Enhance the general user interface with mechanisms to better support
3086 dialog box interfaces, application-defined prompts, the possibility
3087 to use defaults (for example default passwords from somewhere else)
3088 and interrupts/cancellations.
3091 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
3092 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
3095 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
3096 tidy up some unnecessarily weird code in 'sk_new()').
3097 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
3099 *) Change the key loading routines for ENGINEs to use the same kind
3100 callback (pem_password_cb) as all other routines that need this
3104 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
3105 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
3106 than this minimum value is recommended.
3109 *) New random seeder for OpenVMS, using the system process statistics
3110 that are easily reachable.
3113 *) Windows apparently can't transparently handle global
3114 variables defined in DLLs. Initialisations such as:
3116 const ASN1_ITEM *it = &ASN1_INTEGER_it;
3118 wont compile. This is used by the any applications that need to
3119 declare their own ASN1 modules. This was fixed by adding the option
3120 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
3121 needed for static libraries under Win32.
3124 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
3125 setting of purpose and trust fields. New X509_STORE trust and
3126 purpose functions and tidy up setting in other SSL functions.
3129 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
3130 structure. These are inherited by X509_STORE_CTX when it is
3131 initialised. This allows various defaults to be set in the
3132 X509_STORE structure (such as flags for CRL checking and custom
3133 purpose or trust settings) for functions which only use X509_STORE_CTX
3134 internally such as S/MIME.
3136 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
3137 trust settings if they are not set in X509_STORE. This allows X509_STORE
3138 purposes and trust (in S/MIME for example) to override any set by default.
3140 Add command line options for CRL checking to smime, s_client and s_server
3144 *) Initial CRL based revocation checking. If the CRL checking flag(s)
3145 are set then the CRL is looked up in the X509_STORE structure and
3146 its validity and signature checked, then if the certificate is found
3147 in the CRL the verify fails with a revoked error.
3149 Various new CRL related callbacks added to X509_STORE_CTX structure.
3151 Command line options added to 'verify' application to support this.
3153 This needs some additional work, such as being able to handle multiple
3154 CRLs with different times, extension based lookup (rather than just
3155 by subject name) and ultimately more complete V2 CRL extension
3159 *) Add a general user interface API (crypto/ui/). This is designed
3160 to replace things like des_read_password and friends (backward
3161 compatibility functions using this new API are provided).
3162 The purpose is to remove prompting functions from the DES code
3163 section as well as provide for prompting through dialog boxes in
3164 a window system and the like.
3167 *) Add "ex_data" support to ENGINE so implementations can add state at a
3168 per-structure level rather than having to store it globally.
3171 *) Make it possible for ENGINE structures to be copied when retrieved by
3172 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
3173 This causes the "original" ENGINE structure to act like a template,
3174 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
3175 operational state can be localised to each ENGINE structure, despite the
3176 fact they all share the same "methods". New ENGINE structures returned in
3177 this case have no functional references and the return value is the single
3178 structural reference. This matches the single structural reference returned
3179 by ENGINE_by_id() normally, when it is incremented on the pre-existing
3183 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
3184 needs to match any other type at all we need to manually clear the
3188 *) Changes to the "openssl engine" utility to include;
3189 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
3190 about an ENGINE's available control commands.
3191 - executing control commands from command line arguments using the
3192 '-pre' and '-post' switches. '-post' is only used if '-t' is
3193 specified and the ENGINE is successfully initialised. The syntax for
3194 the individual commands are colon-separated, for example;
3195 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
3198 *) New dynamic control command support for ENGINEs. ENGINEs can now
3199 declare their own commands (numbers), names (strings), descriptions,
3200 and input types for run-time discovery by calling applications. A
3201 subset of these commands are implicitly classed as "executable"
3202 depending on their input type, and only these can be invoked through
3203 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
3204 can be based on user input, config files, etc). The distinction is
3205 that "executable" commands cannot return anything other than a boolean
3206 result and can only support numeric or string input, whereas some
3207 discoverable commands may only be for direct use through
3208 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
3209 pointers, or other custom uses. The "executable" commands are to
3210 support parameterisations of ENGINE behaviour that can be
3211 unambiguously defined by ENGINEs and used consistently across any
3212 OpenSSL-based application. Commands have been added to all the
3213 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
3214 control over shared-library paths without source code alterations.
3217 *) Changed all ENGINE implementations to dynamically allocate their
3218 ENGINEs rather than declaring them statically. Apart from this being
3219 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
3220 this also allows the implementations to compile without using the
3221 internal engine_int.h header.
3224 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
3225 'const' value. Any code that should be able to modify a RAND_METHOD
3226 should already have non-const pointers to it (ie. they should only
3227 modify their own ones).
3230 *) Made a variety of little tweaks to the ENGINE code.
3231 - "atalla" and "ubsec" string definitions were moved from header files
3232 to C code. "nuron" string definitions were placed in variables
3233 rather than hard-coded - allowing parameterisation of these values
3234 later on via ctrl() commands.
3235 - Removed unused "#if 0"'d code.
3236 - Fixed engine list iteration code so it uses ENGINE_free() to release
3237 structural references.
3238 - Constified the RAND_METHOD element of ENGINE structures.
3239 - Constified various get/set functions as appropriate and added
3240 missing functions (including a catch-all ENGINE_cpy that duplicates
3241 all ENGINE values onto a new ENGINE except reference counts/state).
3242 - Removed NULL parameter checks in get/set functions. Setting a method
3243 or function to NULL is a way of cancelling out a previously set
3244 value. Passing a NULL ENGINE parameter is just plain stupid anyway
3245 and doesn't justify the extra error symbols and code.
3246 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
3247 flags from engine_int.h to engine.h.
3248 - Changed prototypes for ENGINE handler functions (init(), finish(),
3249 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
3252 *) Implement binary inversion algorithm for BN_mod_inverse in addition
3253 to the algorithm using long division. The binary algorithm can be
3254 used only if the modulus is odd. On 32-bit systems, it is faster
3255 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
3256 roughly 5-15% for 256-bit moduli), so we use it only for moduli
3257 up to 450 bits. In 64-bit environments, the binary algorithm
3258 appears to be advantageous for much longer moduli; here we use it
3259 for moduli up to 2048 bits.
3262 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
3263 could not support the combine flag in choice fields.
3266 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
3267 extensions from a certificate request to the certificate.
3270 *) Allow multiple 'certopt' and 'nameopt' options to be separated
3271 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
3272 file: this allows the display of the certificate about to be
3273 signed to be customised, to allow certain fields to be included
3274 or excluded and extension details. The old system didn't display
3275 multicharacter strings properly, omitted fields not in the policy
3276 and couldn't display additional details such as extensions.
3279 *) Function EC_POINTs_mul for multiple scalar multiplication
3280 of an arbitrary number of elliptic curve points
3281 \sum scalars[i]*points[i],
3282 optionally including the generator defined for the EC_GROUP:
3283 scalar*generator + \sum scalars[i]*points[i].
3285 EC_POINT_mul is a simple wrapper function for the typical case
3286 that the point list has just one item (besides the optional
3290 *) First EC_METHODs for curves over GF(p):
3292 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
3293 operations and provides various method functions that can also
3294 operate with faster implementations of modular arithmetic.
3296 EC_GFp_mont_method() reuses most functions that are part of
3297 EC_GFp_simple_method, but uses Montgomery arithmetic.
3299 [Bodo Moeller; point addition and point doubling
3300 implementation directly derived from source code provided by
3301 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
3303 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
3304 crypto/ec/ec_lib.c):
3306 Curves are EC_GROUP objects (with an optional group generator)
3307 based on EC_METHODs that are built into the library.
3309 Points are EC_POINT objects based on EC_GROUP objects.
3311 Most of the framework would be able to handle curves over arbitrary
3312 finite fields, but as there are no obvious types for fields other
3313 than GF(p), some functions are limited to that for now.
3316 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
3317 that the file contains a complete HTTP response.
3320 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
3321 change the def and num file printf format specifier from "%-40sXXX"
3322 to "%-39s XXX". The latter will always guarantee a space after the
3323 field while the former will cause them to run together if the field
3324 is 40 of more characters long.
3327 *) Constify the cipher and digest 'method' functions and structures
3328 and modify related functions to take constant EVP_MD and EVP_CIPHER
3332 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
3333 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
3336 *) Modify EVP_Digest*() routines so they now return values. Although the
3337 internal software routines can never fail additional hardware versions
3341 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
3343 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
3344 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
3347 ERR_R_NESTED_ASN1_ERROR
3349 ERR_R_MISSING_ASN1_EOS
3350 were 4 .. 9, conflicting with
3351 ERR_LIB_RSA (= ERR_R_RSA_LIB)
3353 ERR_LIB_PEM (= ERR_R_PEM_LIB).
3354 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
3356 Add new error code 'ERR_R_INTERNAL_ERROR'.
3359 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
3363 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
3364 sets the subject name for a new request or supersedes the
3365 subject name in a given request. Formats that can be parsed are
3366 'CN=Some Name, OU=myOU, C=IT'
3368 'CN=Some Name/OU=myOU/C=IT'.
3370 Add options '-batch' and '-verbose' to 'openssl req'.
3371 [Massimiliano Pala <madwolf@hackmasters.net>]
3373 *) Introduce the possibility to access global variables through
3374 functions on platform were that's the best way to handle exporting
3375 global variables in shared libraries. To enable this functionality,
3376 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
3377 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
3378 is normally done by Configure or something similar).
3380 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
3381 in the source file (foo.c) like this:
3383 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
3384 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
3386 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
3387 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
3389 OPENSSL_DECLARE_GLOBAL(int,foo);
3390 #define foo OPENSSL_GLOBAL_REF(foo)
3391 OPENSSL_DECLARE_GLOBAL(double,bar);
3392 #define bar OPENSSL_GLOBAL_REF(bar)
3394 The #defines are very important, and therefore so is including the
3395 header file everywhere where the defined globals are used.
3397 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
3398 of ASN.1 items, but that structure is a bit different.
3400 The largest change is in util/mkdef.pl which has been enhanced with
3401 better and easier to understand logic to choose which symbols should
3402 go into the Windows .def files as well as a number of fixes and code
3403 cleanup (among others, algorithm keywords are now sorted
3404 lexicographically to avoid constant rewrites).
3407 *) In BN_div() keep a copy of the sign of 'num' before writing the
3408 result to 'rm' because if rm==num the value will be overwritten
3409 and produce the wrong result if 'num' is negative: this caused
3410 problems with BN_mod() and BN_nnmod().
3413 *) Function OCSP_request_verify(). This checks the signature on an
3414 OCSP request and verifies the signer certificate. The signer
3415 certificate is just checked for a generic purpose and OCSP request
3419 *) Add OCSP_check_validity() function to check the validity of OCSP
3420 responses. OCSP responses are prepared in real time and may only
3421 be a few seconds old. Simply checking that the current time lies
3422 between thisUpdate and nextUpdate max reject otherwise valid responses
3423 caused by either OCSP responder or client clock inaccuracy. Instead
3424 we allow thisUpdate and nextUpdate to fall within a certain period of
3425 the current time. The age of the response can also optionally be
3426 checked. Two new options -validity_period and -status_age added to
3430 *) If signature or public key algorithm is unrecognized print out its
3431 OID rather that just UNKNOWN.
3434 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
3435 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
3436 ID to be generated from the issuer certificate alone which can then be
3437 passed to OCSP_id_issuer_cmp().
3440 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
3441 ASN1 modules to export functions returning ASN1_ITEM pointers
3442 instead of the ASN1_ITEM structures themselves. This adds several
3443 new macros which allow the underlying ASN1 function/structure to
3444 be accessed transparently. As a result code should not use ASN1_ITEM
3445 references directly (such as &X509_it) but instead use the relevant
3446 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
3447 use of the new ASN1 code on platforms where exporting structures
3448 is problematical (for example in shared libraries) but exporting
3449 functions returning pointers to structures is not.
3452 *) Add support for overriding the generation of SSL/TLS session IDs.
3453 These callbacks can be registered either in an SSL_CTX or per SSL.
3454 The purpose of this is to allow applications to control, if they wish,
3455 the arbitrary values chosen for use as session IDs, particularly as it
3456 can be useful for session caching in multiple-server environments. A
3457 command-line switch for testing this (and any client code that wishes
3458 to use such a feature) has been added to "s_server".
3459 [Geoff Thorpe, Lutz Jaenicke]
3461 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
3462 of the form '#if defined(...) || defined(...) || ...' and
3463 '#if !defined(...) && !defined(...) && ...'. This also avoids
3464 the growing number of special cases it was previously handling.
3467 *) Make all configuration macros available for application by making
3468 sure they are available in opensslconf.h, by giving them names starting
3469 with "OPENSSL_" to avoid conflicts with other packages and by making
3470 sure e_os2.h will cover all platform-specific cases together with
3472 Additionally, it is now possible to define configuration/platform-
3473 specific names (called "system identities"). In the C code, these
3474 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
3475 macro with the name beginning with "OPENSSL_SYS_", which is determined
3476 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
3480 *) New option -set_serial to 'req' and 'x509' this allows the serial
3481 number to use to be specified on the command line. Previously self
3482 signed certificates were hard coded with serial number 0 and the
3483 CA options of 'x509' had to use a serial number in a file which was
3487 *) New options to 'ca' utility to support V2 CRL entry extensions.
3488 Currently CRL reason, invalidity date and hold instruction are
3489 supported. Add new CRL extensions to V3 code and some new objects.
3492 *) New function EVP_CIPHER_CTX_set_padding() this is used to
3493 disable standard block padding (aka PKCS#5 padding) in the EVP
3494 API, which was previously mandatory. This means that the data is
3495 not padded in any way and so the total length much be a multiple
3496 of the block size, otherwise an error occurs.
3499 *) Initial (incomplete) OCSP SSL support.
3502 *) New function OCSP_parse_url(). This splits up a URL into its host,
3503 port and path components: primarily to parse OCSP URLs. New -url
3504 option to ocsp utility.
3507 *) New nonce behavior. The return value of OCSP_check_nonce() now
3508 reflects the various checks performed. Applications can decide
3509 whether to tolerate certain situations such as an absent nonce
3510 in a response when one was present in a request: the ocsp application
3511 just prints out a warning. New function OCSP_add1_basic_nonce()
3512 this is to allow responders to include a nonce in a response even if
3513 the request is nonce-less.
3516 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
3517 skipped when using openssl x509 multiple times on a single input file,
3518 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
3521 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
3522 set string type: to handle setting ASN1_TIME structures. Fix ca
3523 utility to correctly initialize revocation date of CRLs.
3526 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
3527 the clients preferred ciphersuites and rather use its own preferences.
3528 Should help to work around M$ SGC (Server Gated Cryptography) bug in
3529 Internet Explorer by ensuring unchanged hash method during stepup.
3530 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
3533 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
3534 to aes and add a new 'exist' option to print out symbols that don't
3538 *) Additional options to ocsp utility to allow flags to be set and
3539 additional certificates supplied.
3542 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
3543 OCSP client a number of certificate to only verify the response
3547 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
3548 handle the new API. Currently only ECB, CBC modes supported. Add new
3551 Add TLS AES ciphersuites as described in RFC3268, "Advanced
3552 Encryption Standard (AES) Ciphersuites for Transport Layer
3553 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
3554 not enabled by default and were not part of the "ALL" ciphersuite
3555 alias because they were not yet official; they could be
3556 explicitly requested by specifying the "AESdraft" ciphersuite
3557 group alias. In the final release of OpenSSL 0.9.7, the group
3558 alias is called "AES" and is part of "ALL".)
3559 [Ben Laurie, Steve Henson, Bodo Moeller]
3561 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
3562 request to response.
3565 *) Functions for OCSP responders. OCSP_request_onereq_count(),
3566 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
3567 extract information from a certificate request. OCSP_response_create()
3568 creates a response and optionally adds a basic response structure.
3569 OCSP_basic_add1_status() adds a complete single response to a basic
3570 response and returns the OCSP_SINGLERESP structure just added (to allow
3571 extensions to be included for example). OCSP_basic_add1_cert() adds a
3572 certificate to a basic response and OCSP_basic_sign() signs a basic
3573 response with various flags. New helper functions ASN1_TIME_check()
3574 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
3575 (converts ASN1_TIME to GeneralizedTime).
3578 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
3579 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
3580 structure from a certificate. X509_pubkey_digest() digests the public_key
3581 contents: this is used in various key identifiers.
3584 *) Make sk_sort() tolerate a NULL argument.
3585 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
3587 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
3588 passed by the function are trusted implicitly. If any of them signed the
3589 response then it is assumed to be valid and is not verified.
3592 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
3593 to data. This was previously part of the PKCS7 ASN1 code. This
3594 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
3595 [Steve Henson, reported by Kenneth R. Robinette
3596 <support@securenetterm.com>]
3598 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
3599 routines: without these tracing memory leaks is very painful.
3600 Fix leaks in PKCS12 and PKCS7 routines.
3603 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
3604 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
3605 effectively meant GeneralizedTime would never be used. Now it
3606 is initialised to -1 but X509_time_adj() now has to check the value
3607 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
3608 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
3609 [Steve Henson, reported by Kenneth R. Robinette
3610 <support@securenetterm.com>]
3612 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
3613 result in a zero length in the ASN1_INTEGER structure which was
3614 not consistent with the structure when d2i_ASN1_INTEGER() was used
3615 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
3616 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
3617 where it did not print out a minus for negative ASN1_INTEGER.
3620 *) Add summary printout to ocsp utility. The various functions which
3621 convert status values to strings have been renamed to:
3622 OCSP_response_status_str(), OCSP_cert_status_str() and
3623 OCSP_crl_reason_str() and are no longer static. New options
3624 to verify nonce values and to disable verification. OCSP response
3625 printout format cleaned up.
3628 *) Add additional OCSP certificate checks. These are those specified
3629 in RFC2560. This consists of two separate checks: the CA of the
3630 certificate being checked must either be the OCSP signer certificate
3631 or the issuer of the OCSP signer certificate. In the latter case the
3632 OCSP signer certificate must contain the OCSP signing extended key
3633 usage. This check is performed by attempting to match the OCSP
3634 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
3635 in the OCSP_CERTID structures of the response.
3638 *) Initial OCSP certificate verification added to OCSP_basic_verify()
3639 and related routines. This uses the standard OpenSSL certificate
3640 verify routines to perform initial checks (just CA validity) and
3641 to obtain the certificate chain. Then additional checks will be
3642 performed on the chain. Currently the root CA is checked to see
3643 if it is explicitly trusted for OCSP signing. This is used to set
3644 a root CA as a global signing root: that is any certificate that
3645 chains to that CA is an acceptable OCSP signing certificate.
3648 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
3649 extensions from a separate configuration file.
3650 As when reading extensions from the main configuration file,
3651 the '-extensions ...' option may be used for specifying the
3653 [Massimiliano Pala <madwolf@comune.modena.it>]
3655 *) New OCSP utility. Allows OCSP requests to be generated or
3656 read. The request can be sent to a responder and the output
3657 parsed, outputed or printed in text form. Not complete yet:
3658 still needs to check the OCSP response validity.
3661 *) New subcommands for 'openssl ca':
3662 'openssl ca -status <serial>' prints the status of the cert with
3663 the given serial number (according to the index file).
3664 'openssl ca -updatedb' updates the expiry status of certificates
3666 [Massimiliano Pala <madwolf@comune.modena.it>]
3668 *) New '-newreq-nodes' command option to CA.pl. This is like
3669 '-newreq', but calls 'openssl req' with the '-nodes' option
3670 so that the resulting key is not encrypted.
3671 [Damien Miller <djm@mindrot.org>]
3673 *) New configuration for the GNU Hurd.
3674 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
3676 *) Initial code to implement OCSP basic response verify. This
3677 is currently incomplete. Currently just finds the signer's
3678 certificate and verifies the signature on the response.
3681 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
3682 value of OPENSSLDIR. This is available via the new '-d' option
3683 to 'openssl version', and is also included in 'openssl version -a'.
3686 *) Allowing defining memory allocation callbacks that will be given
3687 file name and line number information in additional arguments
3688 (a const char* and an int). The basic functionality remains, as
3689 well as the original possibility to just replace malloc(),
3690 realloc() and free() by functions that do not know about these
3691 additional arguments. To register and find out the current
3692 settings for extended allocation functions, the following
3693 functions are provided:
3695 CRYPTO_set_mem_ex_functions
3696 CRYPTO_set_locked_mem_ex_functions
3697 CRYPTO_get_mem_ex_functions
3698 CRYPTO_get_locked_mem_ex_functions
3700 These work the same way as CRYPTO_set_mem_functions and friends.
3701 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
3702 extended allocation function is enabled.
3703 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
3704 a conventional allocation function is enabled.
3705 [Richard Levitte, Bodo Moeller]
3707 *) Finish off removing the remaining LHASH function pointer casts.
