5 Changes between 0.9.8u and 0.9.8v [xx XXX xxxx]
9 Changes between 0.9.8t and 0.9.8u [12 Mar 2012]
11 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
12 in CMS and PKCS7 code. When RSA decryption fails use a random key for
13 content decryption and always return the same error. Note: this attack
14 needs on average 2^20 messages so it only affects automated senders. The
15 old behaviour can be reenabled in the CMS code by setting the
16 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
17 an MMA defence is not necessary.
18 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
19 this issue. (CVE-2012-0884)
22 *) Fix CVE-2011-4619: make sure we really are receiving a
23 client hello before rejecting multiple SGC restarts. Thanks to
24 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
27 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
29 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
30 Thanks to Antonio Martin, Enterprise Secure Access Research and
31 Development, Cisco Systems, Inc. for discovering this bug and
32 preparing a fix. (CVE-2012-0050)
35 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
37 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
38 of the Vaudenay padding oracle attack on CBC mode encryption
39 which enables an efficient plaintext recovery attack against
40 the OpenSSL implementation of DTLS. Their attack exploits timing
41 differences arising during decryption processing. A research
42 paper describing this attack can be found at:
43 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
44 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
45 Security Group at Royal Holloway, University of London
46 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
47 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
48 for preparing the fix. (CVE-2011-4108)
49 [Robin Seggelmann, Michael Tuexen]
51 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
52 [Ben Laurie, Kasper <ekasper@google.com>]
54 *) Clear bytes used for block padding of SSL 3.0 records.
56 [Adam Langley (Google)]
58 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
59 Kadianakis <desnacked@gmail.com> for discovering this issue and
60 Adam Langley for preparing the fix. (CVE-2011-4619)
61 [Adam Langley (Google)]
63 *) Prevent malformed RFC3779 data triggering an assertion failure.
64 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
65 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
66 [Rob Austein <sra@hactrn.net>]
68 *) Fix ssl_ciph.c set-up race.
69 [Adam Langley (Google)]
71 *) Fix spurious failures in ecdsatest.c.
72 [Emilia Käsper (Google)]
74 *) Fix the BIO_f_buffer() implementation (which was mixing different
75 interpretations of the '..._len' fields).
76 [Adam Langley (Google)]
78 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
79 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
80 threads won't reuse the same blinding coefficients.
82 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
83 lock to call BN_BLINDING_invert_ex, and avoids one use of
84 BN_BLINDING_update for each BN_BLINDING structure (previously,
85 the last update always remained unused).
86 [Emilia Käsper (Google)]
88 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
89 for multi-threaded use of ECDH.
90 [Adam Langley (Google)]
92 *) Fix x509_name_ex_d2i memory leak on bad inputs.
95 *) Add protection against ECDSA timing attacks as mentioned in the paper
96 by Billy Bob Brumley and Nicola Tuveri, see:
98 http://eprint.iacr.org/2011/232.pdf
100 [Billy Bob Brumley and Nicola Tuveri]
102 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
104 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
105 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
107 *) Fix bug in string printing code: if *any* escaping is enabled we must
108 escape the escape character (backslash) or the resulting string is
112 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
114 *) Disable code workaround for ancient and obsolete Netscape browsers
115 and servers: an attacker can use it in a ciphersuite downgrade attack.
116 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
119 *) Fixed J-PAKE implementation error, originally discovered by
120 Sebastien Martini, further info and confirmation from Stefan
121 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
124 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
126 *) Fix extension code to avoid race conditions which can result in a buffer
127 overrun vulnerability: resumed sessions must not be modified as they can
128 be shared by multiple threads. CVE-2010-3864
131 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
134 *) Don't reencode certificate when calculating signature: cache and use
135 the original encoding instead. This makes signature verification of
136 some broken encodings work correctly.
139 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
140 is also one of the inputs.
141 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
143 *) Don't repeatedly append PBE algorithms to table if they already exist.
144 Sort table on each new add. This effectively makes the table read only
145 after all algorithms are added and subsequent calls to PKCS12_pbe_add
149 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
151 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
154 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
155 access or freeing data twice (CVE-2010-0742)
156 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
158 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
159 common in certificates and some applications which only call
160 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
164 Reduce copying into .apps and .test in makevms.com
165 Don't try to use blank CA certificate in CA.com
166 Allow use of C files from original directories in maketests.com
167 [Steven M. Schweda" <sms@antinode.info>]
169 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
171 *) When rejecting SSL/TLS records due to an incorrect version number, never
172 update s->server with a new major version number. As of
173 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
174 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
175 the previous behavior could result in a read attempt at NULL when
176 receiving specific incorrect SSL/TLS records once record payload
177 protection is active. (CVE-2010-0740)
178 [Bodo Moeller, Adam Langley <agl@chromium.org>]
180 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
181 could be crashed if the relevant tables were not present (e.g. chrooted).
182 [Tomas Hoger <thoger@redhat.com>]
184 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
186 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
187 [Martin Olsson, Neel Mehta]
189 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
190 accommodate for stack sorting, always a write lock!).
193 *) On some versions of WIN32 Heap32Next is very slow. This can cause
194 excessive delays in the RAND_poll(): over a minute. As a workaround
195 include a time check in the inner Heap32Next loop too.
198 *) The code that handled flushing of data in SSL/TLS originally used the
199 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
200 the problem outlined in PR#1949. The fix suggested there however can
201 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
202 of Apache). So instead simplify the code to flush unconditionally.
203 This should be fine since flushing with no data to flush is a no op.
206 *) Handle TLS versions 2.0 and later properly and correctly use the
207 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
208 off ancient servers have a habit of sticking around for a while...
211 *) Modify compression code so it frees up structures without using the
212 ex_data callbacks. This works around a problem where some applications
213 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
214 restarting) then use compression (e.g. SSL with compression) later.
215 This results in significant per-connection memory leaks and
216 has caused some security issues including CVE-2008-1678 and
220 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
221 change when encrypting or decrypting.
224 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
225 connect and renegotiate with servers which do not support RI.
226 Until RI is more widely deployed this option is enabled by default.
229 *) Add "missing" ssl ctrls to clear options and mode.
232 *) If client attempts to renegotiate and doesn't support RI respond with
233 a no_renegotiation alert as required by RFC5746. Some renegotiating
234 TLS clients will continue a connection gracefully when they receive
235 the alert. Unfortunately OpenSSL mishandled this alert and would hang
236 waiting for a server hello which it will never receive. Now we treat a
237 received no_renegotiation alert as a fatal error. This is because
238 applications requesting a renegotiation might well expect it to succeed
239 and would have no code in place to handle the server denying it so the
240 only safe thing to do is to terminate the connection.
243 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
244 peer supports secure renegotiation and 0 otherwise. Print out peer
245 renegotiation support in s_client/s_server.
248 *) Replace the highly broken and deprecated SPKAC certification method with
249 the updated NID creation version. This should correctly handle UTF8.
252 *) Implement RFC5746. Re-enable renegotiation but require the extension
253 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
254 turns out to be a bad idea. It has been replaced by
255 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
256 SSL_CTX_set_options(). This is really not recommended unless you
257 know what you are doing.
258 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
260 *) Fixes to stateless session resumption handling. Use initial_ctx when
261 issuing and attempting to decrypt tickets in case it has changed during
262 servername handling. Use a non-zero length session ID when attempting
263 stateless session resumption: this makes it possible to determine if
264 a resumption has occurred immediately after receiving server hello
265 (several places in OpenSSL subtly assume this) instead of later in
269 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
270 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
271 fixes for a few places where the return code is not checked
273 [Julia Lawall <julia@diku.dk>]
275 *) Add --strict-warnings option to Configure script to include devteam
276 warnings in other configurations.
279 *) Add support for --libdir option and LIBDIR variable in makefiles. This
280 makes it possible to install openssl libraries in locations which
281 have names other than "lib", for example "/usr/lib64" which some
283 [Steve Henson, based on patch from Jeremy Utley]
285 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
286 X690 8.9.12 and can produce some misleading textual output of OIDs.
287 [Steve Henson, reported by Dan Kaminsky]
289 *) Delete MD2 from algorithm tables. This follows the recommendation in
290 several standards that it is not used in new applications due to
291 several cryptographic weaknesses. For binary compatibility reasons
292 the MD2 API is still compiled in by default.
295 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
299 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
300 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
302 [Guenter <lists@gknw.net>]
304 *) Fix the server certificate chain building code to use X509_verify_cert(),
305 it used to have an ad-hoc builder which was unable to cope with anything
306 other than a simple chain.
307 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
309 *) Don't check self signed certificate signatures in X509_verify_cert()
310 by default (a flag can override this): it just wastes time without
311 adding any security. As a useful side effect self signed root CAs
312 with non-FIPS digests are now usable in FIPS mode.
315 *) In dtls1_process_out_of_seq_message() the check if the current message
316 is already buffered was missing. For every new message was memory
317 allocated, allowing an attacker to perform an denial of service attack
318 with sending out of seq handshake messages until there is no memory
319 left. Additionally every future messege was buffered, even if the
320 sequence number made no sense and would be part of another handshake.