3708 There should no longer be any prototype-casting required when using
3709 the LHASH abstraction, and any casts that remain are "bugs". See
3710 the callback types and macros at the head of lhash.h for details
3711 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
3714 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
3715 If /dev/[u]random devices are not available or do not return enough
3716 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
3718 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
3719 /etc/entropy will be queried once each in this sequence, quering stops
3720 when enough entropy was collected without querying more sockets.
3723 *) Change the Unix RAND_poll() variant to be able to poll several
3724 random devices, as specified by DEVRANDOM, until a sufficient amount
3725 of data has been collected. We spend at most 10 ms on each file
3726 (select timeout) and read in non-blocking mode. DEVRANDOM now
3727 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
3728 (previously it was just the string "/dev/urandom"), so on typical
3729 platforms the 10 ms delay will never occur.
3730 Also separate out the Unix variant to its own file, rand_unix.c.
3731 For VMS, there's a currently-empty rand_vms.c.
3734 *) Move OCSP client related routines to ocsp_cl.c. These
3735 provide utility functions which an application needing
3736 to issue a request to an OCSP responder and analyse the
3737 response will typically need: as opposed to those which an
3738 OCSP responder itself would need which will be added later.
3740 OCSP_request_sign() signs an OCSP request with an API similar
3741 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
3742 response. OCSP_response_get1_basic() extracts basic response
3743 from response. OCSP_resp_find_status(): finds and extracts status
3744 information from an OCSP_CERTID structure (which will be created
3745 when the request structure is built). These are built from lower
3746 level functions which work on OCSP_SINGLERESP structures but
3747 wont normally be used unless the application wishes to examine
3748 extensions in the OCSP response for example.
3750 Replace nonce routines with a pair of functions.
3751 OCSP_request_add1_nonce() adds a nonce value and optionally
3752 generates a random value. OCSP_check_nonce() checks the
3753 validity of the nonce in an OCSP response.
3756 *) Change function OCSP_request_add() to OCSP_request_add0_id().
3757 This doesn't copy the supplied OCSP_CERTID and avoids the
3758 need to free up the newly created id. Change return type
3759 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
3760 This can then be used to add extensions to the request.
3761 Deleted OCSP_request_new(), since most of its functionality
3762 is now in OCSP_REQUEST_new() (and the case insensitive name
3763 clash) apart from the ability to set the request name which
3764 will be added elsewhere.
3767 *) Update OCSP API. Remove obsolete extensions argument from
3768 various functions. Extensions are now handled using the new
3769 OCSP extension code. New simple OCSP HTTP function which
3770 can be used to send requests and parse the response.
3773 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
3774 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
3775 uses the special reorder version of SET OF to sort the attributes
3776 and reorder them to match the encoded order. This resolves a long
3777 standing problem: a verify on a PKCS7 structure just after signing
3778 it used to fail because the attribute order did not match the
3779 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
3780 it uses the received order. This is necessary to tolerate some broken
3781 software that does not order SET OF. This is handled by encoding
3782 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
3783 to produce the required SET OF.
3786 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
3787 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
3788 files to get correct declarations of the ASN.1 item variables.
3791 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
3792 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
3793 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
3794 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
3795 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
3796 ASN1_ITEM and no wrapper functions.
3799 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
3800 replace the old function pointer based I/O routines. Change most of
3801 the *_d2i_bio() and *_d2i_fp() functions to use these.
3804 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
3805 lines, recognice more "algorithms" that can be deselected, and make
3806 it complain about algorithm deselection that isn't recognised.
3809 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
3810 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
3811 to use new functions. Add NO_ASN1_OLD which can be set to remove
3812 some old style ASN1 functions: this can be used to determine if old
3813 code will still work when these eventually go away.
3816 *) New extension functions for OCSP structures, these follow the
3817 same conventions as certificates and CRLs.
3820 *) New function X509V3_add1_i2d(). This automatically encodes and
3821 adds an extension. Its behaviour can be customised with various
3822 flags to append, replace or delete. Various wrappers added for
3823 certifcates and CRLs.
3826 *) Fix to avoid calling the underlying ASN1 print routine when
3827 an extension cannot be parsed. Correct a typo in the
3828 OCSP_SERVICELOC extension. Tidy up print OCSP format.
3831 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
3832 entries for variables.
3835 *) Add functionality to apps/openssl.c for detecting locking
3836 problems: As the program is single-threaded, all we have
3837 to do is register a locking callback using an array for
3838 storing which locks are currently held by the program.
3841 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
3842 SSL_get_ex_data_X509_STORE_idx(), which is used in
3843 ssl_verify_cert_chain() and thus can be called at any time
3844 during TLS/SSL handshakes so that thread-safety is essential.
3845 Unfortunately, the ex_data design is not at all suited
3846 for multi-threaded use, so it probably should be abolished.
3849 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
3850 [Broadcom, tweaked and integrated by Geoff Thorpe]
3852 *) Move common extension printing code to new function
3853 X509V3_print_extensions(). Reorganise OCSP print routines and
3854 implement some needed OCSP ASN1 functions. Add OCSP extensions.
3857 *) New function X509_signature_print() to remove duplication in some
3861 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
3862 set (this was treated exactly the same as SET OF previously). This
3863 is used to reorder the STACK representing the structure to match the
3864 encoding. This will be used to get round a problem where a PKCS7
3865 structure which was signed could not be verified because the STACK
3866 order did not reflect the encoded order.
3869 *) Reimplement the OCSP ASN1 module using the new code.
3872 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
3873 for its ASN1 operations. The old style function pointers still exist
3874 for now but they will eventually go away.
3877 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
3878 completely replaces the old ASN1 functionality with a table driven
3879 encoder and decoder which interprets an ASN1_ITEM structure describing
3880 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
3881 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
3882 has also been converted to the new form.
3885 *) Change BN_mod_exp_recp so that negative moduli are tolerated
3886 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
3887 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
3888 for negative moduli.
3891 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
3892 of not touching the result's sign bit.
3895 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
3899 *) Changed the LHASH code to use prototypes for callbacks, and created
3900 macros to declare and implement thin (optionally static) functions
3901 that provide type-safety and avoid function pointer casting for the
3902 type-specific callbacks.
3905 *) Added Kerberos Cipher Suites to be used with TLS, as written in
3907 [Veers Staats <staatsvr@asc.hpc.mil>,
3908 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
3910 *) Reformat the FAQ so the different questions and answers can be divided
3911 in sections depending on the subject.
3914 *) Have the zlib compression code load ZLIB.DLL dynamically under
3918 *) New function BN_mod_sqrt for computing square roots modulo a prime
3919 (using the probabilistic Tonelli-Shanks algorithm unless
3920 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
3921 be handled deterministically).
3922 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3924 *) Make BN_mod_inverse faster by explicitly handling small quotients
3925 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
3926 512 bits], about 30% for larger ones [1024 or 2048 bits].)
3929 *) New function BN_kronecker.
3932 *) Fix BN_gcd so that it works on negative inputs; the result is
3933 positive unless both parameters are zero.
3934 Previously something reasonably close to an infinite loop was
3935 possible because numbers could be growing instead of shrinking
3936 in the implementation of Euclid's algorithm.
3939 *) Fix BN_is_word() and BN_is_one() macros to take into account the
3940 sign of the number in question.
3942 Fix BN_is_word(a,w) to work correctly for w == 0.
3944 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
3945 because its test if the absolute value of 'a' equals 'w'.
3946 Note that BN_abs_is_word does *not* handle w == 0 reliably;
3947 it exists mostly for use in the implementations of BN_is_zero(),
3948 BN_is_one(), and BN_is_word().
3951 *) New function BN_swap.
3954 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
3955 the exponentiation functions are more likely to produce reasonable
3956 results on negative inputs.
3959 *) Change BN_mod_mul so that the result is always non-negative.
3960 Previously, it could be negative if one of the factors was negative;
3961 I don't think anyone really wanted that behaviour.
3964 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
3965 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
3966 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
3967 and add new functions:
3976 BN_mod_lshift1_quick
3980 These functions always generate non-negative results.
3982 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
3983 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
3985 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
3986 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
3987 be reduced modulo m.
3988 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3991 The following entry accidentily appeared in the CHANGES file
3992 distributed with OpenSSL 0.9.7. The modifications described in
3993 it do *not* apply to OpenSSL 0.9.7.
3995 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3996 was actually never needed) and in BN_mul(). The removal in BN_mul()
3997 required a small change in bn_mul_part_recursive() and the addition
3998 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3999 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4000 bn_sub_words() and bn_add_words() except they take arrays with
4005 *) In 'openssl passwd', verify passwords read from the terminal
4006 unless the '-salt' option is used (which usually means that
4007 verification would just waste user's time since the resulting
4008 hash is going to be compared with some given password hash)
4009 or the new '-noverify' option is used.
4011 This is an incompatible change, but it does not affect
4012 non-interactive use of 'openssl passwd' (passwords on the command
4013 line, '-stdin' option, '-in ...' option) and thus should not
4017 *) Remove all references to RSAref, since there's no more need for it.
4020 *) Make DSO load along a path given through an environment variable
4021 (SHLIB_PATH) with shl_load().
4024 *) Constify the ENGINE code as a result of BIGNUM constification.
4025 Also constify the RSA code and most things related to it. In a
4026 few places, most notable in the depth of the ASN.1 code, ugly
4027 casts back to non-const were required (to be solved at a later
4031 *) Make it so the openssl application has all engines loaded by default.
4034 *) Constify the BIGNUM routines a little more.
4037 *) Add the following functions:
4039 ENGINE_load_cswift()
4041 ENGINE_load_atalla()
4043 ENGINE_load_builtin_engines()
4045 That way, an application can itself choose if external engines that
4046 are built-in in OpenSSL shall ever be used or not. The benefit is
4047 that applications won't have to be linked with libdl or other dso
4048 libraries unless it's really needed.
4050 Changed 'openssl engine' to load all engines on demand.
4051 Changed the engine header files to avoid the duplication of some
4052 declarations (they differed!).
4055 *) 'openssl engine' can now list capabilities.
4058 *) Better error reporting in 'openssl engine'.
4061 *) Never call load_dh_param(NULL) in s_server.
4064 *) Add engine application. It can currently list engines by name and
4065 identity, and test if they are actually available.
4068 *) Improve RPM specification file by forcing symbolic linking and making
4069 sure the installed documentation is also owned by root.root.
4070 [Damien Miller <djm@mindrot.org>]
4072 *) Give the OpenSSL applications more possibilities to make use of
4073 keys (public as well as private) handled by engines.
4076 *) Add OCSP code that comes from CertCo.
4079 *) Add VMS support for the Rijndael code.
4082 *) Added untested support for Nuron crypto accelerator.
4085 *) Add support for external cryptographic devices. This code was
4086 previously distributed separately as the "engine" branch.
4087 [Geoff Thorpe, Richard Levitte]
4089 *) Rework the filename-translation in the DSO code. It is now possible to
4090 have far greater control over how a "name" is turned into a filename
4091 depending on the operating environment and any oddities about the
4092 different shared library filenames on each system.
4095 *) Support threads on FreeBSD-elf in Configure.
4098 *) Fix for SHA1 assembly problem with MASM: it produces
4099 warnings about corrupt line number information when assembling
4100 with debugging information. This is caused by the overlapping
4102 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
4105 NCONF_get_number() has no error checking at all. As a replacement,
4106 NCONF_get_number_e() is defined (_e for "error checking") and is
4107 promoted strongly. The old NCONF_get_number is kept around for
4108 binary backward compatibility.
4109 Make it possible for methods to load from something other than a BIO,
4110 by providing a function pointer that is given a name instead of a BIO.
4111 For example, this could be used to load configuration data from an
4115 *) Fix for non blocking accept BIOs. Added new I/O special reason
4116 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
4117 with non blocking I/O was not possible because no retry code was
4118 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
4122 *) Added the beginnings of Rijndael support.
4125 *) Fix for bug in DirectoryString mask setting. Add support for
4126 X509_NAME_print_ex() in 'req' and X509_print_ex() function
4127 to allow certificate printing to more controllable, additional
4128 'certopt' option to 'x509' to allow new printing options to be
4132 *) Clean old EAY MD5 hack from e_os.h.
4135 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
4137 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4138 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4139 [Joe Orton, Steve Henson]
4141 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
4143 *) Fix additional bug revealed by the NISCC test suite:
4145 Stop bug triggering large recursion when presented with
4146 certain ASN.1 tags (CVE-2003-0851)
4149 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
4151 *) Fix various bugs revealed by running the NISCC test suite:
4153 Stop out of bounds reads in the ASN1 code when presented with
4154 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4156 If verify callback ignores invalid public key errors don't try to check
4157 certificate signature with the NULL public key.
4161 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4162 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4166 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4167 extra data after the compression methods not only for TLS 1.0
4168 but also for SSL 3.0 (as required by the specification).
4169 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4171 *) Change X509_certificate_type() to mark the key as exported/exportable
4172 when it's 512 *bits* long, not 512 bytes.
4175 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
4177 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4178 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4179 a protocol version number mismatch like a decryption error
4180 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4183 *) Turn on RSA blinding by default in the default implementation
4184 to avoid a timing attack. Applications that don't want it can call
4185 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4186 They would be ill-advised to do so in most cases.
4187 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4189 *) Change RSA blinding code so that it works when the PRNG is not
4190 seeded (in this case, the secret RSA exponent is abused as
4191 an unpredictable seed -- if it is not unpredictable, there
4192 is no point in blinding anyway). Make RSA blinding thread-safe
4193 by remembering the creator's thread ID in rsa->blinding and
4194 having all other threads use local one-time blinding factors
4195 (this requires more computation than sharing rsa->blinding, but
4196 avoids excessive locking; and if an RSA object is not shared
4197 between threads, blinding will still be very fast).
4200 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
4202 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4203 via timing by performing a MAC computation even if incorrrect
4204 block cipher padding has been found. This is a countermeasure
4205 against active attacks where the attacker has to distinguish
4206 between bad padding and a MAC verification error. (CVE-2003-0078)
4208 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4209 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4210 Martin Vuagnoux (EPFL, Ilion)]
4212 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
4214 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
4215 memory from it's contents. This is done with a counter that will
4216 place alternating values in each byte. This can be used to solve
4217 two issues: 1) the removal of calls to memset() by highly optimizing
4218 compilers, and 2) cleansing with other values than 0, since those can
4219 be read through on certain media, for example a swap space on disk.
4222 *) Bugfix: client side session caching did not work with external caching,
4223 because the session->cipher setting was not restored when reloading
4224 from the external cache. This problem was masked, when
4225 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
4226 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
4229 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
4230 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
4231 [Zeev Lieber <zeev-l@yahoo.com>]
4233 *) Undo an undocumented change introduced in 0.9.6e which caused
4234 repeated calls to OpenSSL_add_all_ciphers() and
4235 OpenSSL_add_all_digests() to be ignored, even after calling
4239 *) Change the default configuration reader to deal with last line not
4240 being properly terminated.
4243 *) Change X509_NAME_cmp() so it applies the special rules on handling
4244 DN values that are of type PrintableString, as well as RDNs of type
4245 emailAddress where the value has the type ia5String.
4246 [stefank@valicert.com via Richard Levitte]
4248 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
4249 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
4250 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
4251 the bitwise-OR of the two for use by the majority of applications
4252 wanting this behaviour, and update the docs. The documented
4253 behaviour and actual behaviour were inconsistent and had been
4254 changing anyway, so this is more a bug-fix than a behavioural
4256 [Geoff Thorpe, diagnosed by Nadav Har'El]
4258 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
4259 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
4262 *) Fix initialization code race conditions in
4263 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
4264 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
4265 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
4266 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
4267 ssl2_get_cipher_by_char(),
4268 ssl3_get_cipher_by_char().
4269 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
4271 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
4272 the cached sessions are flushed, as the remove_cb() might use ex_data
4273 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
4274 (see [openssl.org #212]).
4275 [Geoff Thorpe, Lutz Jaenicke]
4277 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
4278 length, instead of the encoding length to d2i_ASN1_OBJECT.
4281 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
4283 *) [In 0.9.6g-engine release:]
4284 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
4285 [Lynn Gazis <lgazis@rainbow.com>]
4287 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
4289 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
4290 and get fix the header length calculation.
4291 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
4292 Alon Kantor <alonk@checkpoint.com> (and others),
4295 *) Use proper error handling instead of 'assertions' in buffer
4296 overflow checks added in 0.9.6e. This prevents DoS (the
4297 assertions could call abort()).
4298 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
4300 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
4302 *) Add various sanity checks to asn1_get_length() to reject
4303 the ASN1 length bytes if they exceed sizeof(long), will appear
4304 negative or the content length exceeds the length of the
4306 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4308 *) Fix cipher selection routines: ciphers without encryption had no flags
4309 for the cipher strength set and where therefore not handled correctly
4310 by the selection routines (PR #130).
4313 *) Fix EVP_dsa_sha macro.
4317 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
4318 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
4319 that was added in OpenSSL 0.9.6d.
4321 As the countermeasure turned out to be incompatible with some
4322 broken SSL implementations, the new option is part of SSL_OP_ALL.
4323 SSL_OP_ALL is usually employed when compatibility with weird SSL
4324 implementations is desired (e.g. '-bugs' option to 's_client' and
4325 's_server'), so the new option is automatically set in many
4329 *) Changes in security patch:
4331 Changes marked "(CHATS)" were sponsored by the Defense Advanced
4332 Research Projects Agency (DARPA) and Air Force Research Laboratory,
4333 Air Force Materiel Command, USAF, under agreement number
4336 *) Add various sanity checks to asn1_get_length() to reject
4337 the ASN1 length bytes if they exceed sizeof(long), will appear
4338 negative or the content length exceeds the length of the
4339 supplied buffer. (CVE-2002-0659)
4340 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4342 *) Assertions for various potential buffer overflows, not known to
4344 [Ben Laurie (CHATS)]
4346 *) Various temporary buffers to hold ASCII versions of integers were
4347 too small for 64 bit platforms. (CVE-2002-0655)
4348 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
4350 *) Remote buffer overflow in SSL3 protocol - an attacker could
4351 supply an oversized session ID to a client. (CVE-2002-0656)
4352 [Ben Laurie (CHATS)]
4354 *) Remote buffer overflow in SSL2 protocol - an attacker could
4355 supply an oversized client master key. (CVE-2002-0656)
4356 [Ben Laurie (CHATS)]
4358 Changes between 0.9.6c and 0.9.6d [9 May 2002]
4360 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
4361 encoded as NULL) with id-dsa-with-sha1.
4362 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
4364 *) Check various X509_...() return values in apps/req.c.
4365 [Nils Larsch <nla@trustcenter.de>]
4367 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
4368 an end-of-file condition would erronously be flagged, when the CRLF
4369 was just at the end of a processed block. The bug was discovered when
4370 processing data through a buffering memory BIO handing the data to a
4371 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
4372 <ptsekov@syntrex.com> and Nedelcho Stanev.
4375 *) Implement a countermeasure against a vulnerability recently found
4376 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
4377 before application data chunks to avoid the use of known IVs
4378 with data potentially chosen by the attacker.
4381 *) Fix length checks in ssl3_get_client_hello().
4384 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
4385 to prevent ssl3_read_internal() from incorrectly assuming that
4386 ssl3_read_bytes() found application data while handshake
4387 processing was enabled when in fact s->s3->in_read_app_data was
4388 merely automatically cleared during the initial handshake.
4389 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
4391 *) Fix object definitions for Private and Enterprise: they were not
4392 recognized in their shortname (=lowercase) representation. Extend
4393 obj_dat.pl to issue an error when using undefined keywords instead
4394 of silently ignoring the problem (Svenning Sorensen
4395 <sss@sss.dnsalias.net>).
4398 *) Fix DH_generate_parameters() so that it works for 'non-standard'
4399 generators, i.e. generators other than 2 and 5. (Previously, the
4400 code did not properly initialise the 'add' and 'rem' values to
4401 BN_generate_prime().)
4403 In the new general case, we do not insist that 'generator' is
4404 actually a primitive root: This requirement is rather pointless;
4405 a generator of the order-q subgroup is just as good, if not
4409 *) Map new X509 verification errors to alerts. Discovered and submitted by
4410 Tom Wu <tom@arcot.com>.
4413 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
4414 returning non-zero before the data has been completely received
4415 when using non-blocking I/O.
4416 [Bodo Moeller; problem pointed out by John Hughes]
4418 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
4419 [Ben Laurie, Lutz Jaenicke]
4421 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
4422 Yoram Zahavi <YoramZ@gilian.com>).
4425 *) Add information about CygWin 1.3 and on, and preserve proper
4426 configuration for the versions before that.
4427 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4429 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
4430 check whether we deal with a copy of a session and do not delete from
4431 the cache in this case. Problem reported by "Izhar Shoshani Levi"
4432 <izhar@checkpoint.com>.