321 So only messages with sequence numbers less than 10 in advance will be
322 buffered. (CVE-2009-1378)
323 [Robin Seggelmann, discovered by Daniel Mentz]
325 *) Records are buffered if they arrive with a future epoch to be
326 processed after finishing the corresponding handshake. There is
327 currently no limitation to this buffer allowing an attacker to perform
328 a DOS attack with sending records with future epochs until there is no
329 memory left. This patch adds the pqueue_size() function to detemine
330 the size of a buffer and limits the record buffer to 100 entries.
332 [Robin Seggelmann, discovered by Daniel Mentz]
334 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
335 parent structure is freed. (CVE-2009-1379)
338 *) Handle non-blocking I/O properly in SSL_shutdown() call.
339 [Darryl Miles <darryl-mailinglists@netbauds.net>]
342 [Ilya O. <vrghost@gmail.com>]
344 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
346 *) Disable renegotiation completely - this fixes a severe security
347 problem (CVE-2009-3555) at the cost of breaking all
348 renegotiation. Renegotiation can be re-enabled by setting
349 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
350 run-time. This is really not recommended unless you know what
354 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
356 *) Don't set val to NULL when freeing up structures, it is freed up by
357 underlying code. If sizeof(void *) > sizeof(long) this can result in
358 zeroing past the valid field. (CVE-2009-0789)
359 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
361 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
362 checked correctly. This would allow some invalid signed attributes to
363 appear to verify correctly. (CVE-2009-0591)
364 [Ivan Nestlerode <inestlerode@us.ibm.com>]
366 *) Reject UniversalString and BMPString types with invalid lengths. This
367 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
368 a legal length. (CVE-2009-0590)
371 *) Set S/MIME signing as the default purpose rather than setting it
372 unconditionally. This allows applications to override it at the store
376 *) Permit restricted recursion of ASN1 strings. This is needed in practice
377 to handle some structures.
380 *) Improve efficiency of mem_gets: don't search whole buffer each time
382 [Jeremy Shapiro <jnshapir@us.ibm.com>]
384 *) New -hex option for openssl rand.
387 *) Print out UTF8String and NumericString when parsing ASN1.
390 *) Support NumericString type for name components.
393 *) Allow CC in the environment to override the automatically chosen
394 compiler. Note that nothing is done to ensure flags work with the
398 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
400 *) Properly check EVP_VerifyFinal() and similar return values
402 [Ben Laurie, Bodo Moeller, Google Security Team]
404 *) Enable TLS extensions by default.
407 *) Allow the CHIL engine to be loaded, whether the application is
408 multithreaded or not. (This does not release the developer from the
409 obligation to set up the dynamic locking callbacks.)
410 [Sander Temme <sander@temme.net>]
412 *) Use correct exit code if there is an error in dgst command.
413 [Steve Henson; problem pointed out by Roland Dirlewanger]
415 *) Tweak Configure so that you need to say "experimental-jpake" to enable
416 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
419 *) Add experimental JPAKE support, including demo authentication in
420 s_client and s_server.
423 *) Set the comparison function in v3_addr_canonize().
424 [Rob Austein <sra@hactrn.net>]
426 *) Add support for XMPP STARTTLS in s_client.
427 [Philip Paeps <philip@freebsd.org>]
429 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
430 to ensure that even with this option, only ciphersuites in the
431 server's preference list will be accepted. (Note that the option
432 applies only when resuming a session, so the earlier behavior was
433 just about the algorithm choice for symmetric cryptography.)
436 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
438 *) Fix NULL pointer dereference if a DTLS server received
439 ChangeCipherSpec as first record (CVE-2009-1386).
442 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
443 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
446 *) The fix in 0.9.8c that supposedly got rid of unsafe
447 double-checked locking was incomplete for RSA blinding,
448 addressing just one layer of what turns out to have been
449 doubly unsafe triple-checked locking.
451 So now fix this for real by retiring the MONT_HELPER macro
452 in crypto/rsa/rsa_eay.c.
454 [Bodo Moeller; problem pointed out by Marius Schilder]
456 *) Various precautionary measures:
458 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
460 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
461 (NB: This would require knowledge of the secret session ticket key
462 to exploit, in which case you'd be SOL either way.)
464 - Change bn_nist.c so that it will properly handle input BIGNUMs
465 outside the expected range.
467 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
470 [Neel Mehta, Bodo Moeller]
472 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
473 the load fails. Useful for distros.
474 [Ben Laurie and the FreeBSD team]
476 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
479 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
482 *) Expand ENGINE to support engine supplied SSL client certificate functions.
484 This work was sponsored by Logica.
487 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
488 keystores. Support for SSL/TLS client authentication too.
489 Not compiled unless enable-capieng specified to Configure.
491 This work was sponsored by Logica.
494 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
495 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
496 attribute creation routines such as certifcate requests and PKCS#12
500 Changes between 0.9.8g and 0.9.8h [28 May 2008]
502 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
503 handshake which could lead to a cilent crash as found using the
504 Codenomicon TLS test suite (CVE-2008-1672)
505 [Steve Henson, Mark Cox]
507 *) Fix double free in TLS server name extensions which could lead to
508 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
511 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
513 Clear the error queue to ensure that error entries left from
514 older function calls do not interfere with the correct operation.
515 [Lutz Jaenicke, Erik de Castro Lopo]
517 *) Remove root CA certificates of commercial CAs:
519 The OpenSSL project does not recommend any specific CA and does not
520 have any policy with respect to including or excluding any CA.
521 Therefore it does not make any sense to ship an arbitrary selection
522 of root CA certificates with the OpenSSL software.
525 *) RSA OAEP patches to fix two separate invalid memory reads.
526 The first one involves inputs when 'lzero' is greater than
527 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
528 before the beginning of from). The second one involves inputs where
529 the 'db' section contains nothing but zeroes (there is a one-byte
530 invalid read after the end of 'db').
531 [Ivan Nestlerode <inestlerode@us.ibm.com>]
533 *) Partial backport from 0.9.9-dev:
535 Introduce bn_mul_mont (dedicated Montgomery multiplication
536 procedure) as a candidate for BIGNUM assembler implementation.
537 While 0.9.9-dev uses assembler for various architectures, only
538 x86_64 is available by default here in the 0.9.8 branch, and
539 32-bit x86 is available through a compile-time setting.
541 To try the 32-bit x86 assembler implementation, use Configure
542 option "enable-montasm" (which exists only for this backport).
544 As "enable-montasm" for 32-bit x86 disclaims code stability
545 anyway, in this constellation we activate additional code
546 backported from 0.9.9-dev for further performance improvements,
547 namely BN_from_montgomery_word. (To enable this otherwise,
548 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
550 [Andy Polyakov (backport partially by Bodo Moeller)]
552 *) Add TLS session ticket callback. This allows an application to set
553 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
554 values. This is useful for key rollover for example where several key
555 sets may exist with different names.
558 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
559 This was broken until now in 0.9.8 releases, such that the only way
560 a registered ENGINE could be used (assuming it initialises
561 successfully on the host) was to explicitly set it as the default
562 for the relevant algorithms. This is in contradiction with 0.9.7
563 behaviour and the documentation. With this fix, when an ENGINE is
564 registered into a given algorithm's table of implementations, the
565 'uptodate' flag is reset so that auto-discovery will be used next
566 time a new context for that algorithm attempts to select an
568 [Ian Lister (tweaked by Geoff Thorpe)]
570 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
571 implemention in the following ways:
573 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
576 Lack of BER streaming support means one pass streaming processing is
577 only supported if data is detached: setting the streaming flag is
578 ignored for embedded content.
580 CMS support is disabled by default and must be explicitly enabled
581 with the enable-cms configuration option.
584 *) Update the GMP engine glue to do direct copies between BIGNUM and
585 mpz_t when openssl and GMP use the same limb size. Otherwise the
586 existing "conversion via a text string export" trick is still used.
587 [Paul Sheer <paulsheer@gmail.com>]
589 *) Zlib compression BIO. This is a filter BIO which compressed and
590 uncompresses any data passed through it.
593 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
594 RFC3394 compatible AES key wrapping.
597 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
598 sets string data without copying. X509_ALGOR_set0() and
599 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
600 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
601 from an X509_ATTRIBUTE structure optionally checking it occurs only
602 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
606 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
607 to get the expected BN_FLG_CONSTTIME behavior.
608 [Bodo Moeller (Google)]
612 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
613 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
614 - added some more tests to do_tests.pl
615 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
616 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
617 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
618 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
619 - various changes to netware.pl to enable gcc-cross builds on Win32
621 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
622 - various changes to fix missing prototype warnings
623 - fixed x86nasm.pl to create correct asm files for NASM COFF output
624 - added AES, WHIRLPOOL and CPUID assembler code to build files
625 - added missing AES assembler make rules to mk1mf.pl
626 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
627 [Guenter Knauf <eflash@gmx.net>]
629 *) Implement certificate status request TLS extension defined in RFC3546.
630 A client can set the appropriate parameters and receive the encoded
631 OCSP response via a callback. A server can query the supplied parameters
632 and set the encoded OCSP response in the callback. Add simplified examples
633 to s_client and s_server.
636 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
639 + Binary incompatibility of ssl_ctx_st structure
640 + DTLS interoperation with non-compliant servers
641 + Don't call get_session_cb() without proposed session
642 + Fix ia64 assembler code
643 [Andy Polyakov, Steve Henson]
645 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
647 *) DTLS Handshake overhaul. There were longstanding issues with
648 OpenSSL DTLS implementation, which were making it impossible for
649 RFC 4347 compliant client to communicate with OpenSSL server.
650 Unfortunately just fixing these incompatibilities would "cut off"
651 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
652 server keeps tolerating non RFC compliant syntax. The opposite is
653 not true, 0.9.8f client can not communicate with earlier server.
654 This update even addresses CVE-2007-4995.
657 *) Changes to avoid need for function casts in OpenSSL: some compilers
658 (gcc 4.2 and later) reject their use.
659 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
662 *) Add RFC4507 support to OpenSSL. This includes the corrections in
663 RFC4507bis. The encrypted ticket format is an encrypted encoded
664 SSL_SESSION structure, that way new session features are automatically
667 If a client application caches session in an SSL_SESSION structure
668 support is transparent because tickets are now stored in the encoded
671 The SSL_CTX structure automatically generates keys for ticket
672 protection in servers so again support should be possible
673 with no application modification.
675 If a client or server wishes to disable RFC4507 support then the option
676 SSL_OP_NO_TICKET can be set.
678 Add a TLS extension debugging callback to allow the contents of any client
679 or server extensions to be examined.
681 This work was sponsored by Google.
684 *) Add initial support for TLS extensions, specifically for the server_name
685 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
686 have new members for a host name. The SSL data structure has an
687 additional member SSL_CTX *initial_ctx so that new sessions can be
688 stored in that context to allow for session resumption, even after the
689 SSL has been switched to a new SSL_CTX in reaction to a client's
690 server_name extension.
692 New functions (subject to change):
695 SSL_get_servername_type()
698 New CTRL codes and macros (subject to change):
700 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
701 - SSL_CTX_set_tlsext_servername_callback()
702 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
703 - SSL_CTX_set_tlsext_servername_arg()
704 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
706 openssl s_client has a new '-servername ...' option.
708 openssl s_server has new options '-servername_host ...', '-cert2 ...',
709 '-key2 ...', '-servername_fatal' (subject to change). This allows
710 testing the HostName extension for a specific single host name ('-cert'
711 and '-key' remain fallbacks for handshakes without HostName
712 negotiation). If the unrecogninzed_name alert has to be sent, this by
713 default is a warning; it becomes fatal with the '-servername_fatal'
716 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
718 *) Add AES and SSE2 assembly language support to VC++ build.
721 *) Mitigate attack on final subtraction in Montgomery reduction.
724 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
725 (which previously caused an internal error).
728 *) Squeeze another 10% out of IGE mode when in != out.
731 *) AES IGE mode speedup.
732 [Dean Gaudet (Google)]
734 *) Add the Korean symmetric 128-bit cipher SEED (see
735 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
736 add SEED ciphersuites from RFC 4162:
738 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
739 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
740 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
741 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
743 To minimize changes between patchlevels in the OpenSSL 0.9.8
744 series, SEED remains excluded from compilation unless OpenSSL
745 is configured with 'enable-seed'.
748 *) Mitigate branch prediction attacks, which can be practical if a
749 single processor is shared, allowing a spy process to extract
750 information. For detailed background information, see
751 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
752 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
753 and Necessary Software Countermeasures"). The core of the change
754 are new versions BN_div_no_branch() and
755 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
756 respectively, which are slower, but avoid the security-relevant
757 conditional branches. These are automatically called by BN_div()
758 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
759 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
760 remove a conditional branch.
762 BN_FLG_CONSTTIME is the new name for the previous
763 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
764 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
765 in the exponent causes BN_mod_exp_mont() to use the alternative
766 implementation in BN_mod_exp_mont_consttime().) The old name
767 remains as a deprecated alias.
769 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
770 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
771 constant-time implementations for more than just exponentiation.
772 Here too the old name is kept as a deprecated alias.
774 BN_BLINDING_new() will now use BN_dup() for the modulus so that
775 the BN_BLINDING structure gets an independent copy of the
776 modulus. This means that the previous "BIGNUM *m" argument to
777 BN_BLINDING_new() and to BN_BLINDING_create_param() now
778 essentially becomes "const BIGNUM *m", although we can't actually
779 change this in the header file before 0.9.9. It allows
780 RSA_setup_blinding() to use BN_with_flags() on the modulus to
781 enable BN_FLG_CONSTTIME.
783 [Matthew D Wood (Intel Corp)]
785 *) In the SSL/TLS server implementation, be strict about session ID
786 context matching (which matters if an application uses a single
787 external cache for different purposes). Previously,
788 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
789 set. This did ensure strict client verification, but meant that,
790 with applications using a single external cache for quite
791 different requirements, clients could circumvent ciphersuite
792 restrictions for a given session ID context by starting a session
793 in a different context.
796 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
797 a ciphersuite string such as "DEFAULT:RSA" cannot enable
798 authentication-only ciphersuites.
801 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
802 not complete and could lead to a possible single byte overflow
803 (CVE-2007-5135) [Ben Laurie]
805 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
807 *) Since AES128 and AES256 (and similarly Camellia128 and
808 Camellia256) share a single mask bit in the logic of
809 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
810 kludge to work properly if AES128 is available and AES256 isn't
811 (or if Camellia128 is available and Camellia256 isn't).
814 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
815 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
816 When a point or a seed is encoded in a BIT STRING, we need to
817 prevent the removal of trailing zero bits to get the proper DER
818 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
819 of a NamedBitList, for which trailing 0 bits need to be removed.)
822 *) Have SSL/TLS server implementation tolerate "mismatched" record
823 protocol version while receiving ClientHello even if the
824 ClientHello is fragmented. (The server can't insist on the
825 particular protocol version it has chosen before the ServerHello
826 message has informed the client about his choice.)
829 *) Add RFC 3779 support.
830 [Rob Austein for ARIN, Ben Laurie]
832 *) Load error codes if they are not already present instead of using a
833 static variable. This allows them to be cleanly unloaded and reloaded.
834 Improve header file function name parsing.
837 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
838 or CAPABILITY handshake as required by RFCs.
841 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
843 *) Introduce limits to prevent malicious keys being able to
844 cause a denial of service. (CVE-2006-2940)
845 [Steve Henson, Bodo Moeller]
847 *) Fix ASN.1 parsing of certain invalid structures that can result
848 in a denial of service. (CVE-2006-2937) [Steve Henson]
850 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
851 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
853 *) Fix SSL client code which could crash if connecting to a
854 malicious SSLv2 server. (CVE-2006-4343)
855 [Tavis Ormandy and Will Drewry, Google Security Team]
857 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
858 match only those. Before that, "AES256-SHA" would be interpreted
859 as a pattern and match "AES128-SHA" too (since AES128-SHA got
860 the same strength classification in 0.9.7h) as we currently only
861 have a single AES bit in the ciphersuite description bitmap.
862 That change, however, also applied to ciphersuite strings such as
863 "RC4-MD5" that intentionally matched multiple ciphersuites --
864 namely, SSL 2.0 ciphersuites in addition to the more common ones
865 from SSL 3.0/TLS 1.0.
867 So we change the selection algorithm again: Naming an explicit
868 ciphersuite selects this one ciphersuite, and any other similar
869 ciphersuite (same bitmap) from *other* protocol versions.
870 Thus, "RC4-MD5" again will properly select both the SSL 2.0
871 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
873 Since SSL 2.0 does not have any ciphersuites for which the
874 128/256 bit distinction would be relevant, this works for now.
875 The proper fix will be to use different bits for AES128 and
876 AES256, which would have avoided the problems from the beginning;
877 however, bits are scarce, so we can only do this in a new release
878 (not just a patchlevel) when we can change the SSL_CIPHER
879 definition to split the single 'unsigned long mask' bitmap into
880 multiple values to extend the available space.
884 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
886 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
887 (CVE-2006-4339) [Ben Laurie and Google Security Team]
889 *) Add AES IGE and biIGE modes.
892 *) Change the Unix randomness entropy gathering to use poll() when
893 possible instead of select(), since the latter has some
894 undesirable limitations.
895 [Darryl Miles via Richard Levitte and Bodo Moeller]
897 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
898 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
899 cannot be implicitly activated as part of, e.g., the "AES" alias.
900 However, please upgrade to OpenSSL 0.9.9[-dev] for
901 non-experimental use of the ECC ciphersuites to get TLS extension
902 support, which is required for curve and point format negotiation
903 to avoid potential handshake problems.
906 *) Disable rogue ciphersuites:
908 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
909 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
910 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
912 The latter two were purportedly from
913 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
916 Also deactivate the remaining ciphersuites from
917 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
918 unofficial, and the ID has long expired.
921 *) Fix RSA blinding Heisenbug (problems sometimes occured on
922 dual-core machines) and other potential thread-safety issues.
925 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
926 versions), which is now available for royalty-free use
927 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
928 Also, add Camellia TLS ciphersuites from RFC 4132.
930 To minimize changes between patchlevels in the OpenSSL 0.9.8
931 series, Camellia remains excluded from compilation unless OpenSSL
932 is configured with 'enable-camellia'.
935 *) Disable the padding bug check when compression is in use. The padding
936 bug check assumes the first packet is of even length, this is not
937 necessarily true if compresssion is enabled and can result in false
938 positives causing handshake failure. The actual bug test is ancient
939 code so it is hoped that implementations will either have fixed it by
940 now or any which still have the bug do not support compression.