4435 *) Do not store session data into the internal session cache, if it
4436 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
4437 flag is set). Proposed by Aslam <aslam@funk.com>.
4440 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
4444 *) [In 0.9.6d-engine release:]
4445 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4446 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4448 *) Add the configuration target linux-s390x.
4449 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
4451 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
4452 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
4453 variable as an indication that a ClientHello message has been
4454 received. As the flag value will be lost between multiple
4455 invocations of ssl3_accept when using non-blocking I/O, the
4456 function may not be aware that a handshake has actually taken
4457 place, thus preventing a new session from being added to the
4460 To avoid this problem, we now set s->new_session to 2 instead of
4461 using a local variable.
4462 [Lutz Jaenicke, Bodo Moeller]
4464 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
4465 if the SSL_R_LENGTH_MISMATCH error is detected.
4466 [Geoff Thorpe, Bodo Moeller]
4468 *) New 'shared_ldflag' column in Configure platform table.
4471 *) Fix EVP_CIPHER_mode macro.
4472 ["Dan S. Camper" <dan@bti.net>]
4474 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
4475 type, we must throw them away by setting rr->length to 0.
4476 [D P Chang <dpc@qualys.com>]
4478 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
4480 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
4481 <Dominikus.Scherkl@biodata.com>. (The previous implementation
4482 worked incorrectly for those cases where range = 10..._2 and
4483 3*range is two bits longer than range.)
4486 *) Only add signing time to PKCS7 structures if it is not already
4490 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
4491 OBJ_ld_ce should be OBJ_id_ce.
4492 Also some ip-pda OIDs in crypto/objects/objects.txt were
4493 incorrect (cf. RFC 3039).
4494 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
4496 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
4497 returns early because it has nothing to do.
4498 [Andy Schneider <andy.schneider@bjss.co.uk>]
4500 *) [In 0.9.6c-engine release:]
4501 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
4502 [Andy Schneider <andy.schneider@bjss.co.uk>]
4504 *) [In 0.9.6c-engine release:]
4505 Add support for Cryptographic Appliance's keyserver technology.
4506 (Use engine 'keyclient')
4507 [Cryptographic Appliances and Geoff Thorpe]
4509 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
4510 is called via tools/c89.sh because arguments have to be
4511 rearranged (all '-L' options must appear before the first object
4513 [Richard Shapiro <rshapiro@abinitio.com>]
4515 *) [In 0.9.6c-engine release:]
4516 Add support for Broadcom crypto accelerator cards, backported
4518 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
4520 *) [In 0.9.6c-engine release:]
4521 Add support for SureWare crypto accelerator cards from
4522 Baltimore Technologies. (Use engine 'sureware')
4523 [Baltimore Technologies and Mark Cox]
4525 *) [In 0.9.6c-engine release:]
4526 Add support for crypto accelerator cards from Accelerated
4527 Encryption Processing, www.aep.ie. (Use engine 'aep')
4528 [AEP Inc. and Mark Cox]
4530 *) Add a configuration entry for gcc on UnixWare.
4531 [Gary Benson <gbenson@redhat.com>]
4533 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
4534 messages are stored in a single piece (fixed-length part and
4535 variable-length part combined) and fix various bugs found on the way.
4538 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
4539 instead. BIO_gethostbyname() does not know what timeouts are
4540 appropriate, so entries would stay in cache even when they have
4542 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
4544 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
4545 faced with a pathologically small ClientHello fragment that does
4546 not contain client_version: Instead of aborting with an error,
4547 simply choose the highest available protocol version (i.e.,
4548 TLS 1.0 unless it is disabled). In practice, ClientHello
4549 messages are never sent like this, but this change gives us
4550 strictly correct behaviour at least for TLS.
4553 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
4554 never resets s->method to s->ctx->method when called from within
4555 one of the SSL handshake functions.
4556 [Bodo Moeller; problem pointed out by Niko Baric]
4558 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
4559 (sent using the client's version number) if client_version is
4560 smaller than the protocol version in use. Also change
4561 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
4562 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
4563 the client will at least see that alert.
4566 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
4570 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
4571 client receives HelloRequest while in a handshake.
4572 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
4574 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
4575 should end in 'break', not 'goto end' which circuments various
4576 cleanups done in state SSL_ST_OK. But session related stuff
4577 must be disabled for SSL_ST_OK in the case that we just sent a
4580 Also avoid some overhead by not calling ssl_init_wbio_buffer()
4581 before just sending a HelloRequest.
4582 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
4584 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
4585 reveal whether illegal block cipher padding was found or a MAC
4586 verification error occured. (Neither SSLerr() codes nor alerts
4587 are directly visible to potential attackers, but the information
4588 may leak via logfiles.)
4590 Similar changes are not required for the SSL 2.0 implementation
4591 because the number of padding bytes is sent in clear for SSL 2.0,
4592 and the extra bytes are just ignored. However ssl/s2_pkt.c
4593 failed to verify that the purported number of padding bytes is in
4597 *) Add OpenUNIX-8 support including shared libraries
4598 (Boyd Lynn Gerber <gerberb@zenez.com>).
4601 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
4602 'wristwatch attack' using huge encoding parameters (cf.
4603 James H. Manger's CRYPTO 2001 paper). Note that the
4604 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
4605 encoding parameters and hence was not vulnerable.
4608 *) BN_sqr() bug fix.
4609 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
4611 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
4612 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
4613 followed by modular reduction.
4614 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
4616 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
4617 equivalent based on BN_pseudo_rand() instead of BN_rand().
4620 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
4621 This function was broken, as the check for a new client hello message
4622 to handle SGC did not allow these large messages.
4623 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
4626 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
4629 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
4630 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
4633 *) Rework the configuration and shared library support for Tru64 Unix.
4634 The configuration part makes use of modern compiler features and
4635 still retains old compiler behavior for those that run older versions
4636 of the OS. The shared library support part includes a variant that
4637 uses the RPATH feature, and is available through the special
4638 configuration target "alpha-cc-rpath", which will never be selected
4640 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
4642 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
4643 with the same message size as in ssl3_get_certificate_request().
4644 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
4645 messages might inadvertently be reject as too long.
4646 [Petr Lampa <lampa@fee.vutbr.cz>]
4648 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
4651 *) Modified SSL library such that the verify_callback that has been set
4652 specificly for an SSL object with SSL_set_verify() is actually being
4653 used. Before the change, a verify_callback set with this function was
4654 ignored and the verify_callback() set in the SSL_CTX at the time of
4655 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
4656 to allow the necessary settings.
4659 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
4660 explicitly to NULL, as at least on Solaris 8 this seems not always to be
4661 done automatically (in contradiction to the requirements of the C
4662 standard). This made problems when used from OpenSSH.
4665 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
4666 dh->length and always used
4668 BN_rand_range(priv_key, dh->p).
4670 BN_rand_range() is not necessary for Diffie-Hellman, and this
4671 specific range makes Diffie-Hellman unnecessarily inefficient if
4672 dh->length (recommended exponent length) is much smaller than the
4673 length of dh->p. We could use BN_rand_range() if the order of
4674 the subgroup was stored in the DH structure, but we only have
4679 BN_rand(priv_key, l, ...)
4681 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
4687 RSA_eay_public_encrypt
4688 RSA_eay_private_decrypt
4689 RSA_eay_private_encrypt (signing)
4690 RSA_eay_public_decrypt (signature verification)
4692 (default implementations for RSA_public_encrypt,
4693 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
4694 always reject numbers >= n.
4697 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
4698 to synchronize access to 'locking_thread'. This is necessary on
4699 systems where access to 'locking_thread' (an 'unsigned long'
4700 variable) is not atomic.
4703 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
4704 *before* setting the 'crypto_lock_rand' flag. The previous code had
4705 a race condition if 0 is a valid thread ID.
4706 [Travis Vitek <vitek@roguewave.com>]
4708 *) Add support for shared libraries under Irix.
4709 [Albert Chin-A-Young <china@thewrittenword.com>]
4711 *) Add configuration option to build on Linux on both big-endian and
4713 [Ralf Baechle <ralf@uni-koblenz.de>]
4715 *) Add the possibility to create shared libraries on HP-UX.
4718 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
4720 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
4721 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
4722 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
4723 PRNG state recovery was possible based on the output of
4724 one PRNG request appropriately sized to gain knowledge on
4725 'md' followed by enough consecutive 1-byte PRNG requests
4726 to traverse all of 'state'.
4728 1. When updating 'md_local' (the current thread's copy of 'md')
4729 during PRNG output generation, hash all of the previous
4730 'md_local' value, not just the half used for PRNG output.
4732 2. Make the number of bytes from 'state' included into the hash
4733 independent from the number of PRNG bytes requested.
4735 The first measure alone would be sufficient to avoid
4736 Markku-Juhani's attack. (Actually it had never occurred
4737 to me that the half of 'md_local' used for chaining was the
4738 half from which PRNG output bytes were taken -- I had always
4739 assumed that the secret half would be used.) The second
4740 measure makes sure that additional data from 'state' is never
4741 mixed into 'md_local' in small portions; this heuristically
4742 further strengthens the PRNG.
4745 *) Fix crypto/bn/asm/mips3.s.
4748 *) When only the key is given to "enc", the IV is undefined. Print out
4749 an error message in this case.
4752 *) Handle special case when X509_NAME is empty in X509 printing routines.
4755 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
4756 positive and less than q.
4759 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
4760 used: it isn't thread safe and the add_lock_callback should handle
4762 [Paul Rose <Paul.Rose@bridge.com>]
4764 *) Verify that incoming data obeys the block size in
4765 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
4769 [Ulf Möller, Bodo Möller]
4771 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
4772 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
4773 when fixing the server behaviour for backwards-compatible 'client
4774 hello' messages. (Note that the attack is impractical against
4775 SSL 3.0 and TLS 1.0 anyway because length and version checking
4776 means that the probability of guessing a valid ciphertext is
4777 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
4780 Before 0.9.5, the countermeasure (hide the error by generating a
4781 random 'decryption result') did not work properly because
4782 ERR_clear_error() was missing, meaning that SSL_get_error() would
4783 detect the supposedly ignored error.
4785 Both problems are now fixed.
4788 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
4789 (previously it was 1024).
4792 *) Fix for compatibility mode trust settings: ignore trust settings
4793 unless some valid trust or reject settings are present.
4796 *) Fix for blowfish EVP: its a variable length cipher.
4799 *) Fix various bugs related to DSA S/MIME verification. Handle missing
4800 parameters in DSA public key structures and return an error in the
4801 DSA routines if parameters are absent.
4804 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
4805 in the current directory if neither $RANDFILE nor $HOME was set.
4806 RAND_file_name() in 0.9.6a returned NULL in this case. This has
4807 caused some confusion to Windows users who haven't defined $HOME.
4808 Thus RAND_file_name() is changed again: e_os.h can define a
4809 DEFAULT_HOME, which will be used if $HOME is not set.
4810 For Windows, we use "C:"; on other platforms, we still require
4811 environment variables.
4813 *) Move 'if (!initialized) RAND_poll()' into regions protected by
4814 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
4815 having multiple threads call RAND_poll() concurrently.
4818 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
4819 combination of a flag and a thread ID variable.
4820 Otherwise while one thread is in ssleay_rand_bytes (which sets the
4821 flag), *other* threads can enter ssleay_add_bytes without obeying
4822 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
4823 that they do not hold after the first thread unsets add_do_not_lock).
4826 *) Change bctest again: '-x' expressions are not available in all
4830 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
4832 *) Fix a couple of memory leaks in PKCS7_dataDecode()
4833 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
4835 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
4836 the default extension for executables, if any. Also, make the perl
4837 scripts that use symlink() to test if it really exists and use "cp"
4838 if it doesn't. All this made OpenSSL compilable and installable in
4842 *) Fix for asn1_GetSequence() for indefinite length constructed data.
4843 If SEQUENCE is length is indefinite just set c->slen to the total
4844 amount of data available.
4845 [Steve Henson, reported by shige@FreeBSD.org]
4846 [This change does not apply to 0.9.7.]
4848 *) Change bctest to avoid here-documents inside command substitution
4849 (workaround for FreeBSD /bin/sh bug).
4850 For compatibility with Ultrix, avoid shell functions (introduced
4851 in the bctest version that searches along $PATH).
4854 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
4855 with des_encrypt() defined on some operating systems, like Solaris
4859 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
4860 On the Importance of Eliminating Errors in Cryptographic
4861 Computations, J. Cryptology 14 (2001) 2, 101-119,
4862 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
4865 *) MIPS assembler BIGNUM division bug fix.
4868 *) Disabled incorrect Alpha assembler code.
4871 *) Fix PKCS#7 decode routines so they correctly update the length
4872 after reading an EOC for the EXPLICIT tag.
4874 [This change does not apply to 0.9.7.]
4876 *) Fix bug in PKCS#12 key generation routines. This was triggered
4877 if a 3DES key was generated with a 0 initial byte. Include
4878 PKCS12_BROKEN_KEYGEN compilation option to retain the old
4879 (but broken) behaviour.
4882 *) Enhance bctest to search for a working bc along $PATH and print
4884 [Tim Rice <tim@multitalents.net> via Richard Levitte]
4886 *) Fix memory leaks in err.c: free err_data string if necessary;
4887 don't write to the wrong index in ERR_set_error_data.
4890 *) Implement ssl23_peek (analogous to ssl23_read), which previously
4894 *) Replace rdtsc with _emit statements for VC++ version 5.
4895 [Jeremy Cooper <jeremy@baymoo.org>]
4897 *) Make it possible to reuse SSLv2 sessions.
4900 *) In copy_email() check for >= 0 as a return value for
4901 X509_NAME_get_index_by_NID() since 0 is a valid index.
4902 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
4904 *) Avoid coredump with unsupported or invalid public keys by checking if
4905 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
4906 PKCS7_verify() fails with non detached data.
4909 *) Don't use getenv in library functions when run as setuid/setgid.
4910 New function OPENSSL_issetugid().
4913 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
4914 due to incorrect handling of multi-threading:
4916 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
4918 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
4920 3. Count how many times MemCheck_off() has been called so that
4921 nested use can be treated correctly. This also avoids
4922 inband-signalling in the previous code (which relied on the
4923 assumption that thread ID 0 is impossible).
4926 *) Add "-rand" option also to s_client and s_server.
4929 *) Fix CPU detection on Irix 6.x.
4930 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
4931 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4933 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
4936 [This change does not apply to 0.9.7.]
4938 *) Use the cached encoding of an X509_NAME structure rather than
4939 copying it. This is apparently the reason for the libsafe "errors"
4940 but the code is actually correct.
4943 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
4944 Bleichenbacher's DSA attack.
4945 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
4946 to be set and top=0 forces the highest bit to be set; top=-1 is new
4947 and leaves the highest bit random.
4948 [Ulf Moeller, Bodo Moeller]
4950 *) In the NCONF_...-based implementations for CONF_... queries
4951 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
4952 a temporary CONF structure with the data component set to NULL
4953 (which gives segmentation faults in lh_retrieve).
4954 Instead, use NULL for the CONF pointer in CONF_get_string and
4955 CONF_get_number (which may use environment variables) and directly
4956 return NULL from CONF_get_section.
4959 *) Fix potential buffer overrun for EBCDIC.
4962 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
4963 keyUsage if basicConstraints absent for a CA.
4966 *) Make SMIME_write_PKCS7() write mail header values with a format that
4967 is more generally accepted (no spaces before the semicolon), since
4968 some programs can't parse those values properly otherwise. Also make
4969 sure BIO's that break lines after each write do not create invalid
4973 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
4974 macros previously used would not encode an empty SEQUENCE OF
4975 and break the signature.
4977 [This change does not apply to 0.9.7.]
4979 *) Zero the premaster secret after deriving the master secret in
4983 *) Add some EVP_add_digest_alias registrations (as found in
4984 OpenSSL_add_all_digests()) to SSL_library_init()
4985 aka OpenSSL_add_ssl_algorithms(). This provides improved
4986 compatibility with peers using X.509 certificates
4987 with unconventional AlgorithmIdentifier OIDs.
4990 *) Fix for Irix with NO_ASM.
4991 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4993 *) ./config script fixes.
4994 [Ulf Moeller, Richard Levitte]
4996 *) Fix 'openssl passwd -1'.
4999 *) Change PKCS12_key_gen_asc() so it can cope with non null
5000 terminated strings whose length is passed in the passlen
5001 parameter, for example from PEM callbacks. This was done
5002 by adding an extra length parameter to asc2uni().
5003 [Steve Henson, reported by <oddissey@samsung.co.kr>]
5005 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
5006 call failed, free the DSA structure.
5009 *) Fix to uni2asc() to cope with zero length Unicode strings.
5010 These are present in some PKCS#12 files.
5013 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
5014 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
5015 when writing a 32767 byte record.
5016 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
5018 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
5019 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
5021 (RSA objects have a reference count access to which is protected
5022 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
5023 so they are meant to be shared between threads.)
5024 [Bodo Moeller, Geoff Thorpe; original patch submitted by
5025 "Reddie, Steven" <Steven.Reddie@ca.com>]
5027 *) Fix a deadlock in CRYPTO_mem_leaks().
5030 *) Use better test patterns in bntest.
5033 *) rand_win.c fix for Borland C.
5036 *) BN_rshift bugfix for n == 0.
5039 *) Add a 'bctest' script that checks for some known 'bc' bugs
5040 so that 'make test' does not abort just because 'bc' is broken.
5043 *) Store verify_result within SSL_SESSION also for client side to
5044 avoid potential security hole. (Re-used sessions on the client side
5045 always resulted in verify_result==X509_V_OK, not using the original
5046 result of the server certificate verification.)
5049 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
5050 SSL3_RT_APPLICATION_DATA, return 0.
5051 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
5055 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
5056 releases, have been re-implemented by renaming the previous
5057 implementations of ssl2_read and ssl3_read to ssl2_read_internal
5058 and ssl3_read_internal, respectively, and adding 'peek' parameters
5059 to them. The new ssl[23]_{read,peek} functions are calls to
5060 ssl[23]_read_internal with the 'peek' flag set appropriately.
5061 A 'peek' parameter has also been added to ssl3_read_bytes, which
5062 does the actual work for ssl3_read_internal.
5065 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
5066 the method-specific "init()" handler. Also clean up ex_data after
5067 calling the method-specific "finish()" handler. Previously, this was
5068 happening the other way round.
5071 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
5072 The previous value, 12, was not always sufficient for BN_mod_exp().
5075 *) Make sure that shared libraries get the internal name engine with
5076 the full version number and not just 0. This should mark the
5077 shared libraries as not backward compatible. Of course, this should
5078 be changed again when we can guarantee backward binary compatibility.
5081 *) Fix typo in get_cert_by_subject() in by_dir.c
5082 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
5084 *) Rework the system to generate shared libraries:
5086 - Make note of the expected extension for the shared libraries and
5087 if there is a need for symbolic links from for example libcrypto.so.0
5088 to libcrypto.so.0.9.7. There is extended info in Configure for
5091 - Make as few rebuilds of the shared libraries as possible.
5093 - Still avoid linking the OpenSSL programs with the shared libraries.
5095 - When installing, install the shared libraries separately from the
5099 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
5101 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
5102 and not in SSL_clear because the latter is also used by the
5103 accept/connect functions; previously, the settings made by
5104 SSL_set_read_ahead would be lost during the handshake.
5105 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
5107 *) Correct util/mkdef.pl to be selective about disabled algorithms.
5108 Previously, it would create entries for disableed algorithms no
5112 *) Added several new manual pages for SSL_* function.
5115 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
5117 *) In ssl23_get_client_hello, generate an error message when faced
5118 with an initial SSL 3.0/TLS record that is too small to contain the
5119 first two bytes of the ClientHello message, i.e. client_version.
5120 (Note that this is a pathologic case that probably has never happened
5121 in real life.) The previous approach was to use the version number
5122 from the record header as a substitute; but our protocol choice
5123 should not depend on that one because it is not authenticated
5124 by the Finished messages.
5127 *) More robust randomness gathering functions for Windows.
5128 [Jeffrey Altman <jaltman@columbia.edu>]
5130 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
5131 not set then we don't setup the error code for issuer check errors
5132 to avoid possibly overwriting other errors which the callback does
5133 handle. If an application does set the flag then we assume it knows
5134 what it is doing and can handle the new informational codes
5138 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
5139 a general "ANY" type, as such it should be able to decode anything
5140 including tagged types. However it didn't check the class so it would
5141 wrongly interpret tagged types in the same way as their universal
5142 counterpart and unknown types were just rejected. Changed so that the
5143 tagged and unknown types are handled in the same way as a SEQUENCE:
5144 that is the encoding is stored intact. There is also a new type
5145 "V_ASN1_OTHER" which is used when the class is not universal, in this
5146 case we have no idea what the actual type is so we just lump them all
5150 *) On VMS, stdout may very well lead to a file that is written to
5151 in a record-oriented fashion. That means that every write() will
5152 write a separate record, which will be read separately by the
5153 programs trying to read from it. This can be very confusing.