943 Changes between 0.9.8a and 0.9.8b [04 May 2006]
945 *) When applying a cipher rule check to see if string match is an explicit
946 cipher suite and only match that one cipher suite if it is.
949 *) Link in manifests for VC++ if needed.
950 [Austin Ziegler <halostatue@gmail.com>]
952 *) Update support for ECC-based TLS ciphersuites according to
953 draft-ietf-tls-ecc-12.txt with proposed changes (but without
954 TLS extensions, which are supported starting with the 0.9.9
955 branch, not in the OpenSSL 0.9.8 branch).
958 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
959 opaque EVP_CIPHER_CTX handling.
962 *) Fixes and enhancements to zlib compression code. We now only use
963 "zlib1.dll" and use the default __cdecl calling convention on Win32
964 to conform with the standards mentioned here:
965 http://www.zlib.net/DLL_FAQ.txt
966 Static zlib linking now works on Windows and the new --with-zlib-include
967 --with-zlib-lib options to Configure can be used to supply the location
968 of the headers and library. Gracefully handle case where zlib library
972 *) Several fixes and enhancements to the OID generation code. The old code
973 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
974 handle numbers larger than ULONG_MAX, truncated printing and had a
975 non standard OBJ_obj2txt() behaviour.
978 *) Add support for building of engines under engine/ as shared libraries
979 under VC++ build system.
982 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
983 Hopefully, we will not see any false combination of paths any more.
986 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
988 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
989 (part of SSL_OP_ALL). This option used to disable the
990 countermeasure against man-in-the-middle protocol-version
991 rollback in the SSL 2.0 server implementation, which is a bad
992 idea. (CVE-2005-2969)
994 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
995 for Information Security, National Institute of Advanced Industrial
996 Science and Technology [AIST], Japan)]
998 *) Add two function to clear and return the verify parameter flags.
1001 *) Keep cipherlists sorted in the source instead of sorting them at
1002 runtime, thus removing the need for a lock.
1005 *) Avoid some small subgroup attacks in Diffie-Hellman.
1006 [Nick Mathewson and Ben Laurie]
1008 *) Add functions for well-known primes.
1011 *) Extended Windows CE support.
1012 [Satoshi Nakamura and Andy Polyakov]
1014 *) Initialize SSL_METHOD structures at compile time instead of during
1015 runtime, thus removing the need for a lock.
1018 *) Make PKCS7_decrypt() work even if no certificate is supplied by
1019 attempting to decrypt each encrypted key in turn. Add support to
1023 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
1025 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
1028 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
1031 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
1032 key into the same file any more.
1035 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
1038 *) Add -utf8 command line and config file option to 'ca'.
1039 [Stefan <stf@udoma.org]
1041 *) Removed the macro des_crypt(), as it seems to conflict with some
1042 libraries. Use DES_crypt().
1045 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
1046 involves renaming the source and generated shared-libs for
1047 both. The engines will accept the corrected or legacy ids
1048 ('ncipher' and '4758_cca' respectively) when binding. NB,
1049 this only applies when building 'shared'.
1050 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
1052 *) Add attribute functions to EVP_PKEY structure. Modify
1053 PKCS12_create() to recognize a CSP name attribute and
1054 use it. Make -CSP option work again in pkcs12 utility.
1057 *) Add new functionality to the bn blinding code:
1058 - automatic re-creation of the BN_BLINDING parameters after
1059 a fixed number of uses (currently 32)
1060 - add new function for parameter creation
1061 - introduce flags to control the update behaviour of the
1062 BN_BLINDING parameters
1063 - hide BN_BLINDING structure
1064 Add a second BN_BLINDING slot to the RSA structure to improve
1065 performance when a single RSA object is shared among several
1069 *) Add support for DTLS.
1070 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
1072 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
1073 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
1076 *) Remove buggy and incompletet DH cert support from
1077 ssl/ssl_rsa.c and ssl/s3_both.c
1080 *) Use SHA-1 instead of MD5 as the default digest algorithm for
1081 the apps/openssl applications.
1084 *) Compile clean with "-Wall -Wmissing-prototypes
1085 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
1086 DEBUG_SAFESTACK must also be set.
1089 *) Change ./Configure so that certain algorithms can be disabled by default.
1090 The new counterpiece to "no-xxx" is "enable-xxx".
1092 The patented RC5 and MDC2 algorithms will now be disabled unless
1093 "enable-rc5" and "enable-mdc2", respectively, are specified.
1095 (IDEA remains enabled despite being patented. This is because IDEA
1096 is frequently required for interoperability, and there is no license
1097 fee for non-commercial use. As before, "no-idea" can be used to
1098 avoid this algorithm.)
1102 *) Add processing of proxy certificates (see RFC 3820). This work was
1103 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
1104 EGEE (Enabling Grids for E-science in Europe).
1107 *) RC4 performance overhaul on modern architectures/implementations, such
1108 as Intel P4, IA-64 and AMD64.
1111 *) New utility extract-section.pl. This can be used specify an alternative
1112 section number in a pod file instead of having to treat each file as
1113 a separate case in Makefile. This can be done by adding two lines to the
1116 =for comment openssl_section:XXX
1118 The blank line is mandatory.
1122 *) New arguments -certform, -keyform and -pass for s_client and s_server
1123 to allow alternative format key and certificate files and passphrase
1127 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
1128 update associated structures and add various utility functions.
1130 Add new policy related verify parameters, include policy checking in
1131 standard verify code. Enhance 'smime' application with extra parameters
1132 to support policy checking and print out.
1135 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
1136 Nehemiah processors. These extensions support AES encryption in hardware
1137 as well as RNG (though RNG support is currently disabled).
1138 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
1140 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
1143 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
1144 [Andy Polyakov and a number of other people]
1146 *) Improved PowerPC platform support. Most notably BIGNUM assembler
1147 implementation contributed by IBM.
1148 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
1150 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
1151 exponent rather than 'unsigned long'. There is a corresponding change to
1152 the new 'rsa_keygen' element of the RSA_METHOD structure.
1153 [Jelte Jansen, Geoff Thorpe]
1155 *) Functionality for creating the initial serial number file is now
1156 moved from CA.pl to the 'ca' utility with a new option -create_serial.
1158 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
1159 number file to 1, which is bound to cause problems. To avoid
1160 the problems while respecting compatibility between different 0.9.7
1161 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
1162 CA.pl for serial number initialization. With the new release 0.9.8,
1163 we can fix the problem directly in the 'ca' utility.)
1166 *) Reduced header interdepencies by declaring more opaque objects in
1167 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
1168 give fewer recursive includes, which could break lazy source code - so
1169 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
1170 developers should define this symbol when building and using openssl to
1171 ensure they track the recommended behaviour, interfaces, [etc], but
1172 backwards-compatible behaviour prevails when this isn't defined.
1175 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
1178 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
1179 This will generate a random key of the appropriate length based on the
1180 cipher context. The EVP_CIPHER can provide its own random key generation
1181 routine to support keys of a specific form. This is used in the des and
1182 3des routines to generate a key of the correct parity. Update S/MIME
1183 code to use new functions and hence generate correct parity DES keys.
1184 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
1185 valid (weak or incorrect parity).
1188 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
1189 as looking them up. This is useful when the verified structure may contain
1190 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
1191 present unless the new PKCS7_NO_CRL flag is asserted.
1194 *) Extend ASN1 oid configuration module. It now additionally accepts the
1197 shortName = some long name, 1.2.3.4
1200 *) Reimplemented the BN_CTX implementation. There is now no more static
1201 limitation on the number of variables it can handle nor the depth of the
1202 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
1203 information can now expand as required, and rather than having a single
1204 static array of bignums, BN_CTX now uses a linked-list of such arrays
1205 allowing it to expand on demand whilst maintaining the usefulness of
1206 BN_CTX's "bundling".
1209 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
1210 to allow all RSA operations to function using a single BN_CTX.
1213 *) Preliminary support for certificate policy evaluation and checking. This
1214 is initially intended to pass the tests outlined in "Conformance Testing
1215 of Relying Party Client Certificate Path Processing Logic" v1.07.
1218 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
1219 remained unused and not that useful. A variety of other little bignum
1220 tweaks and fixes have also been made continuing on from the audit (see
1224 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
1225 associated ASN1, EVP and SSL functions and old ASN1 macros.
1228 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
1229 and this should never fail. So the return value from the use of
1230 BN_set_word() (which can fail due to needless expansion) is now deprecated;
1231 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
1234 *) BN_CTX_get() should return zero-valued bignums, providing the same
1235 initialised value as BN_new().
1236 [Geoff Thorpe, suggested by Ulf Möller]
1238 *) Support for inhibitAnyPolicy certificate extension.
1241 *) An audit of the BIGNUM code is underway, for which debugging code is
1242 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
1243 is considered valid when processing BIGNUMs, and causes execution to
1244 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
1245 further steps are taken to deliberately pollute unused data in BIGNUM
1246 structures to try and expose faulty code further on. For now, openssl will
1247 (in its default mode of operation) continue to tolerate the inconsistent
1248 forms that it has tolerated in the past, but authors and packagers should
1249 consider trying openssl and their own applications when compiled with
1250 these debugging symbols defined. It will help highlight potential bugs in
1251 their own code, and will improve the test coverage for OpenSSL itself. At
1252 some point, these tighter rules will become openssl's default to improve
1253 maintainability, though the assert()s and other overheads will remain only
1254 in debugging configurations. See bn.h for more details.
1255 [Geoff Thorpe, Nils Larsch, Ulf Möller]
1257 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
1258 that can only be obtained through BN_CTX_new() (which implicitly
1259 initialises it). The presence of this function only made it possible
1260 to overwrite an existing structure (and cause memory leaks).
1263 *) Because of the callback-based approach for implementing LHASH as a
1264 template type, lh_insert() adds opaque objects to hash-tables and
1265 lh_doall() or lh_doall_arg() are typically used with a destructor callback
1266 to clean up those corresponding objects before destroying the hash table
1267 (and losing the object pointers). So some over-zealous constifications in
1268 LHASH have been relaxed so that lh_insert() does not take (nor store) the
1269 objects as "const" and the lh_doall[_arg] callback wrappers are not
1270 prototyped to have "const" restrictions on the object pointers they are
1271 given (and so aren't required to cast them away any more).
1274 *) The tmdiff.h API was so ugly and minimal that our own timing utility
1275 (speed) prefers to use its own implementation. The two implementations
1276 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
1277 its object type properly exposed (MS_TM) instead of casting to/from "char
1278 *". This may still change yet if someone realises MS_TM and "ms_time_***"
1279 aren't necessarily the greatest nomenclatures - but this is what was used
1280 internally to the implementation so I've used that for now.
1283 *) Ensure that deprecated functions do not get compiled when
1284 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
1285 the self-tests were still using deprecated key-generation functions so
1286 these have been updated also.
1289 *) Reorganise PKCS#7 code to separate the digest location functionality
1290 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
1291 New function PKCS7_set_digest() to set the digest type for PKCS#7
1292 digestedData type. Add additional code to correctly generate the
1293 digestedData type and add support for this type in PKCS7 initialization
1297 *) New function PKCS7_set0_type_other() this initializes a PKCS7
1298 structure of type "other".
1301 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
1302 sure the loop does correctly stop and breaking ("division by zero")
1303 modulus operations are not performed. The (pre-generated) prime
1304 table crypto/bn/bn_prime.h was already correct, but it could not be
1305 re-generated on some platforms because of the "division by zero"
1306 situation in the script.
1307 [Ralf S. Engelschall]
1309 *) Update support for ECC-based TLS ciphersuites according to
1310 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
1311 SHA-1 now is only used for "small" curves (where the
1312 representation of a field element takes up to 24 bytes); for
1313 larger curves, the field element resulting from ECDH is directly
1314 used as premaster secret.
1315 [Douglas Stebila (Sun Microsystems Laboratories)]
1317 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
1318 curve secp160r1 to the tests.
1319 [Douglas Stebila (Sun Microsystems Laboratories)]
1321 *) Add the possibility to load symbols globally with DSO.
1322 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
1324 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
1325 control of the error stack.
1328 *) Add support for STORE in ENGINE.
1331 *) Add the STORE type. The intention is to provide a common interface
1332 to certificate and key stores, be they simple file-based stores, or
1333 HSM-type store, or LDAP stores, or...
1334 NOTE: The code is currently UNTESTED and isn't really used anywhere.
1337 *) Add a generic structure called OPENSSL_ITEM. This can be used to
1338 pass a list of arguments to any function as well as provide a way
1339 for a function to pass data back to the caller.
1342 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
1343 works like BUF_strdup() but can be used to duplicate a portion of
1344 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
1348 *) Add the function sk_find_ex() which works like sk_find(), but will
1349 return an index to an element even if an exact match couldn't be
1350 found. The index is guaranteed to point at the element where the
1351 searched-for key would be inserted to preserve sorting order.
1354 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
1355 takes an extra flags argument for optional functionality. Currently,
1356 the following flags are defined:
1358 OBJ_BSEARCH_VALUE_ON_NOMATCH
1359 This one gets OBJ_bsearch_ex() to return a pointer to the first
1360 element where the comparing function returns a negative or zero
1363 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
1364 This one gets OBJ_bsearch_ex() to return a pointer to the first
1365 element where the comparing function returns zero. This is useful
1366 if there are more than one element where the comparing function
1370 *) Make it possible to create self-signed certificates with 'openssl ca'
1371 in such a way that the self-signed certificate becomes part of the
1372 CA database and uses the same mechanisms for serial number generation
1373 as all other certificate signing. The new flag '-selfsign' enables
1374 this functionality. Adapt CA.sh and CA.pl.in.
1377 *) Add functionality to check the public key of a certificate request
1378 against a given private. This is useful to check that a certificate
1379 request can be signed by that key (self-signing).
1382 *) Make it possible to have multiple active certificates with the same
1383 subject in the CA index file. This is done only if the keyword
1384 'unique_subject' is set to 'no' in the main CA section (default
1385 if 'CA_default') of the configuration file. The value is saved
1386 with the database itself in a separate index attribute file,
1387 named like the index file with '.attr' appended to the name.
1390 *) Generate muti valued AVAs using '+' notation in config files for
1394 *) Support for nameConstraints certificate extension.
1397 *) Support for policyConstraints certificate extension.
1400 *) Support for policyMappings certificate extension.
1403 *) Make sure the default DSA_METHOD implementation only uses its
1404 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
1405 and change its own handlers to be NULL so as to remove unnecessary
1406 indirection. This lets alternative implementations fallback to the
1407 default implementation more easily.
1410 *) Support for directoryName in GeneralName related extensions
1414 *) Make it possible to link applications using Makefile.shared.
1415 Make that possible even when linking against static libraries!
1418 *) Support for single pass processing for S/MIME signing. This now
1419 means that S/MIME signing can be done from a pipe, in addition
1420 cleartext signing (multipart/signed type) is effectively streaming
1421 and the signed data does not need to be all held in memory.
1423 This is done with a new flag PKCS7_STREAM. When this flag is set
1424 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
1425 is done after the data is output (and digests calculated) in
1426 SMIME_write_PKCS7().
1429 *) Add full support for -rpath/-R, both in shared libraries and
1430 applications, at least on the platforms where it's known how
1434 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
1435 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
1436 will now compute a table of multiples of the generator that
1437 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
1438 faster (notably in the case of a single point multiplication,
1439 scalar * generator).
1440 [Nils Larsch, Bodo Moeller]
1442 *) IPv6 support for certificate extensions. The various extensions
1443 which use the IP:a.b.c.d can now take IPv6 addresses using the
1444 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
1448 *) Added an ENGINE that implements RSA by performing private key
1449 exponentiations with the GMP library. The conversions to and from
1450 GMP's mpz_t format aren't optimised nor are any montgomery forms
1451 cached, and on x86 it appears OpenSSL's own performance has caught up.
1452 However there are likely to be other architectures where GMP could
1453 provide a boost. This ENGINE is not built in by default, but it can be
1454 specified at Configure time and should be accompanied by the necessary
1455 linker additions, eg;
1456 ./config -DOPENSSL_USE_GMP -lgmp
1459 *) "openssl engine" will not display ENGINE/DSO load failure errors when
1460 testing availability of engines with "-t" - the old behaviour is
1461 produced by increasing the feature's verbosity with "-tt".
1464 *) ECDSA routines: under certain error conditions uninitialized BN objects
1465 could be freed. Solution: make sure initialization is performed early
1466 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
1470 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
1471 and DH_METHOD (eg. by ENGINE implementations) to override the normal
1472 software implementations. For DSA and DH, parameter generation can
1473 also be overriden by providing the appropriate method callbacks.
1476 *) Change the "progress" mechanism used in key-generation and
1477 primality testing to functions that take a new BN_GENCB pointer in
1478 place of callback/argument pairs. The new API functions have "_ex"
1479 postfixes and the older functions are reimplemented as wrappers for
1480 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
1481 declarations of the old functions to help (graceful) attempts to
1482 migrate to the new functions. Also, the new key-generation API
1483 functions operate on a caller-supplied key-structure and return
1484 success/failure rather than returning a key or NULL - this is to
1485 help make "keygen" another member function of RSA_METHOD etc.
1487 Example for using the new callback interface:
1489 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
1493 BN_GENCB_set(&my_cb, my_callback, my_arg);
1495 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
1496 /* For the meaning of a, b in calls to my_callback(), see the
1497 * documentation of the function that calls the callback.
1498 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
1499 * my_callback should return 1 if it wants BN_is_prime_ex()
1500 * to continue, or 0 to stop.
1505 *) Change the ZLIB compression method to be stateful, and make it
1506 available to TLS with the number defined in
1507 draft-ietf-tls-compression-04.txt.
1510 *) Add the ASN.1 structures and functions for CertificatePair, which
1511 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
1513 CertificatePair ::= SEQUENCE {
1514 forward [0] Certificate OPTIONAL,
1515 reverse [1] Certificate OPTIONAL,
1516 -- at least one of the pair shall be present -- }
1518 Also implement the PEM functions to read and write certificate
1519 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
1521 This needed to be defined, mostly for the sake of the LDAP
1522 attribute crossCertificatePair, but may prove useful elsewhere as
1526 *) Make it possible to inhibit symlinking of shared libraries in
1527 Makefile.shared, for Cygwin's sake.
1530 *) Extend the BIGNUM API by creating a function
1531 void BN_set_negative(BIGNUM *a, int neg);
1532 and a macro that behave like
1533 int BN_is_negative(const BIGNUM *a);
1535 to avoid the need to access 'a->neg' directly in applications.