5155 The solution is to put a BIO filter in the way that will buffer
5156 text until a linefeed is reached, and then write everything a
5157 line at a time, so every record written will be an actual line,
5158 not chunks of lines and not (usually doesn't happen, but I've
5159 seen it once) several lines in one record. BIO_f_linebuffer() is
5162 Currently, it's a VMS-only method, because that's where it has
5163 been tested well enough.
5166 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
5167 it can return incorrect results.
5168 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
5169 but it was in 0.9.6-beta[12].)
5172 *) Disable the check for content being present when verifying detached
5173 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
5174 include zero length content when signing messages.
5177 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
5178 BIO_ctrl (for BIO pairs).
5181 *) Add DSO method for VMS.
5184 *) Bug fix: Montgomery multiplication could produce results with the
5188 *) Add RPM specification openssl.spec and modify it to build three
5189 packages. The default package contains applications, application
5190 documentation and run-time libraries. The devel package contains
5191 include files, static libraries and function documentation. The
5192 doc package contains the contents of the doc directory. The original
5193 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
5196 *) Add a large number of documentation files for many SSL routines.
5197 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5199 *) Add a configuration entry for Sony News 4.
5200 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
5202 *) Don't set the two most significant bits to one when generating a
5203 random number < q in the DSA library.
5206 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
5207 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
5208 the underlying transport is blocking) if a handshake took place.
5209 (The default behaviour is needed by applications such as s_client
5210 and s_server that use select() to determine when to use SSL_read;
5211 but for applications that know in advance when to expect data, it
5212 just makes things more complicated.)
5215 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
5219 *) Add a few more EBCDIC conditionals that make `req' and `x509'
5220 work better on such systems.
5221 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5223 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
5224 Update PKCS12_parse() so it copies the friendlyName and the
5225 keyid to the certificates aux info.
5228 *) Fix bug in PKCS7_verify() which caused an infinite loop
5229 if there was more than one signature.
5230 [Sven Uszpelkat <su@celocom.de>]
5232 *) Major change in util/mkdef.pl to include extra information
5233 about each symbol, as well as presentig variables as well
5234 as functions. This change means that there's n more need
5235 to rebuild the .num files when some algorithms are excluded.
5238 *) Allow the verify time to be set by an application,
5239 rather than always using the current time.
5242 *) Phase 2 verify code reorganisation. The certificate
5243 verify code now looks up an issuer certificate by a
5244 number of criteria: subject name, authority key id
5245 and key usage. It also verifies self signed certificates
5246 by the same criteria. The main comparison function is
5247 X509_check_issued() which performs these checks.
5249 Lot of changes were necessary in order to support this
5250 without completely rewriting the lookup code.
5252 Authority and subject key identifier are now cached.
5254 The LHASH 'certs' is X509_STORE has now been replaced
5255 by a STACK_OF(X509_OBJECT). This is mainly because an
5256 LHASH can't store or retrieve multiple objects with
5257 the same hash value.
5259 As a result various functions (which were all internal
5260 use only) have changed to handle the new X509_STORE
5261 structure. This will break anything that messed round
5262 with X509_STORE internally.
5264 The functions X509_STORE_add_cert() now checks for an
5265 exact match, rather than just subject name.
5267 The X509_STORE API doesn't directly support the retrieval
5268 of multiple certificates matching a given criteria, however
5269 this can be worked round by performing a lookup first
5270 (which will fill the cache with candidate certificates)
5271 and then examining the cache for matches. This is probably
5272 the best we can do without throwing out X509_LOOKUP
5273 entirely (maybe later...).
5275 The X509_VERIFY_CTX structure has been enhanced considerably.
5277 All certificate lookup operations now go via a get_issuer()
5278 callback. Although this currently uses an X509_STORE it
5279 can be replaced by custom lookups. This is a simple way
5280 to bypass the X509_STORE hackery necessary to make this
5281 work and makes it possible to use more efficient techniques
5282 in future. A very simple version which uses a simple
5283 STACK for its trusted certificate store is also provided
5284 using X509_STORE_CTX_trusted_stack().
5286 The verify_cb() and verify() callbacks now have equivalents
5287 in the X509_STORE_CTX structure.
5289 X509_STORE_CTX also has a 'flags' field which can be used
5290 to customise the verify behaviour.
5293 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
5294 excludes S/MIME capabilities.
5297 *) When a certificate request is read in keep a copy of the
5298 original encoding of the signed data and use it when outputing
5299 again. Signatures then use the original encoding rather than
5300 a decoded, encoded version which may cause problems if the
5301 request is improperly encoded.
5304 *) For consistency with other BIO_puts implementations, call
5305 buffer_write(b, ...) directly in buffer_puts instead of calling
5308 In BIO_puts, increment b->num_write as in BIO_write.
5309 [Peter.Sylvester@EdelWeb.fr]
5311 *) Fix BN_mul_word for the case where the word is 0. (We have to use
5312 BN_zero, we may not return a BIGNUM with an array consisting of
5316 *) Avoid calling abort() from within the library when problems are
5317 detected, except if preprocessor symbols have been defined
5318 (such as REF_CHECK, BN_DEBUG etc.).
5321 *) New openssl application 'rsautl'. This utility can be
5322 used for low level RSA operations. DER public key
5323 BIO/fp routines also added.
5326 *) New Configure entry and patches for compiling on QNX 4.
5327 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
5329 *) A demo state-machine implementation was sponsored by
5330 Nuron (http://www.nuron.com/) and is now available in
5331 demos/state_machine.
5334 *) New options added to the 'dgst' utility for signature
5335 generation and verification.
5338 *) Unrecognized PKCS#7 content types are now handled via a
5339 catch all ASN1_TYPE structure. This allows unsupported
5340 types to be stored as a "blob" and an application can
5341 encode and decode it manually.
5344 *) Fix various signed/unsigned issues to make a_strex.c
5346 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
5348 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
5349 length if passed a buffer. ASN1_INTEGER_to_BN failed
5350 if passed a NULL BN and its argument was negative.
5351 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
5353 *) Modification to PKCS#7 encoding routines to output definite
5354 length encoding. Since currently the whole structures are in
5355 memory there's not real point in using indefinite length
5356 constructed encoding. However if OpenSSL is compiled with
5357 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
5360 *) Added BIO_vprintf() and BIO_vsnprintf().
5363 *) Added more prefixes to parse for in the the strings written
5364 through a logging bio, to cover all the levels that are available
5365 through syslog. The prefixes are now:
5367 PANIC, EMERG, EMR => LOG_EMERG
5368 ALERT, ALR => LOG_ALERT
5369 CRIT, CRI => LOG_CRIT
5370 ERROR, ERR => LOG_ERR
5371 WARNING, WARN, WAR => LOG_WARNING
5372 NOTICE, NOTE, NOT => LOG_NOTICE
5373 INFO, INF => LOG_INFO
5374 DEBUG, DBG => LOG_DEBUG
5376 and as before, if none of those prefixes are present at the
5377 beginning of the string, LOG_ERR is chosen.
5379 On Win32, the LOG_* levels are mapped according to this:
5381 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
5382 LOG_WARNING => EVENTLOG_WARNING_TYPE
5383 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
5387 *) Made it possible to reconfigure with just the configuration
5388 argument "reconf" or "reconfigure". The command line arguments
5389 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
5390 and are retrieved from there when reconfiguring.
5394 [Assar Westerlund <assar@sics.se>, Richard Levitte]
5396 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
5399 *) The obj_dat.pl script was messing up the sorting of object
5400 names. The reason was that it compared the quoted version
5401 of strings as a result "OCSP" > "OCSP Signing" because
5402 " > SPACE. Changed script to store unquoted versions of
5403 names and add quotes on output. It was also omitting some
5404 names from the lookup table if they were given a default
5405 value (that is if SN is missing it is given the same
5406 value as LN and vice versa), these are now added on the
5407 grounds that if an object has a name we should be able to
5408 look it up. Finally added warning output when duplicate
5409 short or long names are found.
5412 *) Changes needed for Tandem NSK.
5413 [Scott Uroff <scott@xypro.com>]
5415 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
5416 RSA_padding_check_SSLv23(), special padding was never detected
5417 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
5418 version rollback attacks was not effective.
5420 In s23_clnt.c, don't use special rollback-attack detection padding
5421 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
5422 client; similarly, in s23_srvr.c, don't do the rollback check if
5423 SSL 2.0 is the only protocol enabled in the server.
5426 *) Make it possible to get hexdumps of unprintable data with 'openssl
5427 asn1parse'. By implication, the functions ASN1_parse_dump() and
5428 BIO_dump_indent() are added.
5431 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
5432 these print out strings and name structures based on various
5433 flags including RFC2253 support and proper handling of
5434 multibyte characters. Added options to the 'x509' utility
5435 to allow the various flags to be set.
5438 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
5439 Also change the functions X509_cmp_current_time() and
5440 X509_gmtime_adj() work with an ASN1_TIME structure,
5441 this will enable certificates using GeneralizedTime in validity
5442 dates to be checked.
5445 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
5446 negative public key encodings) on by default,
5447 NO_NEG_PUBKEY_BUG can be set to disable it.
5450 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
5451 content octets. An i2c_ASN1_OBJECT is unnecessary because
5452 the encoding can be trivially obtained from the structure.
5455 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
5456 not read locks (CRYPTO_r_[un]lock).
5459 *) A first attempt at creating official support for shared
5460 libraries through configuration. I've kept it so the
5461 default is static libraries only, and the OpenSSL programs
5462 are always statically linked for now, but there are
5463 preparations for dynamic linking in place.
5464 This has been tested on Linux and Tru64.
5467 *) Randomness polling function for Win9x, as described in:
5468 Peter Gutmann, Software Generation of Practically Strong
5472 *) Fix so PRNG is seeded in req if using an already existing
5476 *) New options to smime application. -inform and -outform
5477 allow alternative formats for the S/MIME message including
5478 PEM and DER. The -content option allows the content to be
5479 specified separately. This should allow things like Netscape
5480 form signing output easier to verify.
5483 *) Fix the ASN1 encoding of tags using the 'long form'.
5486 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
5487 STRING types. These convert content octets to and from the
5488 underlying type. The actual tag and length octets are
5489 already assumed to have been read in and checked. These
5490 are needed because all other string types have virtually
5491 identical handling apart from the tag. By having versions
5492 of the ASN1 functions that just operate on content octets
5493 IMPLICIT tagging can be handled properly. It also allows
5494 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
5495 and ASN1_INTEGER are identical apart from the tag.
5498 *) Change the handling of OID objects as follows:
5500 - New object identifiers are inserted in objects.txt, following
5501 the syntax given in objects.README.
5502 - objects.pl is used to process obj_mac.num and create a new
5504 - obj_dat.pl is used to create a new obj_dat.h, using the data in
5507 This is currently kind of a hack, and the perl code in objects.pl
5508 isn't very elegant, but it works as I intended. The simplest way
5509 to check that it worked correctly is to look in obj_dat.h and
5510 check the array nid_objs and make sure the objects haven't moved
5511 around (this is important!). Additions are OK, as well as
5512 consistent name changes.
5515 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
5518 *) Addition of the command line parameter '-rand file' to 'openssl req'.
5519 The given file adds to whatever has already been seeded into the
5520 random pool through the RANDFILE configuration file option or
5521 environment variable, or the default random state file.
5524 *) mkstack.pl now sorts each macro group into lexical order.
5525 Previously the output order depended on the order the files
5526 appeared in the directory, resulting in needless rewriting
5530 *) Patches to make OpenSSL compile under Win32 again. Mostly
5531 work arounds for the VC++ problem that it treats func() as
5532 func(void). Also stripped out the parts of mkdef.pl that
5533 added extra typesafe functions: these no longer exist.
5536 *) Reorganisation of the stack code. The macros are now all
5537 collected in safestack.h . Each macro is defined in terms of
5538 a "stack macro" of the form SKM_<name>(type, a, b). The
5539 DEBUG_SAFESTACK is now handled in terms of function casts,
5540 this has the advantage of retaining type safety without the
5541 use of additional functions. If DEBUG_SAFESTACK is not defined
5542 then the non typesafe macros are used instead. Also modified the
5543 mkstack.pl script to handle the new form. Needs testing to see
5544 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
5545 the default if no major problems. Similar behaviour for ASN1_SET_OF
5546 and PKCS12_STACK_OF.
5549 *) When some versions of IIS use the 'NET' form of private key the
5550 key derivation algorithm is different. Normally MD5(password) is
5551 used as a 128 bit RC4 key. In the modified case
5552 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
5553 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
5554 as the old Netscape_RSA functions except they have an additional
5555 'sgckey' parameter which uses the modified algorithm. Also added
5556 an -sgckey command line option to the rsa utility. Thanks to
5557 Adrian Peck <bertie@ncipher.com> for posting details of the modified
5558 algorithm to openssl-dev.
5561 *) The evp_local.h macros were using 'c.##kname' which resulted in
5562 invalid expansion on some systems (SCO 5.0.5 for example).
5563 Corrected to 'c.kname'.
5564 [Phillip Porch <root@theporch.com>]
5566 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
5567 a STACK of email addresses from a certificate or request, these look
5568 in the subject name and the subject alternative name extensions and
5569 omit any duplicate addresses.
5572 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
5573 This makes DSA verification about 2 % faster.
5576 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
5577 (meaning that now 2^5 values will be precomputed, which is only 4 KB
5578 plus overhead for 1024 bit moduli).
5579 This makes exponentiations about 0.5 % faster for 1024 bit
5580 exponents (as measured by "openssl speed rsa2048").
5583 *) Rename memory handling macros to avoid conflicts with other
5585 Malloc => OPENSSL_malloc
5586 Malloc_locked => OPENSSL_malloc_locked
5587 Realloc => OPENSSL_realloc
5588 Free => OPENSSL_free
5591 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
5592 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
5595 *) CygWin32 support.
5596 [John Jarvie <jjarvie@newsguy.com>]
5598 *) The type-safe stack code has been rejigged. It is now only compiled
5599 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
5600 by default all type-specific stack functions are "#define"d back to
5601 standard stack functions. This results in more streamlined output
5602 but retains the type-safety checking possibilities of the original
5606 *) The STACK code has been cleaned up, and certain type declarations
5607 that didn't make a lot of sense have been brought in line. This has
5608 also involved a cleanup of sorts in safestack.h to more correctly
5609 map type-safe stack functions onto their plain stack counterparts.
5610 This work has also resulted in a variety of "const"ifications of
5611 lots of the code, especially "_cmp" operations which should normally
5612 be prototyped with "const" parameters anyway.
5615 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
5616 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
5617 (The PRNG state consists of two parts, the large pool 'state' and 'md',
5618 where all of 'md' is used each time the PRNG is used, but 'state'
5619 is used only indexed by a cyclic counter. As entropy may not be
5620 well distributed from the beginning, 'md' is important as a
5621 chaining variable. However, the output function chains only half
5622 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
5623 all of 'md', and seeding with STATE_SIZE dummy bytes will result
5624 in all of 'state' being rewritten, with the new values depending
5625 on virtually all of 'md'. This overcomes the 80 bit limitation.)
5628 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
5629 the handshake is continued after ssl_verify_cert_chain();
5630 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
5631 can lead to 'unexplainable' connection aborts later.
5632 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
5634 *) Major EVP API cipher revision.
5635 Add hooks for extra EVP features. This allows various cipher
5636 parameters to be set in the EVP interface. Support added for variable
5637 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
5638 setting of RC2 and RC5 parameters.
5640 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
5643 Remove lots of duplicated code from the EVP library. For example *every*
5644 cipher init() function handles the 'iv' in the same way according to the
5645 cipher mode. They also all do nothing if the 'key' parameter is NULL and
5646 for CFB and OFB modes they zero ctx->num.
5648 New functionality allows removal of S/MIME code RC2 hack.
5650 Most of the routines have the same form and so can be declared in terms
5653 By shifting this to the top level EVP_CipherInit() it can be removed from
5654 all individual ciphers. If the cipher wants to handle IVs or keys
5655 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
5658 Change lots of functions like EVP_EncryptUpdate() to now return a
5659 value: although software versions of the algorithms cannot fail
5660 any installed hardware versions can.
5663 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
5664 this option is set, tolerate broken clients that send the negotiated
5665 protocol version number instead of the requested protocol version
5669 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
5670 i.e. non-zero for export ciphersuites, zero otherwise.
5671 Previous versions had this flag inverted, inconsistent with
5672 rsa_tmp_cb (..._TMP_RSA_CB).
5673 [Bodo Moeller; problem reported by Amit Chopra]
5675 *) Add missing DSA library text string. Work around for some IIS
5676 key files with invalid SEQUENCE encoding.
5679 *) Add a document (doc/standards.txt) that list all kinds of standards
5680 and so on that are implemented in OpenSSL.
5683 *) Enhance c_rehash script. Old version would mishandle certificates
5684 with the same subject name hash and wouldn't handle CRLs at all.
5685 Added -fingerprint option to crl utility, to support new c_rehash
5689 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
5692 *) Fix for SSL server purpose checking. Server checking was
5693 rejecting certificates which had extended key usage present
5694 but no ssl client purpose.
5695 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
5697 *) Make PKCS#12 code work with no password. The PKCS#12 spec
5698 is a little unclear about how a blank password is handled.
5699 Since the password in encoded as a BMPString with terminating
5700 double NULL a zero length password would end up as just the
5701 double NULL. However no password at all is different and is
5702 handled differently in the PKCS#12 key generation code. NS
5703 treats a blank password as zero length. MSIE treats it as no
5704 password on export: but it will try both on import. We now do
5705 the same: PKCS12_parse() tries zero length and no password if
5706 the password is set to "" or NULL (NULL is now a valid password:
5707 it wasn't before) as does the pkcs12 application.
5710 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
5711 perror when PEM_read_bio_X509_REQ fails, the error message must
5712 be obtained from the error queue.
5715 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
5716 it in ERR_remove_state if appropriate, and change ERR_get_state
5717 accordingly to avoid race conditions (this is necessary because
5718 thread_hash is no longer constant once set).
5721 *) Bugfix for linux-elf makefile.one.
5724 *) RSA_get_default_method() will now cause a default
5725 RSA_METHOD to be chosen if one doesn't exist already.
5726 Previously this was only set during a call to RSA_new()
5727 or RSA_new_method(NULL) meaning it was possible for
5728 RSA_get_default_method() to return NULL.
5731 *) Added native name translation to the existing DSO code
5732 that will convert (if the flag to do so is set) filenames
5733 that are sufficiently small and have no path information
5734 into a canonical native form. Eg. "blah" converted to
5735 "libblah.so" or "blah.dll" etc.
5738 *) New function ERR_error_string_n(e, buf, len) which is like
5739 ERR_error_string(e, buf), but writes at most 'len' bytes
5740 including the 0 terminator. For ERR_error_string_n, 'buf'
5742 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
5744 *) CONF library reworked to become more general. A new CONF
5745 configuration file reader "class" is implemented as well as a
5746 new functions (NCONF_*, for "New CONF") to handle it. The now
5747 old CONF_* functions are still there, but are reimplemented to
5748 work in terms of the new functions. Also, a set of functions
5749 to handle the internal storage of the configuration data is
5750 provided to make it easier to write new configuration file
5751 reader "classes" (I can definitely see something reading a
5752 configuration file in XML format, for example), called _CONF_*,
5753 or "the configuration storage API"...
5755 The new configuration file reading functions are:
5757 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
5758 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
5760 NCONF_default, NCONF_WIN32
5762 NCONF_dump_fp, NCONF_dump_bio
5764 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
5765 NCONF_new creates a new CONF object. This works in the same way
5766 as other interfaces in OpenSSL, like the BIO interface.
5767 NCONF_dump_* dump the internal storage of the configuration file,
5768 which is useful for debugging. All other functions take the same
5769 arguments as the old CONF_* functions wth the exception of the
5770 first that must be a `CONF *' instead of a `LHASH *'.
5772 To make it easer to use the new classes with the old CONF_* functions,
5773 the function CONF_set_default_method is provided.
5776 *) Add '-tls1' option to 'openssl ciphers', which was already
5777 mentioned in the documentation but had not been implemented.
5778 (This option is not yet really useful because even the additional
5779 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
5782 *) Initial DSO code added into libcrypto for letting OpenSSL (and
5783 OpenSSL-based applications) load shared libraries and bind to
5784 them in a portable way.
5785 [Geoff Thorpe, with contributions from Richard Levitte]
5787 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
5789 *) Make sure _lrotl and _lrotr are only used with MSVC.
5791 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
5792 (the default implementation of RAND_status).
5794 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
5795 to '-clrext' (= clear extensions), as intended and documented.
5796 [Bodo Moeller; inconsistency pointed out by Michael Attili
5799 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
5800 was larger than the MD block size.
5801 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
5803 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
5804 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
5805 using the passed key: if the passed key was a private key the result
5806 of X509_print(), for example, would be to print out all the private key
5810 *) des_quad_cksum() byte order bug fix.
5811 [Ulf Möller, using the problem description in krb4-0.9.7, where
5812 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
5814 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
5816 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
5818 *) For easily testing in shell scripts whether some command
5819 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
5820 returns with exit code 0 iff no command of the given name is available.
5821 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
5822 the output goes to stdout and nothing is printed to stderr.
5823 Additional arguments are always ignored.
5825 Since for each cipher there is a command of the same name,
5826 the 'no-cipher' compilation switches can be tested this way.
5828 ('openssl no-XXX' is not able to detect pseudo-commands such
5829 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
5832 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
5835 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
5836 is set; it will be thrown away anyway because each handshake creates
5838 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
5839 to parameters -- in previous versions (since OpenSSL 0.9.3) the
5840 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
5841 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
5844 *) New s_client option -ign_eof: EOF at stdin is ignored, and
5845 'Q' and 'R' lose their special meanings (quit/renegotiate).
5846 This is part of what -quiet does; unlike -quiet, -ign_eof
5847 does not suppress any output.
5850 *) Add compatibility options to the purpose and trust code. The
5851 purpose X509_PURPOSE_ANY is "any purpose" which automatically
5852 accepts a certificate or CA, this was the previous behaviour,
5853 with all the associated security issues.
5855 X509_TRUST_COMPAT is the old trust behaviour: only and
5856 automatically trust self signed roots in certificate store. A
5857 new trust setting X509_TRUST_DEFAULT is used to specify that
5858 a purpose has no associated trust setting and it should instead
5859 use the value in the default purpose.
5862 *) Fix the PKCS#8 DSA private key code so it decodes keys again
5863 and fix a memory leak.
5866 *) In util/mkerr.pl (which implements 'make errors'), preserve
5867 reason strings from the previous version of the .c file, as
5868 the default to have only downcase letters (and digits) in
5869 automatically generated reasons codes is not always appropriate.
5872 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
5873 using strerror. Previously, ERR_reason_error_string() returned
5874 library names as reason strings for SYSerr; but SYSerr is a special
5875 case where small numbers are errno values, not library numbers.
5878 *) Add '-dsaparam' option to 'openssl dhparam' application. This
5879 converts DSA parameters into DH parameters. (When creating parameters,
5880 DSA_generate_parameters is used.)
5883 *) Include 'length' (recommended exponent length) in C code generated
5884 by 'openssl dhparam -C'.
5887 *) The second argument to set_label in perlasm was already being used
5888 so couldn't be used as a "file scope" flag. Moved to third argument
5892 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
5893 instead of RAND_bytes for encryption IVs and salts.
5896 *) Include RAND_status() into RAND_METHOD instead of implementing
5897 it only for md_rand.c Otherwise replacing the PRNG by calling
5898 RAND_set_rand_method would be impossible.
5901 *) Don't let DSA_generate_key() enter an infinite loop if the random
5902 number generation fails.
5905 *) New 'rand' application for creating pseudo-random output.
5908 *) Added configuration support for Linux/IA64
5909 [Rolf Haberrecker <rolf@suse.de>]
5911 *) Assembler module support for Mingw32.
5914 *) Shared library support for HPUX (in shlib/).
5915 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
5917 *) Shared library support for Solaris gcc.
5918 [Lutz Behnke <behnke@trustcenter.de>]
5920 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
5922 *) PKCS7_encrypt() was adding text MIME headers twice because they
5923 were added manually and by SMIME_crlf_copy().
5926 *) In bntest.c don't call BN_rand with zero bits argument.
5927 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
5929 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
5930 case was implemented. This caused BN_div_recp() to fail occasionally.
5933 *) Add an optional second argument to the set_label() in the perl
5934 assembly language builder. If this argument exists and is set
5935 to 1 it signals that the assembler should use a symbol whose
5936 scope is the entire file, not just the current function. This
5937 is needed with MASM which uses the format label:: for this scope.
5938 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
5940 *) Change the ASN1 types so they are typedefs by default. Before
5941 almost all types were #define'd to ASN1_STRING which was causing
5942 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
5946 *) Change names of new functions to the new get1/get0 naming
5947 convention: After 'get1', the caller owns a reference count
5948 and has to call ..._free; 'get0' returns a pointer to some
5949 data structure without incrementing reference counters.
5950 (Some of the existing 'get' functions increment a reference
5951 counter, some don't.)
5952 Similarly, 'set1' and 'add1' functions increase reference
5953 counters or duplicate objects.
5956 *) Allow for the possibility of temp RSA key generation failure:
5957 the code used to assume it always worked and crashed on failure.
5960 *) Fix potential buffer overrun problem in BIO_printf().
5961 [Ulf Möller, using public domain code by Patrick Powell; problem
5962 pointed out by David Sacerdote <das33@cornell.edu>]
5964 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
5965 RAND_egd() and RAND_status(). In the command line application,
5966 the EGD socket can be specified like a seed file using RANDFILE
5970 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
5971 Some CAs (e.g. Verisign) distribute certificates in this form.
5974 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
5975 list to exclude them. This means that no special compilation option
5976 is needed to use anonymous DH: it just needs to be included in the
5980 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
5981 EVP_MD_type. The old functionality is available in a new macro called
5982 EVP_MD_md(). Change code that uses it and update docs.
5985 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
5986 where the 'void *' argument is replaced by a function pointer argument.
5987 Previously 'void *' was abused to point to functions, which works on
5988 many platforms, but is not correct. As these functions are usually
5989 called by macros defined in OpenSSL header files, most source code
5990 should work without changes.
5993 *) <openssl/opensslconf.h> (which is created by Configure) now contains
5994 sections with information on -D... compiler switches used for
5995 compiling the library so that applications can see them. To enable
5996 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
5997 must be defined. E.g.,
5998 #define OPENSSL_ALGORITHM_DEFINES
5999 #include <openssl/opensslconf.h>
6000 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
6001 [Richard Levitte, Ulf and Bodo Möller]
6003 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
6007 *) Change the 'other' type in certificate aux info to a STACK_OF
6008 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
6009 the required ASN1 format: arbitrary types determined by an OID.
6012 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
6013 argument to 'req'. This is not because the function is newer or
6014 better than others it just uses the work 'NEW' in the certificate
6015 request header lines. Some software needs this.
6018 *) Reorganise password command line arguments: now passwords can be
6019 obtained from various sources. Delete the PEM_cb function and make
6020 it the default behaviour: i.e. if the callback is NULL and the
6021 usrdata argument is not NULL interpret it as a null terminated pass
6022 phrase. If usrdata and the callback are NULL then the pass phrase
6023 is prompted for as usual.
6026 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
6027 the support is automatically enabled. The resulting binaries will
6028 autodetect the card and use it if present.
6029 [Ben Laurie and Compaq Inc.]
6031 *) Work around for Netscape hang bug. This sends certificate request
6032 and server done in one record. Since this is perfectly legal in the
6033 SSL/TLS protocol it isn't a "bug" option and is on by default. See
6034 the bugs/SSLv3 entry for more info.
6037 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
6040 *) Add -rand argument to smime and pkcs12 applications and read/write
6044 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
6047 *) Add command line password options to the remaining applications.
6050 *) Bug fix for BN_div_recp() for numerators with an even number of
6054 *) More tests in bntest.c, and changed test_bn output.
6057 *) ./config recognizes MacOS X now.
6060 *) Bug fix for BN_div() when the first words of num and divsor are
6061 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
6064 *) Add support for various broken PKCS#8 formats, and command line
6065 options to produce them.
6068 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
6069 get temporary BIGNUMs from a BN_CTX.
6072 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
6076 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
6077 include a #define from the old name to the new. The original intent
6078 was that statically linked binaries could for example just call
6079 SSLeay_add_all_ciphers() to just add ciphers to the table and not
6080 link with digests. This never worked becayse SSLeay_add_all_digests()
6081 and SSLeay_add_all_ciphers() were in the same source file so calling
6082 one would link with the other. They are now in separate source files.
6085 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
6088 *) Use a less unusual form of the Miller-Rabin primality test (it used
6089 a binary algorithm for exponentiation integrated into the Miller-Rabin
6090 loop, our standard modexp algorithms are faster).
6093 *) Support for the EBCDIC character set completed.
6094 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
6096 *) Source code cleanups: use const where appropriate, eliminate casts,
6097 use void * instead of char * in lhash.
6100 *) Bugfix: ssl3_send_server_key_exchange was not restartable
6101 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
6102 this the server could overwrite ephemeral keys that the client
6106 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
6107 using 50 iterations of the Rabin-Miller test.
6109 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
6110 iterations of the Rabin-Miller test as required by the appendix
6111 to FIPS PUB 186[-1]) instead of DSA_is_prime.
6112 As BN_is_prime_fasttest includes trial division, DSA parameter
6113 generation becomes much faster.
6115 This implies a change for the callback functions in DSA_is_prime
6116 and DSA_generate_parameters: The callback function is called once
6117 for each positive witness in the Rabin-Miller test, not just
6118 occasionally in the inner loop; and the parameters to the
6119 callback function now provide an iteration count for the outer
6120 loop rather than for the current invocation of the inner loop.
6121 DSA_generate_parameters additionally can call the callback
6122 function with an 'iteration count' of -1, meaning that a
6123 candidate has passed the trial division test (when q is generated
6124 from an application-provided seed, trial division is skipped).
6127 *) New function BN_is_prime_fasttest that optionally does trial
6128 division before starting the Rabin-Miller test and has
6129 an additional BN_CTX * argument (whereas BN_is_prime always
6130 has to allocate at least one BN_CTX).
6131 'callback(1, -1, cb_arg)' is called when a number has passed the
6132 trial division stage.
6135 *) Fix for bug in CRL encoding. The validity dates weren't being handled
6139 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
6142 *) New function BN_pseudo_rand().
6145 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
6146 bignum version of BN_from_montgomery() with the working code from
6147 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
6151 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
6152 made it impossible to use the same SSL_SESSION data structure in
6153 SSL2 clients in multiple threads.
6156 *) The return value of RAND_load_file() no longer counts bytes obtained
6157 by stat(). RAND_load_file(..., -1) is new and uses the complete file
6158 to seed the PRNG (previously an explicit byte count was required).
6159 [Ulf Möller, Bodo Möller]
6161 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
6162 used (char *) instead of (void *) and had casts all over the place.
6165 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
6168 *) Retain source code compatibility for BN_prime_checks macro:
6169 BN_is_prime(..., BN_prime_checks, ...) now uses
6170 BN_prime_checks_for_size to determine the appropriate number of
6171 Rabin-Miller iterations.
6174 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
6175 DH_CHECK_P_NOT_SAFE_PRIME.
6176 (Check if this is true? OpenPGP calls them "strong".)
6179 *) Merge the functionality of "dh" and "gendh" programs into a new program
6180 "dhparam". The old programs are retained for now but will handle DH keys
6181 (instead of parameters) in future.
6184 *) Make the ciphers, s_server and s_client programs check the return values
6185 when a new cipher list is set.
6188 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
6189 ciphers. Before when the 56bit ciphers were enabled the sorting was
6192 The syntax for the cipher sorting has been extended to support sorting by
6193 cipher-strength (using the strength_bits hard coded in the tables).
6194 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
6196 Fix a bug in the cipher-command parser: when supplying a cipher command
6197 string with an "undefined" symbol (neither command nor alphanumeric
6198 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
6199 an error is flagged.
6201 Due to the strength-sorting extension, the code of the
6202 ssl_create_cipher_list() function was completely rearranged. I hope that
6203 the readability was also increased :-)
6204 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6206 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
6207 for the first serial number and places 2 in the serial number file. This
6208 avoids problems when the root CA is created with serial number zero and
6209 the first user certificate has the same issuer name and serial number
6213 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
6214 the new code. Add documentation for this stuff.
6217 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
6218 X509_*() to X509at_*() on the grounds that they don't handle X509
6219 structures and behave in an analagous way to the X509v3 functions:
6220 they shouldn't be called directly but wrapper functions should be used
6223 So we also now have some wrapper functions that call the X509at functions
6224 when passed certificate requests. (TO DO: similar things can be done with
6225 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
6226 things. Some of these need some d2i or i2d and print functionality
6227 because they handle more complex structures.)
6230 *) Add missing #ifndefs that caused missing symbols when building libssl
6231 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
6232 NO_RSA in ssl/s2*.c.
6233 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
6235 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
6236 has a return value which indicates the quality of the random data
6237 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
6238 error queue. New function RAND_pseudo_bytes() generates output that is
6239 guaranteed to be unique but not unpredictable. RAND_add is like
6240 RAND_seed, but takes an extra argument for an entropy estimate
6241 (RAND_seed always assumes full entropy).
6244 *) Do more iterations of Rabin-Miller probable prime test (specifically,
6245 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
6246 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
6247 in crypto/bn/bn_prime.c for the complete table). This guarantees a
6248 false-positive rate of at most 2^-80 for random input.
6251 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
6254 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
6255 in the 0.9.5 release), this returns the chain
6256 from an X509_CTX structure with a dup of the stack and all
6257 the X509 reference counts upped: so the stack will exist
6258 after X509_CTX_cleanup() has been called. Modify pkcs12.c
6261 Also make SSL_SESSION_print() print out the verify return
6265 *) Add manpage for the pkcs12 command. Also change the default
6266 behaviour so MAC iteration counts are used unless the new
6267 -nomaciter option is used. This improves file security and
6268 only older versions of MSIE (4.0 for example) need it.
6271 *) Honor the no-xxx Configure options when creating .DEF files.
6274 *) Add PKCS#10 attributes to field table: challengePassword,
6275 unstructuredName and unstructuredAddress. These are taken from
6276 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
6277 international characters are used.
6279 More changes to X509_ATTRIBUTE code: allow the setting of types
6280 based on strings. Remove the 'loc' parameter when adding
6281 attributes because these will be a SET OF encoding which is sorted
6285 *) Initial changes to the 'req' utility to allow request generation
6286 automation. This will allow an application to just generate a template
6287 file containing all the field values and have req construct the
6290 Initial support for X509_ATTRIBUTE handling. Stacks of these are
6291 used all over the place including certificate requests and PKCS#7
6292 structures. They are currently handled manually where necessary with
6293 some primitive wrappers for PKCS#7. The new functions behave in a
6294 manner analogous to the X509 extension functions: they allow
6295 attributes to be looked up by NID and added.
6297 Later something similar to the X509V3 code would be desirable to
6298 automatically handle the encoding, decoding and printing of the
6299 more complex types. The string types like challengePassword can
6300 be handled by the string table functions.
6302 Also modified the multi byte string table handling. Now there is
6303 a 'global mask' which masks out certain types. The table itself
6304 can use the flag STABLE_NO_MASK to ignore the mask setting: this
6305 is useful when for example there is only one permissible type
6306 (as in countryName) and using the mask might result in no valid
6310 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
6311 SSL_get_peer_finished to allow applications to obtain the latest
6312 Finished messages sent to the peer or expected from the peer,
6313 respectively. (SSL_get_peer_finished is usually the Finished message
6314 actually received from the peer, otherwise the protocol will be aborted.)
6316 As the Finished message are message digests of the complete handshake
6317 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
6318 be used for external authentication procedures when the authentication
6319 provided by SSL/TLS is not desired or is not enough.
6322 *) Enhanced support for Alpha Linux is added. Now ./config checks if
6323 the host supports BWX extension and if Compaq C is present on the
6324 $PATH. Just exploiting of the BWX extension results in 20-30%
6325 performance kick for some algorithms, e.g. DES and RC4 to mention
6326 a couple. Compaq C in turn generates ~20% faster code for MD5 and
6330 *) Add support for MS "fast SGC". This is arguably a violation of the
6331 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
6332 weak crypto and after checking the certificate is SGC a second one
6333 with strong crypto. MS SGC stops the first handshake after receiving
6334 the server certificate message and sends a second client hello. Since
6335 a server will typically do all the time consuming operations before
6336 expecting any further messages from the client (server key exchange
6337 is the most expensive) there is little difference between the two.
6339 To get OpenSSL to support MS SGC we have to permit a second client
6340 hello message after we have sent server done. In addition we have to
6341 reset the MAC if we do get this second client hello.
6344 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
6345 if a DER encoded private key is RSA or DSA traditional format. Changed
6346 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
6347 format DER encoded private key. Newer code should use PKCS#8 format which
6348 has the key type encoded in the ASN1 structure. Added DER private key
6349 support to pkcs8 application.
6352 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
6353 ciphersuites has been selected (as required by the SSL 3/TLS 1
6354 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
6355 is set, we interpret this as a request to violate the specification
6356 (the worst that can happen is a handshake failure, and 'correct'
6357 behaviour would result in a handshake failure anyway).
6360 *) In SSL_CTX_add_session, take into account that there might be multiple
6361 SSL_SESSION structures with the same session ID (e.g. when two threads
6362 concurrently obtain them from an external cache).
6363 The internal cache can handle only one SSL_SESSION with a given ID,
6364 so if there's a conflict, we now throw out the old one to achieve
6368 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
6369 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
6370 some routines that use cipher OIDs: some ciphers do not have OIDs
6371 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
6375 *) Simplify the trust setting structure and code. Now we just have
6376 two sequences of OIDs for trusted and rejected settings. These will
6377 typically have values the same as the extended key usage extension
6378 and any application specific purposes.
6380 The trust checking code now has a default behaviour: it will just
6381 check for an object with the same NID as the passed id. Functions can
6382 be provided to override either the default behaviour or the behaviour
6383 for a given id. SSL client, server and email already have functions
6384 in place for compatibility: they check the NID and also return "trusted"
6385 if the certificate is self signed.
6388 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
6389 traditional format into an EVP_PKEY structure.
6392 *) Add a password callback function PEM_cb() which either prompts for
6393 a password if usr_data is NULL or otherwise assumes it is a null
6394 terminated password. Allow passwords to be passed on command line
6395 environment or config files in a few more utilities.
6398 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
6399 keys. Add some short names for PKCS#8 PBE algorithms and allow them
6400 to be specified on the command line for the pkcs8 and pkcs12 utilities.
6401 Update documentation.
6404 *) Support for ASN1 "NULL" type. This could be handled before by using
6405 ASN1_TYPE but there wasn't any function that would try to read a NULL
6406 and produce an error if it couldn't. For compatibility we also have
6407 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
6408 don't allocate anything because they don't need to.
6411 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
6413 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
6415 *) Rebuild of the memory allocation routines used by OpenSSL code and
6416 possibly others as well. The purpose is to make an interface that
6417 provide hooks so anyone can build a separate set of allocation and
6418 deallocation routines to be used by OpenSSL, for example memory
6419 pool implementations, or something else, which was previously hard
6420 since Malloc(), Realloc() and Free() were defined as macros having
6421 the values malloc, realloc and free, respectively (except for Win32
6422 compilations). The same is provided for memory debugging code.
6423 OpenSSL already comes with functionality to find memory leaks, but
6424 this gives people a chance to debug other memory problems.
6426 With these changes, a new set of functions and macros have appeared:
6428 CRYPTO_set_mem_debug_functions() [F]
6429 CRYPTO_get_mem_debug_functions() [F]
6430 CRYPTO_dbg_set_options() [F]
6431 CRYPTO_dbg_get_options() [F]
6432 CRYPTO_malloc_debug_init() [M]
6434 The memory debug functions are NULL by default, unless the library
6435 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
6436 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
6437 gives the standard debugging functions that come with OpenSSL) or
6438 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
6439 provided by the library user) must be used. When the standard
6440 debugging functions are used, CRYPTO_dbg_set_options can be used to
6441 request additional information:
6442 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
6443 the CRYPTO_MDEBUG_xxx macro when compiling the library.
6445 Also, things like CRYPTO_set_mem_functions will always give the
6446 expected result (the new set of functions is used for allocation
6447 and deallocation) at all times, regardless of platform and compiler
6450 To finish it up, some functions that were never use in any other
6451 way than through macros have a new API and new semantic:
6454 CRYPTO_dbg_realloc()
6457 All macros of value have retained their old syntax.
6458 [Richard Levitte and Bodo Moeller]
6460 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
6461 ordering of SMIMECapabilities wasn't in "strength order" and there
6462 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
6466 *) Some ASN1 types with illegal zero length encoding (INTEGER,
6467 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
6468 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
6470 *) Merge in my S/MIME library for OpenSSL. This provides a simple
6471 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
6472 functionality to handle multipart/signed properly) and a utility
6473 called 'smime' to call all this stuff. This is based on code I
6474 originally wrote for Celo who have kindly allowed it to be
6475 included in OpenSSL.