1538 *) Implement fast modular reduction for pseudo-Mersenne primes
1539 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
1540 EC_GROUP_new_curve_GFp() will now automatically use this
1542 [Nils Larsch <nla@trustcenter.de>]
1544 *) Add new lock type (CRYPTO_LOCK_BN).
1547 *) Change the ENGINE framework to automatically load engines
1548 dynamically from specific directories unless they could be
1549 found to already be built in or loaded. Move all the
1550 current engines except for the cryptodev one to a new
1552 The engines in engines/ are built as shared libraries if
1553 the "shared" options was given to ./Configure or ./config.
1554 Otherwise, they are inserted in libcrypto.a.
1555 /usr/local/ssl/engines is the default directory for dynamic
1556 engines, but that can be overriden at configure time through
1557 the usual use of --prefix and/or --openssldir, and at run
1558 time with the environment variable OPENSSL_ENGINES.
1559 [Geoff Thorpe and Richard Levitte]
1561 *) Add Makefile.shared, a helper makefile to build shared
1562 libraries. Addapt Makefile.org.
1565 *) Add version info to Win32 DLLs.
1566 [Peter 'Luna' Runestig" <peter@runestig.com>]
1568 *) Add new 'medium level' PKCS#12 API. Certificates and keys
1569 can be added using this API to created arbitrary PKCS#12
1570 files while avoiding the low level API.
1572 New options to PKCS12_create(), key or cert can be NULL and
1573 will then be omitted from the output file. The encryption
1574 algorithm NIDs can be set to -1 for no encryption, the mac
1575 iteration count can be set to 0 to omit the mac.
1577 Enhance pkcs12 utility by making the -nokeys and -nocerts
1578 options work when creating a PKCS#12 file. New option -nomac
1579 to omit the mac, NONE can be set for an encryption algorithm.
1580 New code is modified to use the enhanced PKCS12_create()
1581 instead of the low level API.
1584 *) Extend ASN1 encoder to support indefinite length constructed
1585 encoding. This can output sequences tags and octet strings in
1586 this form. Modify pk7_asn1.c to support indefinite length
1587 encoding. This is experimental and needs additional code to
1588 be useful, such as an ASN1 bio and some enhanced streaming
1591 Extend template encode functionality so that tagging is passed
1592 down to the template encoder.
1595 *) Let 'openssl req' fail if an argument to '-newkey' is not
1596 recognized instead of using RSA as a default.
1599 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
1600 As these are not official, they are not included in "ALL";
1601 the "ECCdraft" ciphersuite group alias can be used to select them.
1602 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
1604 *) Add ECDH engine support.
1605 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
1607 *) Add ECDH in new directory crypto/ecdh/.
1608 [Douglas Stebila (Sun Microsystems Laboratories)]
1610 *) Let BN_rand_range() abort with an error after 100 iterations
1611 without success (which indicates a broken PRNG).
1614 *) Change BN_mod_sqrt() so that it verifies that the input value
1615 is really the square of the return value. (Previously,
1616 BN_mod_sqrt would show GIGO behaviour.)
1619 *) Add named elliptic curves over binary fields from X9.62, SECG,
1620 and WAP/WTLS; add OIDs that were still missing.
1622 [Sheueling Chang Shantz and Douglas Stebila
1623 (Sun Microsystems Laboratories)]
1625 *) Extend the EC library for elliptic curves over binary fields
1626 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
1629 EC_GF2m_simple_method
1633 EC_GROUP_new_curve_GF2m
1634 EC_GROUP_set_curve_GF2m
1635 EC_GROUP_get_curve_GF2m
1636 EC_POINT_set_affine_coordinates_GF2m
1637 EC_POINT_get_affine_coordinates_GF2m
1638 EC_POINT_set_compressed_coordinates_GF2m
1640 Point compression for binary fields is disabled by default for
1641 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
1644 As binary polynomials are represented as BIGNUMs, various members
1645 of the EC_GROUP and EC_POINT data structures can be shared
1646 between the implementations for prime fields and binary fields;
1647 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
1648 are essentially identical to their ..._GFp counterparts.
1649 (For simplicity, the '..._GFp' prefix has been dropped from
1650 various internal method names.)
1652 An internal 'field_div' method (similar to 'field_mul' and
1653 'field_sqr') has been added; this is used only for binary fields.
1655 [Sheueling Chang Shantz and Douglas Stebila
1656 (Sun Microsystems Laboratories)]
1658 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
1659 through methods ('mul', 'precompute_mult').
1661 The generic implementations (now internally called 'ec_wNAF_mul'
1662 and 'ec_wNAF_precomputed_mult') remain the default if these
1663 methods are undefined.
1665 [Sheueling Chang Shantz and Douglas Stebila
1666 (Sun Microsystems Laboratories)]
1668 *) New function EC_GROUP_get_degree, which is defined through
1669 EC_METHOD. For curves over prime fields, this returns the bit
1670 length of the modulus.
1672 [Sheueling Chang Shantz and Douglas Stebila
1673 (Sun Microsystems Laboratories)]
1675 *) New functions EC_GROUP_dup, EC_POINT_dup.
1676 (These simply call ..._new and ..._copy).
1678 [Sheueling Chang Shantz and Douglas Stebila
1679 (Sun Microsystems Laboratories)]
1681 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
1682 Polynomials are represented as BIGNUMs (where the sign bit is not
1683 used) in the following functions [macros]:
1686 BN_GF2m_sub [= BN_GF2m_add]
1687 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
1688 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
1689 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
1691 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
1692 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
1693 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
1694 BN_GF2m_cmp [= BN_ucmp]
1696 (Note that only the 'mod' functions are actually for fields GF(2^m).
1697 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
1699 For some functions, an the irreducible polynomial defining a
1700 field can be given as an 'unsigned int[]' with strictly
1701 decreasing elements giving the indices of those bits that are set;
1702 i.e., p[] represents the polynomial
1703 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
1705 p[0] > p[1] > ... > p[k] = 0.
1706 This applies to the following functions:
1711 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
1712 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
1714 BN_GF2m_mod_sqrt_arr
1715 BN_GF2m_mod_solve_quad_arr
1719 Conversion can be performed by the following functions:
1724 bntest.c has additional tests for binary polynomial arithmetic.
1726 Two implementations for BN_GF2m_mod_div() are available.
1727 The default algorithm simply uses BN_GF2m_mod_inv() and
1728 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
1729 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
1730 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
1732 [Sheueling Chang Shantz and Douglas Stebila
1733 (Sun Microsystems Laboratories)]
1735 *) Add new error code 'ERR_R_DISABLED' that can be used when some
1736 functionality is disabled at compile-time.
1737 [Douglas Stebila <douglas.stebila@sun.com>]
1739 *) Change default behaviour of 'openssl asn1parse' so that more
1740 information is visible when viewing, e.g., a certificate:
1742 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
1743 mode the content of non-printable OCTET STRINGs is output in a
1744 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
1745 avoid the appearance of a printable string.
1746 [Nils Larsch <nla@trustcenter.de>]
1748 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
1750 EC_GROUP_set_asn1_flag()
1751 EC_GROUP_get_asn1_flag()
1752 EC_GROUP_set_point_conversion_form()
1753 EC_GROUP_get_point_conversion_form()
1754 These control ASN1 encoding details:
1755 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
1756 has been set to OPENSSL_EC_NAMED_CURVE.
1757 - Points are encoded in uncompressed form by default; options for
1758 asn1_for are as for point2oct, namely
1759 POINT_CONVERSION_COMPRESSED
1760 POINT_CONVERSION_UNCOMPRESSED
1761 POINT_CONVERSION_HYBRID
1763 Also add 'seed' and 'seed_len' members to EC_GROUP with access
1766 EC_GROUP_get0_seed()
1767 EC_GROUP_get_seed_len()
1768 This is used only for ASN1 purposes (so far).
1769 [Nils Larsch <nla@trustcenter.de>]
1771 *) Add 'field_type' member to EC_METHOD, which holds the NID
1772 of the appropriate field type OID. The new function
1773 EC_METHOD_get_field_type() returns this value.
1774 [Nils Larsch <nla@trustcenter.de>]
1779 EC_POINT_point2hex()
1780 EC_POINT_hex2point()
1781 providing useful interfaces to EC_POINT_point2oct() and
1782 EC_POINT_oct2point().
1783 [Nils Larsch <nla@trustcenter.de>]
1785 *) Change internals of the EC library so that the functions
1786 EC_GROUP_set_generator()
1787 EC_GROUP_get_generator()
1788 EC_GROUP_get_order()
1789 EC_GROUP_get_cofactor()
1790 are implemented directly in crypto/ec/ec_lib.c and not dispatched
1791 to methods, which would lead to unnecessary code duplication when
1792 adding different types of curves.
1793 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
1795 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
1796 arithmetic, and such that modified wNAFs are generated
1797 (which avoid length expansion in many cases).
1800 *) Add a function EC_GROUP_check_discriminant() (defined via
1801 EC_METHOD) that verifies that the curve discriminant is non-zero.
1803 Add a function EC_GROUP_check() that makes some sanity tests
1804 on a EC_GROUP, its generator and order. This includes
1805 EC_GROUP_check_discriminant().
1806 [Nils Larsch <nla@trustcenter.de>]
1808 *) Add ECDSA in new directory crypto/ecdsa/.
1810 Add applications 'openssl ecparam' and 'openssl ecdsa'
1811 (these are based on 'openssl dsaparam' and 'openssl dsa').
1813 ECDSA support is also included in various other files across the
1814 library. Most notably,
1815 - 'openssl req' now has a '-newkey ecdsa:file' option;
1816 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
1817 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
1818 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
1819 them suitable for ECDSA where domain parameters must be
1820 extracted before the specific public key;
1821 - ECDSA engine support has been added.
1822 [Nils Larsch <nla@trustcenter.de>]
1824 *) Include some named elliptic curves, and add OIDs from X9.62,
1825 SECG, and WAP/WTLS. Each curve can be obtained from the new
1827 EC_GROUP_new_by_curve_name(),
1828 and the list of available named curves can be obtained with
1829 EC_get_builtin_curves().
1830 Also add a 'curve_name' member to EC_GROUP objects, which can be
1832 EC_GROUP_set_curve_name()
1833 EC_GROUP_get_curve_name()
1834 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
1836 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1837 was actually never needed) and in BN_mul(). The removal in BN_mul()
1838 required a small change in bn_mul_part_recursive() and the addition
1839 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1840 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1841 bn_sub_words() and bn_add_words() except they take arrays with
1845 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
1847 *) Cleanse PEM buffers before freeing them since they may contain
1849 [Benjamin Bennett <ben@psc.edu>]
1851 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1852 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1853 authentication-only ciphersuites.
1856 *) Since AES128 and AES256 share a single mask bit in the logic of
1857 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1858 kludge to work properly if AES128 is available and AES256 isn't.
1861 *) Expand security boundary to match 1.1.1 module.
1864 *) Remove redundant features: hash file source, editing of test vectors
1865 modify fipsld to use external fips_premain.c signature.
1868 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
1869 run algorithm test programs.
1872 *) Make algorithm test programs more tolerant of whitespace.
1875 *) Have SSL/TLS server implementation tolerate "mismatched" record
1876 protocol version while receiving ClientHello even if the
1877 ClientHello is fragmented. (The server can't insist on the
1878 particular protocol version it has chosen before the ServerHello
1879 message has informed the client about his choice.)
1882 *) Load error codes if they are not already present instead of using a
1883 static variable. This allows them to be cleanly unloaded and reloaded.
1886 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
1888 *) Introduce limits to prevent malicious keys being able to
1889 cause a denial of service. (CVE-2006-2940)
1890 [Steve Henson, Bodo Moeller]
1892 *) Fix ASN.1 parsing of certain invalid structures that can result
1893 in a denial of service. (CVE-2006-2937) [Steve Henson]
1895 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1896 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1898 *) Fix SSL client code which could crash if connecting to a
1899 malicious SSLv2 server. (CVE-2006-4343)
1900 [Tavis Ormandy and Will Drewry, Google Security Team]
1902 *) Change ciphersuite string processing so that an explicit
1903 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
1904 will no longer include "AES128-SHA"), and any other similar
1905 ciphersuite (same bitmap) from *other* protocol versions (so that
1906 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
1907 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
1908 changes from 0.9.8b and 0.9.8d.
1911 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
1913 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1914 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1916 *) Change the Unix randomness entropy gathering to use poll() when
1917 possible instead of select(), since the latter has some
1918 undesirable limitations.
1919 [Darryl Miles via Richard Levitte and Bodo Moeller]
1921 *) Disable rogue ciphersuites:
1923 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1924 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1925 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1927 The latter two were purportedly from
1928 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1931 Also deactive the remaining ciphersuites from
1932 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1933 unofficial, and the ID has long expired.
1936 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1937 dual-core machines) and other potential thread-safety issues.
1940 Changes between 0.9.7i and 0.9.7j [04 May 2006]
1942 *) Adapt fipsld and the build system to link against the validated FIPS
1943 module in FIPS mode.
1946 *) Fixes for VC++ 2005 build under Windows.
1949 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
1950 from a Windows bash shell such as MSYS. It is autodetected from the
1951 "config" script when run from a VC++ environment. Modify standard VC++
1952 build to use fipscanister.o from the GNU make build.
1955 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
1957 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
1958 The value now differs depending on if you build for FIPS or not.
1959 BEWARE! A program linked with a shared FIPSed libcrypto can't be
1960 safely run with a non-FIPSed libcrypto, as it may crash because of
1961 the difference induced by this change.
1964 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
1966 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1967 (part of SSL_OP_ALL). This option used to disable the
1968 countermeasure against man-in-the-middle protocol-version
1969 rollback in the SSL 2.0 server implementation, which is a bad
1970 idea. (CVE-2005-2969)
1972 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1973 for Information Security, National Institute of Advanced Industrial
1974 Science and Technology [AIST], Japan)]
1976 *) Minimal support for X9.31 signatures and PSS padding modes. This is
1977 mainly for FIPS compliance and not fully integrated at this stage.
1980 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
1981 the exponentiation using a fixed-length exponent. (Otherwise,
1982 the information leaked through timing could expose the secret key
1983 after many signatures; cf. Bleichenbacher's attack on DSA with
1987 *) Make a new fixed-window mod_exp implementation the default for
1988 RSA, DSA, and DH private-key operations so that the sequence of
1989 squares and multiplies and the memory access pattern are
1990 independent of the particular secret key. This will mitigate
1991 cache-timing and potential related attacks.
1993 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
1994 and this is automatically used by BN_mod_exp_mont() if the new flag
1995 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
1996 will use this BN flag for private exponents unless the flag
1997 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
1998 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
2000 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
2002 *) Change the client implementation for SSLv23_method() and
2003 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
2004 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
2005 (Previously, the SSL 2.0 backwards compatible Client Hello
2006 message format would be used even with SSL_OP_NO_SSLv2.)
2009 *) Add support for smime-type MIME parameter in S/MIME messages which some
2013 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
2014 a threadsafe manner. Modify rsa code to use new function and add calls
2015 to dsa and dh code (which had race conditions before).
2018 *) Include the fixed error library code in the C error file definitions
2019 instead of fixing them up at runtime. This keeps the error code
2020 structures constant.
2023 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
2025 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
2028 *) Fixes for newer kerberos headers. NB: the casts are needed because
2029 the 'length' field is signed on one version and unsigned on another
2030 with no (?) obvious way to tell the difference, without these VC++
2031 complains. Also the "definition" of FAR (blank) is no longer included
2032 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
2033 some needed definitions.
2036 *) Undo Cygwin change.
2039 *) Added support for proxy certificates according to RFC 3820.
2040 Because they may be a security thread to unaware applications,
2041 they must be explicitely allowed in run-time. See
2042 docs/HOWTO/proxy_certificates.txt for further information.
2045 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
2047 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
2048 server and client random values. Previously
2049 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
2050 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
2052 This change has negligible security impact because:
2054 1. Server and client random values still have 24 bytes of pseudo random
2057 2. Server and client random values are sent in the clear in the initial
2060 3. The master secret is derived using the premaster secret (48 bytes in
2061 size for static RSA ciphersuites) as well as client server and random
2064 The OpenSSL team would like to thank the UK NISCC for bringing this issue
2067 [Stephen Henson, reported by UK NISCC]
2069 *) Use Windows randomness collection on Cygwin.
2072 *) Fix hang in EGD/PRNGD query when communication socket is closed
2073 prematurely by EGD/PRNGD.
2074 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
2076 *) Prompt for pass phrases when appropriate for PKCS12 input format.
2079 *) Back-port of selected performance improvements from development
2080 branch, as well as improved support for PowerPC platforms.
2083 *) Add lots of checks for memory allocation failure, error codes to indicate
2084 failure and freeing up memory if a failure occurs.
2085 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
2087 *) Add new -passin argument to dgst.
2090 *) Perform some character comparisons of different types in X509_NAME_cmp:
2091 this is needed for some certificates that reencode DNs into UTF8Strings
2092 (in violation of RFC3280) and can't or wont issue name rollover
2096 *) Make an explicit check during certificate validation to see that
2097 the CA setting in each certificate on the chain is correct. As a
2098 side effect always do the following basic checks on extensions,
2099 not just when there's an associated purpose to the check:
2101 - if there is an unhandled critical extension (unless the user
2102 has chosen to ignore this fault)
2103 - if the path length has been exceeded (if one is set at all)
2104 - that certain extensions fit the associated purpose (if one has
2108 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
2110 *) Avoid a race condition when CRLs are checked in a multi threaded
2111 environment. This would happen due to the reordering of the revoked
2112 entries during signature checking and serial number lookup. Now the
2113 encoding is cached and the serial number sort performed under a lock.
2114 Add new STACK function sk_is_sorted().
2117 *) Add Delta CRL to the extension code.
2120 *) Various fixes to s3_pkt.c so alerts are sent properly.
2121 [David Holmes <d.holmes@f5.com>]
2123 *) Reduce the chances of duplicate issuer name and serial numbers (in
2124 violation of RFC3280) using the OpenSSL certificate creation utilities.
2125 This is done by creating a random 64 bit value for the initial serial
2126 number when a serial number file is created or when a self signed
2127 certificate is created using 'openssl req -x509'. The initial serial
2128 number file is created using 'openssl x509 -next_serial' in CA.pl
2129 rather than being initialized to 1.