6478 *) Add variants des_set_key_checked and des_set_key_unchecked of
6479 des_set_key (aka des_key_sched). Global variable des_check_key
6480 decides which of these is called by des_set_key; this way
6481 des_check_key behaves as it always did, but applications and
6482 the library itself, which was buggy for des_check_key == 1,
6483 have a cleaner way to pick the version they need.
6486 *) New function PKCS12_newpass() which changes the password of a
6490 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
6491 dynamic mix. In both cases the ids can be used as an index into the
6492 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
6493 functions so they accept a list of the field values and the
6494 application doesn't need to directly manipulate the X509_TRUST
6498 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
6502 *) Modify the way the V3 extension code looks up extensions. This now
6503 works in a similar way to the object code: we have some "standard"
6504 extensions in a static table which is searched with OBJ_bsearch()
6505 and the application can add dynamic ones if needed. The file
6506 crypto/x509v3/ext_dat.h now has the info: this file needs to be
6507 updated whenever a new extension is added to the core code and kept
6508 in ext_nid order. There is a simple program 'tabtest.c' which checks
6509 this. New extensions are not added too often so this file can readily
6510 be maintained manually.
6512 There are two big advantages in doing things this way. The extensions
6513 can be looked up immediately and no longer need to be "added" using
6514 X509V3_add_standard_extensions(): this function now does nothing.
6515 [Side note: I get *lots* of email saying the extension code doesn't
6516 work because people forget to call this function]
6517 Also no dynamic allocation is done unless new extensions are added:
6518 so if we don't add custom extensions there is no need to call
6519 X509V3_EXT_cleanup().
6522 *) Modify enc utility's salting as follows: make salting the default. Add a
6523 magic header, so unsalted files fail gracefully instead of just decrypting
6524 to garbage. This is because not salting is a big security hole, so people
6525 should be discouraged from doing it.
6528 *) Fixes and enhancements to the 'x509' utility. It allowed a message
6529 digest to be passed on the command line but it only used this
6530 parameter when signing a certificate. Modified so all relevant
6531 operations are affected by the digest parameter including the
6532 -fingerprint and -x509toreq options. Also -x509toreq choked if a
6533 DSA key was used because it didn't fix the digest.
6536 *) Initial certificate chain verify code. Currently tests the untrusted
6537 certificates for consistency with the verify purpose (which is set
6538 when the X509_STORE_CTX structure is set up) and checks the pathlength.
6540 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
6541 this is because it will reject chains with invalid extensions whereas
6542 every previous version of OpenSSL and SSLeay made no checks at all.
6544 Trust code: checks the root CA for the relevant trust settings. Trust
6545 settings have an initial value consistent with the verify purpose: e.g.
6546 if the verify purpose is for SSL client use it expects the CA to be
6547 trusted for SSL client use. However the default value can be changed to
6548 permit custom trust settings: one example of this would be to only trust
6549 certificates from a specific "secure" set of CAs.
6551 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
6552 which should be used for version portability: especially since the
6553 verify structure is likely to change more often now.
6555 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
6556 to set them. If not set then assume SSL clients will verify SSL servers
6559 Two new options to the verify program: -untrusted allows a set of
6560 untrusted certificates to be passed in and -purpose which sets the
6561 intended purpose of the certificate. If a purpose is set then the
6562 new chain verify code is used to check extension consistency.
6565 *) Support for the authority information access extension.
6568 *) Modify RSA and DSA PEM read routines to transparently handle
6569 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
6570 public keys in a format compatible with certificate
6571 SubjectPublicKeyInfo structures. Unfortunately there were already
6572 functions called *_PublicKey_* which used various odd formats so
6573 these are retained for compatibility: however the DSA variants were
6574 never in a public release so they have been deleted. Changed dsa/rsa
6575 utilities to handle the new format: note no releases ever handled public
6576 keys so we should be OK.
6578 The primary motivation for this change is to avoid the same fiasco
6579 that dogs private keys: there are several incompatible private key
6580 formats some of which are standard and some OpenSSL specific and
6581 require various evil hacks to allow partial transparent handling and
6582 even then it doesn't work with DER formats. Given the option anything
6583 other than PKCS#8 should be dumped: but the other formats have to
6584 stay in the name of compatibility.
6586 With public keys and the benefit of hindsight one standard format
6587 is used which works with EVP_PKEY, RSA or DSA structures: though
6588 it clearly returns an error if you try to read the wrong kind of key.
6590 Added a -pubkey option to the 'x509' utility to output the public key.
6591 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
6592 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
6593 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
6594 that do the same as the EVP_PKEY_assign_*() except they up the
6595 reference count of the added key (they don't "swallow" the
6599 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
6600 CRLs would fail if the file contained no certificates or no CRLs:
6601 added a new function to read in both types and return the number
6602 read: this means that if none are read it will be an error. The
6603 DER versions of the certificate and CRL reader would always fail
6604 because it isn't possible to mix certificates and CRLs in DER format
6605 without choking one or the other routine. Changed this to just read
6606 a certificate: this is the best we can do. Also modified the code
6607 in apps/verify.c to take notice of return codes: it was previously
6608 attempting to read in certificates from NULL pointers and ignoring
6609 any errors: this is one reason why the cert and CRL reader seemed
6610 to work. It doesn't check return codes from the default certificate
6611 routines: these may well fail if the certificates aren't installed.
6614 *) Code to support otherName option in GeneralName.
6617 *) First update to verify code. Change the verify utility
6618 so it warns if it is passed a self signed certificate:
6619 for consistency with the normal behaviour. X509_verify
6620 has been modified to it will now verify a self signed
6621 certificate if *exactly* the same certificate appears
6622 in the store: it was previously impossible to trust a
6623 single self signed certificate. This means that:
6624 openssl verify ss.pem
6625 now gives a warning about a self signed certificate but
6626 openssl verify -CAfile ss.pem ss.pem
6630 *) For servers, store verify_result in SSL_SESSION data structure
6631 (and add it to external session representation).
6632 This is needed when client certificate verifications fails,
6633 but an application-provided verification callback (set by
6634 SSL_CTX_set_cert_verify_callback) allows accepting the session
6635 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
6636 but returns 1): When the session is reused, we have to set
6637 ssl->verify_result to the appropriate error code to avoid
6639 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
6641 *) Fix a bug in the new PKCS#7 code: it didn't consider the
6642 case in PKCS7_dataInit() where the signed PKCS7 structure
6643 didn't contain any existing data because it was being created.
6644 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
6646 *) Add a salt to the key derivation routines in enc.c. This
6647 forms the first 8 bytes of the encrypted file. Also add a
6648 -S option to allow a salt to be input on the command line.
6651 *) New function X509_cmp(). Oddly enough there wasn't a function
6652 to compare two certificates. We do this by working out the SHA1
6653 hash and comparing that. X509_cmp() will be needed by the trust
6657 *) SSL_get1_session() is like SSL_get_session(), but increments
6658 the reference count in the SSL_SESSION returned.
6659 [Geoff Thorpe <geoff@eu.c2.net>]
6661 *) Fix for 'req': it was adding a null to request attributes.
6662 Also change the X509_LOOKUP and X509_INFO code to handle
6663 certificate auxiliary information.
6666 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
6670 *) Add the possibility to add extra information to the memory leak
6671 detecting output, to form tracebacks, showing from where each
6672 allocation was originated: CRYPTO_push_info("constant string") adds
6673 the string plus current file name and line number to a per-thread
6674 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
6675 is like calling CYRPTO_pop_info() until the stack is empty.
6676 Also updated memory leak detection code to be multi-thread-safe.
6679 *) Add options -text and -noout to pkcs7 utility and delete the
6680 encryption options which never did anything. Update docs.
6683 *) Add options to some of the utilities to allow the pass phrase
6684 to be included on either the command line (not recommended on
6685 OSes like Unix) or read from the environment. Update the
6686 manpages and fix a few bugs.
6689 *) Add a few manpages for some of the openssl commands.
6692 *) Fix the -revoke option in ca. It was freeing up memory twice,
6693 leaking and not finding already revoked certificates.
6696 *) Extensive changes to support certificate auxiliary information.
6697 This involves the use of X509_CERT_AUX structure and X509_AUX
6698 functions. An X509_AUX function such as PEM_read_X509_AUX()
6699 can still read in a certificate file in the usual way but it
6700 will also read in any additional "auxiliary information". By
6701 doing things this way a fair degree of compatibility can be
6702 retained: existing certificates can have this information added
6703 using the new 'x509' options.
6705 Current auxiliary information includes an "alias" and some trust
6706 settings. The trust settings will ultimately be used in enhanced
6707 certificate chain verification routines: currently a certificate
6708 can only be trusted if it is self signed and then it is trusted
6712 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
6713 The problem was that one of the replacement routines had not been working
6714 since SSLeay releases. For now the offending routine has been replaced
6715 with non-optimised assembler. Even so, this now gives around 95%
6716 performance improvement for 1024 bit RSA signs.
6719 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
6720 handling. Most clients have the effective key size in bits equal to
6721 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
6722 A few however don't do this and instead use the size of the decrypted key
6723 to determine the RC2 key length and the AlgorithmIdentifier to determine
6724 the effective key length. In this case the effective key length can still
6725 be 40 bits but the key length can be 168 bits for example. This is fixed
6726 by manually forcing an RC2 key into the EVP_PKEY structure because the
6727 EVP code can't currently handle unusual RC2 key sizes: it always assumes
6728 the key length and effective key length are equal.
6731 *) Add a bunch of functions that should simplify the creation of
6732 X509_NAME structures. Now you should be able to do:
6733 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
6734 and have it automatically work out the correct field type and fill in
6735 the structures. The more adventurous can try:
6736 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
6737 and it will (hopefully) work out the correct multibyte encoding.
6740 *) Change the 'req' utility to use the new field handling and multibyte
6741 copy routines. Before the DN field creation was handled in an ad hoc
6742 way in req, ca, and x509 which was rather broken and didn't support
6743 BMPStrings or UTF8Strings. Since some software doesn't implement
6744 BMPStrings or UTF8Strings yet, they can be enabled using the config file
6745 using the dirstring_type option. See the new comment in the default
6746 openssl.cnf for more info.
6749 *) Make crypto/rand/md_rand.c more robust:
6750 - Assure unique random numbers after fork().
6751 - Make sure that concurrent threads access the global counter and
6752 md serializably so that we never lose entropy in them
6753 or use exactly the same state in multiple threads.
6754 Access to the large state is not always serializable because
6755 the additional locking could be a performance killer, and
6756 md should be large enough anyway.
6759 *) New file apps/app_rand.c with commonly needed functionality
6760 for handling the random seed file.
6762 Use the random seed file in some applications that previously did not:
6764 dsaparam -genkey (which also ignored its '-rand' option),
6767 x509 (when signing).
6768 Except on systems with /dev/urandom, it is crucial to have a random
6769 seed file at least for key creation, DSA signing, and for DH exchanges;
6770 for RSA signatures we could do without one.
6772 gendh and gendsa (unlike genrsa) used to read only the first byte
6773 of each file listed in the '-rand' option. The function as previously
6774 found in genrsa is now in app_rand.c and is used by all programs
6775 that support '-rand'.
6778 *) In RAND_write_file, use mode 0600 for creating files;
6779 don't just chmod when it may be too late.
6782 *) Report an error from X509_STORE_load_locations
6783 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
6786 *) New function ASN1_mbstring_copy() this copies a string in either
6787 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
6788 into an ASN1_STRING type. A mask of permissible types is passed
6789 and it chooses the "minimal" type to use or an error if not type
6793 *) Add function equivalents to the various macros in asn1.h. The old
6794 macros are retained with an M_ prefix. Code inside the library can
6795 use the M_ macros. External code (including the openssl utility)
6796 should *NOT* in order to be "shared library friendly".
6799 *) Add various functions that can check a certificate's extensions
6800 to see if it usable for various purposes such as SSL client,
6801 server or S/MIME and CAs of these types. This is currently
6802 VERY EXPERIMENTAL but will ultimately be used for certificate chain
6803 verification. Also added a -purpose flag to x509 utility to
6804 print out all the purposes.
6807 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
6811 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
6812 for, obtain and decode and extension and obtain its critical flag.
6813 This allows all the necessary extension code to be handled in a
6814 single function call.
6817 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
6818 platforms. See crypto/rc4/rc4_enc.c for further details.
6821 *) New -noout option to asn1parse. This causes no output to be produced
6822 its main use is when combined with -strparse and -out to extract data
6823 from a file (which may not be in ASN.1 format).
6826 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
6827 when producing the local key id.
6828 [Richard Levitte <levitte@stacken.kth.se>]
6830 *) New option -dhparam in s_server. This allows a DH parameter file to be
6831 stated explicitly. If it is not stated then it tries the first server
6832 certificate file. The previous behaviour hard coded the filename
6836 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
6837 a public key to be input or output. For example:
6838 openssl rsa -in key.pem -pubout -out pubkey.pem
6839 Also added necessary DSA public key functions to handle this.
6842 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
6843 in the message. This was handled by allowing
6844 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
6845 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
6847 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
6848 to the end of the strings whereas this didn't. This would cause problems
6849 if strings read with d2i_ASN1_bytes() were later modified.
6850 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
6852 *) Fix for base64 decode bug. When a base64 bio reads only one line of
6853 data and it contains EOF it will end up returning an error. This is
6854 caused by input 46 bytes long. The cause is due to the way base64
6855 BIOs find the start of base64 encoded data. They do this by trying a
6856 trial decode on each line until they find one that works. When they
6857 do a flag is set and it starts again knowing it can pass all the
6858 data directly through the decoder. Unfortunately it doesn't reset
6859 the context it uses. This means that if EOF is reached an attempt
6860 is made to pass two EOFs through the context and this causes the
6861 resulting error. This can also cause other problems as well. As is
6862 usual with these problems it takes *ages* to find and the fix is
6863 trivial: move one line.
6864 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
6866 *) Ugly workaround to get s_client and s_server working under Windows. The
6867 old code wouldn't work because it needed to select() on sockets and the
6868 tty (for keypresses and to see if data could be written). Win32 only
6869 supports select() on sockets so we select() with a 1s timeout on the
6870 sockets and then see if any characters are waiting to be read, if none
6871 are present then we retry, we also assume we can always write data to
6872 the tty. This isn't nice because the code then blocks until we've
6873 received a complete line of data and it is effectively polling the
6874 keyboard at 1s intervals: however it's quite a bit better than not
6875 working at all :-) A dedicated Windows application might handle this
6876 with an event loop for example.
6879 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
6880 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
6881 will be called when RSA_sign() and RSA_verify() are used. This is useful
6882 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
6883 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
6884 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
6885 This necessitated the support of an extra signature type NID_md5_sha1
6886 for SSL signatures and modifications to the SSL library to use it instead
6887 of calling RSA_public_decrypt() and RSA_private_encrypt().
6890 *) Add new -verify -CAfile and -CApath options to the crl program, these
6891 will lookup a CRL issuers certificate and verify the signature in a
6892 similar way to the verify program. Tidy up the crl program so it
6893 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
6894 less strict. It will now permit CRL extensions even if it is not
6895 a V2 CRL: this will allow it to tolerate some broken CRLs.
6898 *) Initialize all non-automatic variables each time one of the openssl
6899 sub-programs is started (this is necessary as they may be started
6900 multiple times from the "OpenSSL>" prompt).
6901 [Lennart Bang, Bodo Moeller]
6903 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
6904 removing all other RSA functionality (this is what NO_RSA does). This
6905 is so (for example) those in the US can disable those operations covered
6906 by the RSA patent while allowing storage and parsing of RSA keys and RSA
6910 *) Non-copying interface to BIO pairs.
6911 (still largely untested)
6914 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
6915 ASCII string. This was handled independently in various places before.
6918 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
6919 UTF8 strings a character at a time.
6922 *) Use client_version from client hello to select the protocol
6923 (s23_srvr.c) and for RSA client key exchange verification
6924 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
6927 *) Add various utility functions to handle SPKACs, these were previously
6928 handled by poking round in the structure internals. Added new function
6929 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
6930 print, verify and generate SPKACs. Based on an original idea from
6931 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
6934 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
6937 *) Allow the config file extension section to be overwritten on the
6938 command line. Based on an original idea from Massimiliano Pala
6939 <madwolf@comune.modena.it>. The new option is called -extensions
6940 and can be applied to ca, req and x509. Also -reqexts to override
6941 the request extensions in req and -crlexts to override the crl extensions
6945 *) Add new feature to the SPKAC handling in ca. Now you can include
6946 the same field multiple times by preceding it by "XXXX." for example:
6949 this is the same syntax as used in the req config file.
6952 *) Allow certificate extensions to be added to certificate requests. These
6953 are specified in a 'req_extensions' option of the req section of the
6954 config file. They can be printed out with the -text option to req but
6955 are otherwise ignored at present.
6958 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
6959 data read consists of only the final block it would not decrypted because
6960 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
6961 A misplaced 'break' also meant the decrypted final block might not be
6962 copied until the next read.
6965 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
6966 a few extra parameters to the DH structure: these will be useful if
6967 for example we want the value of 'q' or implement X9.42 DH.
6970 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
6971 provides hooks that allow the default DSA functions or functions on a
6972 "per key" basis to be replaced. This allows hardware acceleration and
6973 hardware key storage to be handled without major modification to the
6974 library. Also added low level modexp hooks and CRYPTO_EX structure and
6975 associated functions.
6978 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
6979 as "read only": it can't be written to and the buffer it points to will
6980 not be freed. Reading from a read only BIO is much more efficient than
6981 a normal memory BIO. This was added because there are several times when
6982 an area of memory needs to be read from a BIO. The previous method was
6983 to create a memory BIO and write the data to it, this results in two
6984 copies of the data and an O(n^2) reading algorithm. There is a new
6985 function BIO_new_mem_buf() which creates a read only memory BIO from
6986 an area of memory. Also modified the PKCS#7 routines to use read only
6990 *) Bugfix: ssl23_get_client_hello did not work properly when called in
6991 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
6992 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
6993 but a retry condition occured while trying to read the rest.
6996 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
6997 NID_pkcs7_encrypted by default: this was wrong since this should almost
6998 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
6999 the encrypted data type: this is a more sensible place to put it and it
7000 allows the PKCS#12 code to be tidied up that duplicated this
7004 *) Changed obj_dat.pl script so it takes its input and output files on
7005 the command line. This should avoid shell escape redirection problems
7009 *) Initial support for certificate extension requests, these are included
7010 in things like Xenroll certificate requests. Included functions to allow
7011 extensions to be obtained and added.
7014 *) -crlf option to s_client and s_server for sending newlines as
7015 CRLF (as required by many protocols).
7018 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
7020 *) Install libRSAglue.a when OpenSSL is built with RSAref.
7021 [Ralf S. Engelschall]
7023 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
7024 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
7026 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
7030 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
7031 DH parameters/keys (q is lost during that conversion, but the resulting
7032 DH parameters contain its length).
7034 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
7035 much faster than DH_generate_parameters (which creates parameters
7036 where p = 2*q + 1), and also the smaller q makes DH computations
7037 much more efficient (160-bit exponentiation instead of 1024-bit
7038 exponentiation); so this provides a convenient way to support DHE
7039 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
7040 utter importance to use
7041 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7043 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7044 when such DH parameters are used, because otherwise small subgroup
7045 attacks may become possible!
7048 *) Avoid memory leak in i2d_DHparams.
7051 *) Allow the -k option to be used more than once in the enc program:
7052 this allows the same encrypted message to be read by multiple recipients.
7055 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
7056 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
7057 it will always use the numerical form of the OID, even if it has a short
7061 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
7062 method only got called if p,q,dmp1,dmq1,iqmp components were present,
7063 otherwise bn_mod_exp was called. In the case of hardware keys for example
7064 no private key components need be present and it might store extra data
7065 in the RSA structure, which cannot be accessed from bn_mod_exp.
7066 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
7067 private key operations.
7070 *) Added support for SPARC Linux.
7073 *) pem_password_cb function type incompatibly changed from
7074 typedef int pem_password_cb(char *buf, int size, int rwflag);
7076 ....(char *buf, int size, int rwflag, void *userdata);
7077 so that applications can pass data to their callbacks:
7078 The PEM[_ASN1]_{read,write}... functions and macros now take an
7079 additional void * argument, which is just handed through whenever
7080 the password callback is called.
7081 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
7083 New function SSL_CTX_set_default_passwd_cb_userdata.
7085 Compatibility note: As many C implementations push function arguments
7086 onto the stack in reverse order, the new library version is likely to
7087 interoperate with programs that have been compiled with the old
7088 pem_password_cb definition (PEM_whatever takes some data that
7089 happens to be on the stack as its last argument, and the callback
7090 just ignores this garbage); but there is no guarantee whatsoever that
7093 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
7094 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
7095 problems not only on Windows, but also on some Unix platforms.
7096 To avoid problematic command lines, these definitions are now in an
7097 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
7098 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
7101 *) MIPS III/IV assembler module is reimplemented.
7104 *) More DES library cleanups: remove references to srand/rand and
7105 delete an unused file.
7108 *) Add support for the the free Netwide assembler (NASM) under Win32,
7109 since not many people have MASM (ml) and it can be hard to obtain.
7110 This is currently experimental but it seems to work OK and pass all
7111 the tests. Check out INSTALL.W32 for info.
7114 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
7115 without temporary keys kept an extra copy of the server key,
7116 and connections with temporary keys did not free everything in case
7120 *) New function RSA_check_key and new openssl rsa option -check
7121 for verifying the consistency of RSA keys.
7122 [Ulf Moeller, Bodo Moeller]
7124 *) Various changes to make Win32 compile work:
7125 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
7126 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
7127 comparison" warnings.
7128 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
7131 *) Add a debugging option to PKCS#5 v2 key generation function: when
7132 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
7133 derived keys are printed to stderr.
7136 *) Copy the flags in ASN1_STRING_dup().
7137 [Roman E. Pavlov <pre@mo.msk.ru>]
7139 *) The x509 application mishandled signing requests containing DSA
7140 keys when the signing key was also DSA and the parameters didn't match.
7142 It was supposed to omit the parameters when they matched the signing key:
7143 the verifying software was then supposed to automatically use the CA's
7144 parameters if they were absent from the end user certificate.
7146 Omitting parameters is no longer recommended. The test was also
7147 the wrong way round! This was probably due to unusual behaviour in
7148 EVP_cmp_parameters() which returns 1 if the parameters match.
7149 This meant that parameters were omitted when they *didn't* match and
7150 the certificate was useless. Certificates signed with 'ca' didn't have
7152 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
7154 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
7155 The interface is as follows:
7156 Applications can use
7157 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
7158 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
7159 "off" is now the default.
7160 The library internally uses
7161 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
7162 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
7163 to disable memory-checking temporarily.
7165 Some inconsistent states that previously were possible (and were
7166 even the default) are now avoided.
7168 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
7169 with each memory chunk allocated; this is occasionally more helpful
7170 than just having a counter.
7172 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
7174 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
7178 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
7179 which largely parallels "options", but is for changing API behaviour,
7180 whereas "options" are about protocol behaviour.
7181 Initial "mode" flags are:
7183 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
7184 a single record has been written.
7185 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
7186 retries use the same buffer location.
7187 (But all of the contents must be
7191 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
7194 *) Fix problems with no-hmac etc.
7195 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
7197 *) New functions RSA_get_default_method(), RSA_set_method() and
7198 RSA_get_method(). These allows replacement of RSA_METHODs without having
7199 to mess around with the internals of an RSA structure.
7202 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
7203 Also really enable memory leak checks in openssl.c and in some
7205 [Chad C. Mulligan, Bodo Moeller]
7207 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
7208 up the length of negative integers. This has now been simplified to just
7209 store the length when it is first determined and use it later, rather
7210 than trying to keep track of where data is copied and updating it to
7212 [Steve Henson, reported by Brien Wheeler
7213 <bwheeler@authentica-security.com>]
7215 *) Add a new function PKCS7_signatureVerify. This allows the verification
7216 of a PKCS#7 signature but with the signing certificate passed to the
7217 function itself. This contrasts with PKCS7_dataVerify which assumes the
7218 certificate is present in the PKCS#7 structure. This isn't always the
7219 case: certificates can be omitted from a PKCS#7 structure and be
7220 distributed by "out of band" means (such as a certificate database).
7223 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
7224 function prototypes in pem.h, also change util/mkdef.pl to add the
7225 necessary function names.
7228 *) mk1mf.pl (used by Windows builds) did not properly read the
7229 options set by Configure in the top level Makefile, and Configure
7230 was not even able to write more than one option correctly.
7231 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
7234 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
7235 file to be loaded from a BIO or FILE pointer. The BIO version will
7236 for example allow memory BIOs to contain config info.
7239 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
7240 Whoever hopes to achieve shared-library compatibility across versions
7241 must use this, not the compile-time macro.
7242 (Exercise 0.9.4: Which is the minimum library version required by
7244 Note: All this applies only to multi-threaded programs, others don't
7248 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
7249 through a BIO pair triggered the default case, i.e.
7250 SSLerr(...,SSL_R_UNKNOWN_STATE).
7253 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
7254 can use the SSL library even if none of the specific BIOs is
7258 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
7259 for the encoded length.
7260 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
7262 *) Add initial documentation of the X509V3 functions.
7265 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
7266 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
7267 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
7268 secure PKCS#8 private key format with a high iteration count.
7271 *) Fix determination of Perl interpreter: A perl or perl5
7272 _directory_ in $PATH was also accepted as the interpreter.
7273 [Ralf S. Engelschall]
7275 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
7276 wrong with it but it was very old and did things like calling
7277 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
7281 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
7282 to use the new extension code.
7285 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
7286 with macros. This should make it easier to change their form, add extra
7287 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
7291 *) Add to configuration table a new entry that can specify an alternative
7292 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
7293 according to Mark Crispin <MRC@Panda.COM>.
7297 *) DES CBC did not update the IV. Weird.
7300 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
7301 Changing the behaviour of the former might break existing programs --
7302 where IV updating is needed, des_ncbc_encrypt can be used.
7305 *) When bntest is run from "make test" it drives bc to check its
7306 calculations, as well as internally checking them. If an internal check
7307 fails, it needs to cause bc to give a non-zero result or make test carries
7308 on without noticing the failure. Fixed.
7311 *) DES library cleanups.
7314 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
7315 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
7316 ciphers. NOTE: although the key derivation function has been verified
7317 against some published test vectors it has not been extensively tested
7318 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
7322 *) Instead of "mkdir -p", which is not fully portable, use new
7323 Perl script "util/mkdir-p.pl".
7326 *) Rewrite the way password based encryption (PBE) is handled. It used to
7327 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
7328 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
7329 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
7330 the 'parameter' field of the AlgorithmIdentifier is passed to the
7331 underlying key generation function so it must do its own ASN1 parsing.
7332 This has also changed the EVP_PBE_CipherInit() function which now has a
7333 'parameter' argument instead of literal salt and iteration count values
7334 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
7337 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
7338 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
7339 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
7340 KEY" because this clashed with PKCS#8 unencrypted string. Since this
7341 value was just used as a "magic string" and not used directly its
7342 value doesn't matter.
7345 *) Introduce some semblance of const correctness to BN. Shame C doesn't
7349 *) "linux-sparc64" configuration (ultrapenguin).
7350 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
7351 "linux-sparc" configuration.
7352 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
7354 *) config now generates no-xxx options for missing ciphers.
7357 *) Support the EBCDIC character set (work in progress).
7358 File ebcdic.c not yet included because it has a different license.
7359 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7361 *) Support BS2000/OSD-POSIX.
7362 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7364 *) Make callbacks for key generation use void * instead of char *.
7367 *) Make S/MIME samples compile (not yet tested).
7370 *) Additional typesafe stacks.
7373 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
7377 Changes between 0.9.3 and 0.9.3a [29 May 1999]
7379 *) New configuration variant "sco5-gcc".
7381 *) Updated some demos.
7382 [Sean O Riordain, Wade Scholine]
7384 *) Add missing BIO_free at exit of pkcs12 application.
7387 *) Fix memory leak in conf.c.
7390 *) Updates for Win32 to assembler version of MD5.
7393 *) Set #! path to perl in apps/der_chop to where we found it
7394 instead of using a fixed path.
7397 *) SHA library changes for irix64-mips4-cc.
7400 *) Improvements for VMS support.
7404 Changes between 0.9.2b and 0.9.3 [24 May 1999]
7406 *) Bignum library bug fix. IRIX 6 passes "make test" now!
7407 This also avoids the problems with SC4.2 and unpatched SC5.
7408 [Andy Polyakov <appro@fy.chalmers.se>]
7410 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
7411 These are required because of the typesafe stack would otherwise break
7412 existing code. If old code used a structure member which used to be STACK
7413 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
7414 sk_num or sk_value it would produce an error because the num, data members
7415 are not present in STACK_OF. Now it just produces a warning. sk_set
7416 replaces the old method of assigning a value to sk_value
7417 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
7418 that does this will no longer work (and should use sk_set instead) but
7419 this could be regarded as a "questionable" behaviour anyway.
7422 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
7423 correctly handle encrypted S/MIME data.
7426 *) Change type of various DES function arguments from des_cblock
7427 (which means, in function argument declarations, pointer to char)
7428 to des_cblock * (meaning pointer to array with 8 char elements),
7429 which allows the compiler to do more typechecking; it was like
7430 that back in SSLeay, but with lots of ugly casts.
7432 Introduce new type const_des_cblock.
7435 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
7436 problems: find RecipientInfo structure that matches recipient certificate
7437 and initialise the ASN1 structures properly based on passed cipher.
7440 *) Belatedly make the BN tests actually check the results.
7443 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
7444 to and from BNs: it was completely broken. New compilation option
7445 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
7446 key elements as negative integers.
7449 *) Reorganize and speed up MD5.
7450 [Andy Polyakov <appro@fy.chalmers.se>]
7453 [Richard Levitte <richard@levitte.org>]
7455 *) New option -out to asn1parse to allow the parsed structure to be
7456 output to a file. This is most useful when combined with the -strparse
7457 option to examine the output of things like OCTET STRINGS.
7460 *) Make SSL library a little more fool-proof by not requiring any longer
7461 that SSL_set_{accept,connect}_state be called before
7462 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
7463 in many applications because usually everything *appeared* to work as
7464 intended anyway -- now it really works as intended).
7467 *) Move openssl.cnf out of lib/.
7470 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
7471 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
7472 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
7473 [Ralf S. Engelschall]
7475 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
7476 handle PKCS#7 enveloped data properly.
7477 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
7479 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
7480 copying pointers. The cert_st handling is changed by this in
7481 various ways (and thus what used to be known as ctx->default_cert
7482 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
7483 any longer when s->cert does not give us what we need).
7484 ssl_cert_instantiate becomes obsolete by this change.
7485 As soon as we've got the new code right (possibly it already is?),
7486 we have solved a couple of bugs of the earlier code where s->cert
7487 was used as if it could not have been shared with other SSL structures.
7489 Note that using the SSL API in certain dirty ways now will result
7490 in different behaviour than observed with earlier library versions:
7491 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
7492 does not influence s as it used to.
7494 In order to clean up things more thoroughly, inside SSL_SESSION
7495 we don't use CERT any longer, but a new structure SESS_CERT
7496 that holds per-session data (if available); currently, this is
7497 the peer's certificate chain and, for clients, the server's certificate
7498 and temporary key. CERT holds only those values that can have
7499 meaningful defaults in an SSL_CTX.
7502 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
7503 from the internal representation. Various PKCS#7 fixes: remove some
7504 evil casts and set the enc_dig_alg field properly based on the signing
7508 *) Allow PKCS#12 password to be set from the command line or the
7509 environment. Let 'ca' get its config file name from the environment
7510 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
7514 *) Allow certificate policies extension to use an IA5STRING for the
7515 organization field. This is contrary to the PKIX definition but
7516 VeriSign uses it and IE5 only recognises this form. Document 'x509'
7520 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
7521 without disallowing inline assembler and the like for non-pedantic builds.
7524 *) Support Borland C++ builder.
7525 [Janez Jere <jj@void.si>, modified by Ulf Möller]
7530 *) SHA-1 cleanups and performance enhancements.
7531 [Andy Polyakov <appro@fy.chalmers.se>]
7533 *) Sparc v8plus assembler for the bignum library.
7534 [Andy Polyakov <appro@fy.chalmers.se>]
7536 *) Accept any -xxx and +xxx compiler options in Configure.
7539 *) Update HPUX configuration.
7542 *) Add missing sk_<type>_unshift() function to safestack.h
7543 [Ralf S. Engelschall]
7545 *) New function SSL_CTX_use_certificate_chain_file that sets the
7546 "extra_cert"s in addition to the certificate. (This makes sense
7547 only for "PEM" format files, as chains as a whole are not
7551 *) Support verify_depth from the SSL API.
7552 x509_vfy.c had what can be considered an off-by-one-error:
7553 Its depth (which was not part of the external interface)
7554 was actually counting the number of certificates in a chain;
7555 now it really counts the depth.
7558 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
7559 instead of X509err, which often resulted in confusing error
7560 messages since the error codes are not globally unique
7561 (e.g. an alleged error in ssl3_accept when a certificate
7562 didn't match the private key).
7564 *) New function SSL_CTX_set_session_id_context that allows to set a default
7565 value (so that you don't need SSL_set_session_id_context for each
7566 connection using the SSL_CTX).
7569 *) OAEP decoding bug fix.
7572 *) Support INSTALL_PREFIX for package builders, as proposed by
7576 *) New Configure options "threads" and "no-threads". For systems
7577 where the proper compiler options are known (currently Solaris
7578 and Linux), "threads" is the default.
7581 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
7584 *) Install various scripts to $(OPENSSLDIR)/misc, not to
7585 $(INSTALLTOP)/bin -- they shouldn't clutter directories
7586 such as /usr/local/bin.
7589 *) "make linux-shared" to build shared libraries.
7590 [Niels Poppe <niels@netbox.org>]
7592 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
7595 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
7596 extension adding in x509 utility.
7599 *) Remove NOPROTO sections and error code comments.
7602 *) Partial rewrite of the DEF file generator to now parse the ANSI
7606 *) New Configure options --prefix=DIR and --openssldir=DIR.
7609 *) Complete rewrite of the error code script(s). It is all now handled
7610 by one script at the top level which handles error code gathering,
7611 header rewriting and C source file generation. It should be much better
7612 than the old method: it now uses a modified version of Ulf's parser to
7613 read the ANSI prototypes in all header files (thus the old K&R definitions
7614 aren't needed for error creation any more) and do a better job of
7615 translating function codes into names. The old 'ASN1 error code imbedded
7616 in a comment' is no longer necessary and it doesn't use .err files which
7617 have now been deleted. Also the error code call doesn't have to appear all
7618 on one line (which resulted in some large lines...).
7621 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
7624 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
7625 0 (which usually indicates a closed connection), but continue reading.
7628 *) Fix some race conditions.
7631 *) Add support for CRL distribution points extension. Add Certificate
7632 Policies and CRL distribution points documentation.
7635 *) Move the autogenerated header file parts to crypto/opensslconf.h.
7638 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
7639 8 of keying material. Merlin has also confirmed interop with this fix
7640 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
7641 [Merlin Hughes <merlin@baltimore.ie>]
7643 *) Fix lots of warnings.
7644 [Richard Levitte <levitte@stacken.kth.se>]
7646 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
7647 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
7648 [Richard Levitte <levitte@stacken.kth.se>]
7650 *) Fix problems with sizeof(long) == 8.
7651 [Andy Polyakov <appro@fy.chalmers.se>]
7653 *) Change functions to ANSI C.
7656 *) Fix typos in error codes.
7657 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
7659 *) Remove defunct assembler files from Configure.
7662 *) SPARC v8 assembler BIGNUM implementation.
7663 [Andy Polyakov <appro@fy.chalmers.se>]
7665 *) Support for Certificate Policies extension: both print and set.
7666 Various additions to support the r2i method this uses.
7669 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
7670 return a const string when you are expecting an allocated buffer.
7673 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
7674 types DirectoryString and DisplayText.
7677 *) Add code to allow r2i extensions to access the configuration database,
7678 add an LHASH database driver and add several ctx helper functions.
7681 *) Fix an evil bug in bn_expand2() which caused various BN functions to
7682 fail when they extended the size of a BIGNUM.
7685 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
7686 support typesafe stack.
7689 *) Fix typo in SSL_[gs]et_options().
7690 [Nils Frostberg <nils@medcom.se>]
7692 *) Delete various functions and files that belonged to the (now obsolete)
7693 old X509V3 handling code.
7696 *) New Configure option "rsaref".
7699 *) Don't auto-generate pem.h.
7702 *) Introduce type-safe ASN.1 SETs.
7705 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
7706 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
7708 *) Introduce type-safe STACKs. This will almost certainly break lots of code
7709 that links with OpenSSL (well at least cause lots of warnings), but fear
7710 not: the conversion is trivial, and it eliminates loads of evil casts. A
7711 few STACKed things have been converted already. Feel free to convert more.
7712 In the fullness of time, I'll do away with the STACK type altogether.
7715 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
7716 specified in <certfile> by updating the entry in the index.txt file.
7717 This way one no longer has to edit the index.txt file manually for
7718 revoking a certificate. The -revoke option does the gory details now.
7719 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
7721 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
7722 `-text' option at all and this way the `-noout -text' combination was
7723 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
7724 [Ralf S. Engelschall]
7726 *) Make sure a corresponding plain text error message exists for the
7727 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
7728 verify callback function determined that a certificate was revoked.
7729 [Ralf S. Engelschall]
7731 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
7732 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
7733 all available cipers including rc5, which was forgotten until now.
7734 In order to let the testing shell script know which algorithms
7735 are available, a new (up to now undocumented) command
7736 "openssl list-cipher-commands" is used.
7739 *) Bugfix: s_client occasionally would sleep in select() when
7740 it should have checked SSL_pending() first.
7743 *) New functions DSA_do_sign and DSA_do_verify to provide access to
7744 the raw DSA values prior to ASN.1 encoding.
7747 *) Tweaks to Configure
7748 [Niels Poppe <niels@netbox.org>]
7750 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
7754 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
7757 *) New config option to avoid instructions that are illegal on the 80386.
7758 The default code is faster, but requires at least a 486.
7761 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
7762 SSL2_SERVER_VERSION (not used at all) macros, which are now the
7763 same as SSL2_VERSION anyway.
7766 *) New "-showcerts" option for s_client.
7769 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
7770 application. Various cleanups and fixes.
7773 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
7774 modify error routines to work internally. Add error codes and PBE init
7775 to library startup routines.
7778 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
7779 packing functions to asn1 and evp. Changed function names and error
7780 codes along the way.
7783 *) PKCS12 integration: and so it begins... First of several patches to
7784 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
7785 objects to objects.h
7788 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
7789 and display support for Thawte strong extranet extension.
7792 *) Add LinuxPPC support.
7793 [Jeff Dubrule <igor@pobox.org>]
7795 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
7796 bn_div_words in alpha.s.
7797 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
7799 *) Make sure the RSA OAEP test is skipped under -DRSAref because
7800 OAEP isn't supported when OpenSSL is built with RSAref.
7801 [Ulf Moeller <ulf@fitug.de>]
7803 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
7804 so they no longer are missing under -DNOPROTO.
7805 [Soren S. Jorvang <soren@t.dk>]
7808 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
7810 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
7811 doesn't work when the session is reused. Coming soon!
7814 *) Fix a security hole, that allows sessions to be reused in the wrong
7815 context thus bypassing client cert protection! All software that uses
7816 client certs and session caches in multiple contexts NEEDS PATCHING to
7817 allow session reuse! A fuller solution is in the works.
7818 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
7820 *) Some more source tree cleanups (removed obsolete files
7821 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
7822 permission on "config" script to be executable) and a fix for the INSTALL
7824 [Ulf Moeller <ulf@fitug.de>]
7826 *) Remove some legacy and erroneous uses of malloc, free instead of
7828 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
7830 *) Make rsa_oaep_test return non-zero on error.
7831 [Ulf Moeller <ulf@fitug.de>]
7833 *) Add support for native Solaris shared libraries. Configure
7834 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
7835 if someone would make that last step automatic.
7836 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
7838 *) ctx_size was not built with the right compiler during "make links". Fixed.
7841 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
7842 except NULL ciphers". This means the default cipher list will no longer
7843 enable NULL ciphers. They need to be specifically enabled e.g. with
7844 the string "DEFAULT:eNULL".
7847 *) Fix to RSA private encryption routines: if p < q then it would
7848 occasionally produce an invalid result. This will only happen with
7849 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
7852 *) Be less restrictive and allow also `perl util/perlpath.pl
7853 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
7854 because this way one can also use an interpreter named `perl5' (which is
7855 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
7856 installed as `perl').
7857 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7859 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
7860 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7862 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
7863 advapi32.lib to Win32 build and change the pem test comparision
7864 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
7865 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
7866 and crypto/des/ede_cbcm_enc.c.
7869 *) DES quad checksum was broken on big-endian architectures. Fixed.
7872 *) Comment out two functions in bio.h that aren't implemented. Fix up the
7873 Win32 test batch file so it (might) work again. The Win32 test batch file
7874 is horrible: I feel ill....
7877 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
7878 in e_os.h. Audit of header files to check ANSI and non ANSI
7879 sections: 10 functions were absent from non ANSI section and not exported
7880 from Windows DLLs. Fixed up libeay.num for new functions.
7883 *) Make `openssl version' output lines consistent.
7884 [Ralf S. Engelschall]
7886 *) Fix Win32 symbol export lists for BIO functions: Added
7887 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
7888 to ms/libeay{16,32}.def.
7889 [Ralf S. Engelschall]
7891 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
7892 fine under Unix and passes some trivial tests I've now added. But the
7893 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
7894 added to make sure no one expects that this stuff really works in the
7895 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
7896 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
7898 [Ralf S. Engelschall]
7900 *) Fix the generation of two part addresses in perl.
7901 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
7903 *) Add config entry for Linux on MIPS.
7904 [John Tobey <jtobey@channel1.com>]
7906 *) Make links whenever Configure is run, unless we are on Windoze.
7909 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
7910 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
7914 *) Add a useful kludge to allow package maintainers to specify compiler and
7915 other platforms details on the command line without having to patch the
7916 Configure script everytime: One now can use ``perl Configure
7917 <id>:<details>'', i.e. platform ids are allowed to have details appended
7918 to them (seperated by colons). This is treated as there would be a static
7919 pre-configured entry in Configure's %table under key <id> with value
7920 <details> and ``perl Configure <id>'' is called. So, when you want to
7921 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
7922 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
7923 now, which overrides the FreeBSD-elf entry on-the-fly.
7924 [Ralf S. Engelschall]
7926 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
7929 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
7930 on the `perl Configure ...' command line. This way one can compile
7931 OpenSSL libraries with Position Independent Code (PIC) which is needed
7932 for linking it into DSOs.
7933 [Ralf S. Engelschall]
7935 *) Remarkably, export ciphers were totally broken and no-one had noticed!
7939 *) Cleaned up the LICENSE document: The official contact for any license
7940 questions now is the OpenSSL core team under openssl-core@openssl.org.
7941 And add a paragraph about the dual-license situation to make sure people
7942 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
7943 to the OpenSSL toolkit.
7944 [Ralf S. Engelschall]
7946 *) General source tree makefile cleanups: Made `making xxx in yyy...'
7947 display consistent in the source tree and replaced `/bin/rm' by `rm'.
7948 Additonally cleaned up the `make links' target: Remove unnecessary
7949 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
7950 to speed processing and no longer clutter the display with confusing
7951 stuff. Instead only the actually done links are displayed.
7952 [Ralf S. Engelschall]
7954 *) Permit null encryption ciphersuites, used for authentication only. It used
7955 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
7956 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
7960 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
7961 signed attributes when verifying signatures (this would break them),
7962 the detached data encoding was wrong and public keys obtained using
7963 X509_get_pubkey() weren't freed.
7966 *) Add text documentation for the BUFFER functions. Also added a work around
7967 to a Win95 console bug. This was triggered by the password read stuff: the
7968 last character typed gets carried over to the next fread(). If you were
7969 generating a new cert request using 'req' for example then the last
7970 character of the passphrase would be CR which would then enter the first
7974 *) Added the new `Includes OpenSSL Cryptography Software' button as
7975 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
7976 button and can be used by applications based on OpenSSL to show the
7977 relationship to the OpenSSL project.
7978 [Ralf S. Engelschall]
7980 *) Remove confusing variables in function signatures in files
7981 ssl/ssl_lib.c and ssl/ssl.h.
7982 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7984 *) Don't install bss_file.c under PREFIX/include/
7985 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7987 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
7988 functions that return function pointers and has support for NT specific
7989 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
7990 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
7991 unsigned to signed types: this was killing the Win32 compile.
7994 *) Add new certificate file to stack functions,
7995 SSL_add_dir_cert_subjects_to_stack() and
7996 SSL_add_file_cert_subjects_to_stack(). These largely supplant
7997 SSL_load_client_CA_file(), and can be used to add multiple certs easily
7998 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
7999 This means that Apache-SSL and similar packages don't have to mess around
8000 to add as many CAs as they want to the preferred list.
8003 *) Experiment with doxygen documentation. Currently only partially applied to
8005 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
8006 openssl.doxy as the configuration file.
8009 *) Get rid of remaining C++-style comments which strict C compilers hate.
8010 [Ralf S. Engelschall, pointed out by Carlos Amengual]
8012 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
8013 compiled in by default: it has problems with large keys.
8016 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
8017 DH private keys and/or callback functions which directly correspond to
8018 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
8019 is needed for applications which have to configure certificates on a
8020 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
8022 For the RSA certificate situation is makes no difference, but
8023 for the DSA certificate situation this fixes the "no shared cipher"
8024 problem where the OpenSSL cipher selection procedure failed because the
8025 temporary keys were not overtaken from the context and the API provided
8026 no way to reconfigure them.
8027 The new functions now let applications reconfigure the stuff and they
8028 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
8029 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
8030 non-public-API function ssl_cert_instantiate() is used as a helper
8031 function and also to reduce code redundancy inside ssl_rsa.c.
8032 [Ralf S. Engelschall]
8034 *) Move s_server -dcert and -dkey options out of the undocumented feature
8035 area because they are useful for the DSA situation and should be
8036 recognized by the users.
8037 [Ralf S. Engelschall]
8039 *) Fix the cipher decision scheme for export ciphers: the export bits are
8040 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
8041 SSL_EXP_MASK. So, the original variable has to be used instead of the
8042 already masked variable.
8043 [Richard Levitte <levitte@stacken.kth.se>]
8045 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
8046 [Richard Levitte <levitte@stacken.kth.se>]
8048 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
8049 from `int' to `unsigned int' because it's a length and initialized by
8050 EVP_DigestFinal() which expects an `unsigned int *'.
8051 [Richard Levitte <levitte@stacken.kth.se>]
8053 *) Don't hard-code path to Perl interpreter on shebang line of Configure
8054 script. Instead use the usual Shell->Perl transition trick.
8055 [Ralf S. Engelschall]
8057 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
8058 (in addition to RSA certificates) to match the behaviour of `openssl dsa
8059 -noout -modulus' as it's already the case for `openssl rsa -noout
8060 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
8061 currently the public key is printed (a decision which was already done by
8062 `openssl dsa -modulus' in the past) which serves a similar purpose.
8063 Additionally the NO_RSA no longer completely removes the whole -modulus
8064 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
8066 [Ralf S. Engelschall]
8068 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
8069 BIO. See the source (crypto/evp/bio_ok.c) for more info.
8070 [Arne Ansper <arne@ats.cyber.ee>]
8072 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
8073 to be added. Now both 'req' and 'ca' can use new objects defined in the
8077 *) Add cool BIO that does syslog (or event log on NT).
8078 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
8080 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
8081 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
8082 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
8083 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
8086 *) Add preliminary config info for new extension code.
8089 *) Make RSA_NO_PADDING really use no padding.
8090 [Ulf Moeller <ulf@fitug.de>]
8092 *) Generate errors when private/public key check is done.
8095 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
8096 for some CRL extensions and new objects added.
8099 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
8100 key usage extension and fuller support for authority key id.
8103 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
8104 padding method for RSA, which is recommended for new applications in PKCS
8105 #1 v2.0 (RFC 2437, October 1998).
8106 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
8107 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
8108 against Bleichbacher's attack on RSA.
8109 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
8112 *) Updates to the new SSL compression code
8113 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8115 *) Fix so that the version number in the master secret, when passed
8116 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
8117 (because the server will not accept higher), that the version number
8118 is 0x03,0x01, not 0x03,0x00
8119 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8121 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
8122 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
8123 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
8126 *) Support for RAW extensions where an arbitrary extension can be
8127 created by including its DER encoding. See apps/openssl.cnf for
8131 *) Make sure latest Perl versions don't interpret some generated C array
8132 code as Perl array code in the crypto/err/err_genc.pl script.
8133 [Lars Weber <3weber@informatik.uni-hamburg.de>]
8135 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
8136 not many people have the assembler. Various Win32 compilation fixes and
8137 update to the INSTALL.W32 file with (hopefully) more accurate Win32
8141 *) Modify configure script 'Configure' to automatically create crypto/date.h
8142 file under Win32 and also build pem.h from pem.org. New script
8143 util/mkfiles.pl to create the MINFO file on environments that can't do a
8144 'make files': perl util/mkfiles.pl >MINFO should work.
8147 *) Major rework of DES function declarations, in the pursuit of correctness
8148 and purity. As a result, many evil casts evaporated, and some weirdness,
8149 too. You may find this causes warnings in your code. Zapping your evil
8150 casts will probably fix them. Mostly.
8153 *) Fix for a typo in asn1.h. Bug fix to object creation script
8154 obj_dat.pl. It considered a zero in an object definition to mean
8155 "end of object": none of the objects in objects.h have any zeros
8156 so it wasn't spotted.
8157 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
8159 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
8160 Masking (CBCM). In the absence of test vectors, the best I have been able
8161 to do is check that the decrypt undoes the encrypt, so far. Send me test
8162 vectors if you have them.
8165 *) Correct calculation of key length for export ciphers (too much space was
8166 allocated for null ciphers). This has not been tested!
8169 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
8170 message is now correct (it understands "crypto" and "ssl" on its
8171 command line). There is also now an "update" option. This will update
8172 the util/ssleay.num and util/libeay.num files with any new functions.
8174 perl util/mkdef.pl crypto ssl update
8175 it will update them.
8178 *) Overhauled the Perl interface (perl/*):
8179 - ported BN stuff to OpenSSL's different BN library
8180 - made the perl/ source tree CVS-aware
8181 - renamed the package from SSLeay to OpenSSL (the files still contain
8182 their history because I've copied them in the repository)
8183 - removed obsolete files (the test scripts will be replaced
8184 by better Test::Harness variants in the future)
8185 [Ralf S. Engelschall]
8187 *) First cut for a very conservative source tree cleanup:
8188 1. merge various obsolete readme texts into doc/ssleay.txt
8189 where we collect the old documents and readme texts.
8190 2. remove the first part of files where I'm already sure that we no
8191 longer need them because of three reasons: either they are just temporary
8192 files which were left by Eric or they are preserved original files where
8193 I've verified that the diff is also available in the CVS via "cvs diff
8194 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
8195 the crypto/md/ stuff).
8196 [Ralf S. Engelschall]
8198 *) More extension code. Incomplete support for subject and issuer alt
8199 name, issuer and authority key id. Change the i2v function parameters
8200 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
8201 what that's for :-) Fix to ASN1 macro which messed up
8202 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
8205 *) Preliminary support for ENUMERATED type. This is largely copied from the
8209 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
8210 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8212 *) Make sure `make rehash' target really finds the `openssl' program.
8213 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8215 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
8216 like to hear about it if this slows down other processors.
8219 *) Add CygWin32 platform information to Configure script.
8220 [Alan Batie <batie@aahz.jf.intel.com>]
8222 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
8223 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
8225 *) New program nseq to manipulate netscape certificate sequences
8228 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
8232 *) Fixes to BN code. Previously the default was to define BN_RECURSION
8233 but the BN code had some problems that would cause failures when
8234 doing certificate verification and some other functions.
8235 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8237 *) Add ASN1 and PEM code to support netscape certificate sequences.
8240 *) Add ASN1 and PEM code to support netscape certificate sequences.
8243 *) Add several PKIX and private extended key usage OIDs.
8246 *) Modify the 'ca' program to handle the new extension code. Modify
8247 openssl.cnf for new extension format, add comments.
8250 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
8251 and add a sample to openssl.cnf so req -x509 now adds appropriate
8255 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
8256 error code, add initial support to X509_print() and x509 application.
8259 *) Takes a deep breath and start addding X509 V3 extension support code. Add
8260 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
8261 stuff is currently isolated and isn't even compiled yet.
8264 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
8265 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
8266 Removed the versions check from X509 routines when loading extensions:
8267 this allows certain broken certificates that don't set the version
8268 properly to be processed.
8271 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
8272 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
8273 can still be regenerated with "make depend".
8276 *) Spelling mistake in C version of CAST-128.
8277 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
8279 *) Changes to the error generation code. The perl script err-code.pl
8280 now reads in the old error codes and retains the old numbers, only
8281 adding new ones if necessary. It also only changes the .err files if new
8282 codes are added. The makefiles have been modified to only insert errors
8283 when needed (to avoid needlessly modifying header files). This is done
8284 by only inserting errors if the .err file is newer than the auto generated
8285 C file. To rebuild all the error codes from scratch (the old behaviour)
8286 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
8287 or delete all the .err files.
8290 *) CAST-128 was incorrectly implemented for short keys. The C version has
8291 been fixed, but is untested. The assembler versions are also fixed, but
8292 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
8293 to regenerate it if needed.
8294 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
8295 Hagino <itojun@kame.net>]
8297 *) File was opened incorrectly in randfile.c.
8298 [Ulf Möller <ulf@fitug.de>]
8300 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
8301 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
8302 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
8303 al: it's just almost always a UTCTime. Note this patch adds new error
8304 codes so do a "make errors" if there are problems.
8307 *) Correct Linux 1 recognition in config.
8308 [Ulf Möller <ulf@fitug.de>]
8310 *) Remove pointless MD5 hash when using DSA keys in ca.
8311 [Anonymous <nobody@replay.com>]
8313 *) Generate an error if given an empty string as a cert directory. Also
8314 generate an error if handed NULL (previously returned 0 to indicate an
8315 error, but didn't set one).
8316 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
8318 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
8321 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
8322 parameters. This was causing a warning which killed off the Win32 compile.
8325 *) Remove C++ style comments from crypto/bn/bn_local.h.
8326 [Neil Costigan <neil.costigan@celocom.com>]
8328 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
8329 based on a text string, looking up short and long names and finally
8330 "dot" format. The "dot" format stuff didn't work. Added new function
8331 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
8332 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
8333 OID is not part of the table.
8336 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
8337 X509_LOOKUP_by_alias().
8340 *) Sort openssl functions by name.
8343 *) Get the gendsa program working (hopefully) and add it to app list. Remove
8344 encryption from sample DSA keys (in case anyone is interested the password
8348 *) Make _all_ *_free functions accept a NULL pointer.
8349 [Frans Heymans <fheymans@isaserver.be>]
8351 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
8353 [Anonymous <nobody@replay.com>]
8355 *) s_server should send the CAfile as acceptable CAs, not its own cert.
8356 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8358 *) Don't blow it for numeric -newkey arguments to apps/req.
8359 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8361 *) Temp key "for export" tests were wrong in s3_srvr.c.
8362 [Anonymous <nobody@replay.com>]
8364 *) Add prototype for temp key callback functions
8365 SSL_CTX_set_tmp_{rsa,dh}_callback().
8368 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
8369 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
8372 *) X509_name_add_entry() freed the wrong thing after an error.
8373 [Arne Ansper <arne@ats.cyber.ee>]
8375 *) rsa_eay.c would attempt to free a NULL context.
8376 [Arne Ansper <arne@ats.cyber.ee>]
8378 *) BIO_s_socket() had a broken should_retry() on Windoze.
8379 [Arne Ansper <arne@ats.cyber.ee>]
8381 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
8382 [Arne Ansper <arne@ats.cyber.ee>]
8384 *) Make sure the already existing X509_STORE->depth variable is initialized
8385 in X509_STORE_new(), but document the fact that this variable is still
8386 unused in the certificate verification process.
8387 [Ralf S. Engelschall]
8389 *) Fix the various library and apps files to free up pkeys obtained from
8390 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
8393 *) Fix reference counting in X509_PUBKEY_get(). This makes
8394 demos/maurice/example2.c work, amongst others, probably.
8395 [Steve Henson and Ben Laurie]
8397 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
8398 `openssl' and second, the shortcut symlinks for the `openssl <command>'
8399 are no longer created. This way we have a single and consistent command
8400 line interface `openssl <command>', similar to `cvs <command>'.
8401 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
8403 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
8404 BIT STRING wrapper always have zero unused bits.
8407 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
8410 *) Make the top-level INSTALL documentation easier to understand.
8413 *) Makefiles updated to exit if an error occurs in a sub-directory
8414 make (including if user presses ^C) [Paul Sutton]
8416 *) Make Montgomery context stuff explicit in RSA data structure.
8419 *) Fix build order of pem and err to allow for generated pem.h.
8422 *) Fix renumbering bug in X509_NAME_delete_entry().
8425 *) Enhanced the err-ins.pl script so it makes the error library number
8426 global and can add a library name. This is needed for external ASN1 and
8427 other error libraries.
8430 *) Fixed sk_insert which never worked properly.
8433 *) Fix ASN1 macros so they can handle indefinite length construted
8434 EXPLICIT tags. Some non standard certificates use these: they can now
8438 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
8439 into a single doc/ssleay.txt bundle. This way the information is still
8440 preserved but no longer messes up this directory. Now it's new room for
8441 the new set of documenation files.
8442 [Ralf S. Engelschall]
8444 *) SETs were incorrectly DER encoded. This was a major pain, because they
8445 shared code with SEQUENCEs, which aren't coded the same. This means that
8446 almost everything to do with SETs or SEQUENCEs has either changed name or
8447 number of arguments.
8448 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
8450 *) Fix test data to work with the above.
8453 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
8454 was already fixed by Eric for 0.9.1 it seems.
8455 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
8457 *) Autodetect FreeBSD3.
8460 *) Fix various bugs in Configure. This affects the following platforms:
8464 unixware-2.0-pentium
8468 *) Eliminate generated files from CVS. Reorder tests to regenerate files
8469 before they are needed.
8472 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
8476 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
8478 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
8479 changed SSLeay to OpenSSL in version strings.
8480 [Ralf S. Engelschall]
8482 *) Some fixups to the top-level documents.
8485 *) Fixed the nasty bug where rsaref.h was not found under compile-time
8486 because the symlink to include/ was missing.
8487 [Ralf S. Engelschall]
8489 *) Incorporated the popular no-RSA/DSA-only patches
8490 which allow to compile a RSA-free SSLeay.
8491 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
8493 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
8494 when "ssleay" is still not found.
8495 [Ralf S. Engelschall]
8497 *) Added more platforms to Configure: Cray T3E, HPUX 11,
8498 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
8500 *) Updated the README file.
8501 [Ralf S. Engelschall]
8503 *) Added various .cvsignore files in the CVS repository subdirs
8504 to make a "cvs update" really silent.
8505 [Ralf S. Engelschall]
8507 *) Recompiled the error-definition header files and added
8508 missing symbols to the Win32 linker tables.
8509 [Ralf S. Engelschall]
8511 *) Cleaned up the top-level documents;
8512 o new files: CHANGES and LICENSE
8513 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
8514 o merged COPYRIGHT into LICENSE
8515 o removed obsolete TODO file
8516 o renamed MICROSOFT to INSTALL.W32
8517 [Ralf S. Engelschall]
8519 *) Removed dummy files from the 0.9.1b source tree:
8520 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
8521 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
8522 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
8523 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
8524 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
8525 [Ralf S. Engelschall]
8527 *) Added various platform portability fixes.
8530 *) The Genesis of the OpenSSL rpject:
8531 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
8532 Young and Tim J. Hudson created while they were working for C2Net until
8534 [The OpenSSL Project]
8537 Changes between 0.9.0b and 0.9.1b [not released]
8539 *) Updated a few CA certificates under certs/
8542 *) Changed some BIGNUM api stuff.
8545 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
8546 DGUX x86, Linux Alpha, etc.
8549 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
8550 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
8554 *) Add -strparse option to asn1pars program which parses nested
8556 [Dr Stephen Henson <shenson@bigfoot.com>]
8558 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
8561 *) DSA fix for "ca" program.
8564 *) Added "-genkey" option to "dsaparam" program.
8567 *) Added RIPE MD160 (rmd160) message digest.
8570 *) Added -a (all) option to "ssleay version" command.
8573 *) Added PLATFORM define which is the id given to Configure.
8576 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
8579 *) Extended the ASN.1 parser routines.
8582 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
8585 *) Added a BN_CTX to the BN library.
8588 *) Fixed the weak key values in DES library
8591 *) Changed API in EVP library for cipher aliases.
8594 *) Added support for RC2/64bit cipher.
8597 *) Converted the lhash library to the crypto/mem.c functions.
8600 *) Added more recognized ASN.1 object ids.
8603 *) Added more RSA padding checks for SSL/TLS.
8606 *) Added BIO proxy/filter functionality.
8609 *) Added extra_certs to SSL_CTX which can be used
8610 send extra CA certificates to the client in the CA cert chain sending
8611 process. It can be configured with SSL_CTX_add_extra_chain_cert().
8614 *) Now Fortezza is denied in the authentication phase because
8615 this is key exchange mechanism is not supported by SSLeay at all.
8618 *) Additional PKCS1 checks.
8621 *) Support the string "TLSv1" for all TLS v1 ciphers.
8624 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
8625 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
8628 *) Fixed a few memory leaks.
8631 *) Fixed various code and comment typos.
8634 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
8635 bytes sent in the client random.
8636 [Edward Bishop <ebishop@spyglass.com>]
projects / thirdparty / openssl.git / blob