2132 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
2134 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
2135 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
2136 [Joe Orton, Steve Henson]
2138 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
2140 [Joe Orton, Steve Henson]
2142 *) Make it possible to have multiple active certificates with the same
2143 subject in the CA index file. This is done only if the keyword
2144 'unique_subject' is set to 'no' in the main CA section (default
2145 if 'CA_default') of the configuration file. The value is saved
2146 with the database itself in a separate index attribute file,
2147 named like the index file with '.attr' appended to the name.
2150 *) X509 verify fixes. Disable broken certificate workarounds when
2151 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
2152 keyUsage extension present. Don't accept CRLs with unhandled critical
2153 extensions: since verify currently doesn't process CRL extensions this
2154 rejects a CRL with *any* critical extensions. Add new verify error codes
2158 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
2159 A clarification of RFC2560 will require the use of OCTET STRINGs and
2160 some implementations cannot handle the current raw format. Since OpenSSL
2161 copies and compares OCSP nonces as opaque blobs without any attempt at
2162 parsing them this should not create any compatibility issues.
2165 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
2166 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
2167 this HMAC (and other) operations are several times slower than OpenSSL
2171 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
2172 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2174 *) Use the correct content when signing type "other".
2177 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
2179 *) Fix various bugs revealed by running the NISCC test suite:
2181 Stop out of bounds reads in the ASN1 code when presented with
2182 invalid tags (CVE-2003-0543 and CVE-2003-0544).
2184 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
2186 If verify callback ignores invalid public key errors don't try to check
2187 certificate signature with the NULL public key.
2191 *) New -ignore_err option in ocsp application to stop the server
2192 exiting on the first error in a request.
2195 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
2196 if the server requested one: as stated in TLS 1.0 and SSL 3.0
2200 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
2201 extra data after the compression methods not only for TLS 1.0
2202 but also for SSL 3.0 (as required by the specification).
2203 [Bodo Moeller; problem pointed out by Matthias Loepfe]
2205 *) Change X509_certificate_type() to mark the key as exported/exportable
2206 when it's 512 *bits* long, not 512 bytes.
2209 *) Change AES_cbc_encrypt() so it outputs exact multiple of
2210 blocks during encryption.
2213 *) Various fixes to base64 BIO and non blocking I/O. On write
2214 flushes were not handled properly if the BIO retried. On read
2215 data was not being buffered properly and had various logic bugs.
2216 This also affects blocking I/O when the data being decoded is a
2220 *) Various S/MIME bugfixes and compatibility changes:
2221 output correct application/pkcs7 MIME type if
2222 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
2223 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
2224 of files as .eml work). Correctly handle very long lines in MIME
2228 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
2230 *) Countermeasure against the Klima-Pokorny-Rosa extension of
2231 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
2232 a protocol version number mismatch like a decryption error
2233 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
2236 *) Turn on RSA blinding by default in the default implementation
2237 to avoid a timing attack. Applications that don't want it can call
2238 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
2239 They would be ill-advised to do so in most cases.
2240 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
2242 *) Change RSA blinding code so that it works when the PRNG is not
2243 seeded (in this case, the secret RSA exponent is abused as
2244 an unpredictable seed -- if it is not unpredictable, there
2245 is no point in blinding anyway). Make RSA blinding thread-safe
2246 by remembering the creator's thread ID in rsa->blinding and
2247 having all other threads use local one-time blinding factors
2248 (this requires more computation than sharing rsa->blinding, but
2249 avoids excessive locking; and if an RSA object is not shared
2250 between threads, blinding will still be very fast).
2253 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
2254 ENGINE as defaults for all supported algorithms irrespective of
2255 the 'flags' parameter. 'flags' is now honoured, so applications
2256 should make sure they are passing it correctly.
2259 *) Target "mingw" now allows native Windows code to be generated in
2260 the Cygwin environment as well as with the MinGW compiler.
2263 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
2265 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
2266 via timing by performing a MAC computation even if incorrrect
2267 block cipher padding has been found. This is a countermeasure
2268 against active attacks where the attacker has to distinguish
2269 between bad padding and a MAC verification error. (CVE-2003-0078)
2271 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
2272 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
2273 Martin Vuagnoux (EPFL, Ilion)]
2275 *) Make the no-err option work as intended. The intention with no-err
2276 is not to have the whole error stack handling routines removed from
2277 libcrypto, it's only intended to remove all the function name and
2278 reason texts, thereby removing some of the footprint that may not
2279 be interesting if those errors aren't displayed anyway.
2281 NOTE: it's still possible for any application or module to have it's
2282 own set of error texts inserted. The routines are there, just not
2283 used by default when no-err is given.
2286 *) Add support for FreeBSD on IA64.
2287 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
2289 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
2290 Kerberos function mit_des_cbc_cksum(). Before this change,
2291 the value returned by DES_cbc_cksum() was like the one from
2292 mit_des_cbc_cksum(), except the bytes were swapped.
2293 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
2295 *) Allow an application to disable the automatic SSL chain building.
2296 Before this a rather primitive chain build was always performed in
2297 ssl3_output_cert_chain(): an application had no way to send the
2298 correct chain if the automatic operation produced an incorrect result.
2300 Now the chain builder is disabled if either:
2302 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
2304 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
2306 The reasoning behind this is that an application would not want the
2307 auto chain building to take place if extra chain certificates are
2308 present and it might also want a means of sending no additional
2309 certificates (for example the chain has two certificates and the
2313 *) Add the possibility to build without the ENGINE framework.
2314 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2316 *) Under Win32 gmtime() can return NULL: check return value in
2317 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
2320 *) DSA routines: under certain error conditions uninitialized BN objects
2321 could be freed. Solution: make sure initialization is performed early
2322 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
2323 Nils Larsch <nla@trustcenter.de> via PR#459)
2326 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
2327 checked on reconnect on the client side, therefore session resumption
2328 could still fail with a "ssl session id is different" error. This
2329 behaviour is masked when SSL_OP_ALL is used due to
2330 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
2331 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2332 followup to PR #377.
2335 *) IA-32 assembler support enhancements: unified ELF targets, support
2336 for SCO/Caldera platforms, fix for Cygwin shared build.
2339 *) Add support for FreeBSD on sparc64. As a consequence, support for
2340 FreeBSD on non-x86 processors is separate from x86 processors on
2341 the config script, much like the NetBSD support.
2342 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
2344 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
2346 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
2349 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
2350 code (06) was taken as the first octet of the session ID and the last
2351 octet was ignored consequently. As a result SSLv2 client side session
2352 caching could not have worked due to the session ID mismatch between
2354 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2358 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
2359 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
2363 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
2364 seems that in spite of existing for more than a year, many application
2365 author have done nothing to provide the necessary callbacks, which
2366 means that this particular engine will not work properly anywhere.
2367 This is a very unfortunate situation which forces us, in the name
2368 of usability, to give the hw_ncipher.c a static lock, which is part
2370 NOTE: This is for the 0.9.7 series ONLY. This hack will never
2371 appear in 0.9.8 or later. We EXPECT application authors to have
2372 dealt properly with this when 0.9.8 is released (unless we actually
2373 make such changes in the libcrypto locking code that changes will
2374 have to be made anyway).
2377 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
2378 octets have been read, EOF or an error occurs. Without this change
2379 some truncated ASN1 structures will not produce an error.
2382 *) Disable Heimdal support, since it hasn't been fully implemented.
2383 Still give the possibility to force the use of Heimdal, but with
2384 warnings and a request that patches get sent to openssl-dev.
2387 *) Add the VC-CE target, introduce the WINCE sysname, and add
2388 INSTALL.WCE and appropriate conditionals to make it build.
2389 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2391 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
2392 cygssl-x.y.z.dll, where x, y and z are the major, minor and
2393 edit numbers of the version.
2394 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2396 *) Introduce safe string copy and catenation functions
2397 (BUF_strlcpy() and BUF_strlcat()).
2398 [Ben Laurie (CHATS) and Richard Levitte]
2400 *) Avoid using fixed-size buffers for one-line DNs.
2401 [Ben Laurie (CHATS)]
2403 *) Add BUF_MEM_grow_clean() to avoid information leakage when
2404 resizing buffers containing secrets, and use where appropriate.
2405 [Ben Laurie (CHATS)]
2407 *) Avoid using fixed size buffers for configuration file location.
2408 [Ben Laurie (CHATS)]
2410 *) Avoid filename truncation for various CA files.
2411 [Ben Laurie (CHATS)]
2413 *) Use sizeof in preference to magic numbers.
2414 [Ben Laurie (CHATS)]
2416 *) Avoid filename truncation in cert requests.
2417 [Ben Laurie (CHATS)]
2419 *) Add assertions to check for (supposedly impossible) buffer
2421 [Ben Laurie (CHATS)]
2423 *) Don't cache truncated DNS entries in the local cache (this could
2424 potentially lead to a spoofing attack).
2425 [Ben Laurie (CHATS)]
2427 *) Fix various buffers to be large enough for hex/decimal
2428 representations in a platform independent manner.
2429 [Ben Laurie (CHATS)]
2431 *) Add CRYPTO_realloc_clean() to avoid information leakage when
2432 resizing buffers containing secrets, and use where appropriate.
2433 [Ben Laurie (CHATS)]
2435 *) Add BIO_indent() to avoid much slightly worrying code to do
2437 [Ben Laurie (CHATS)]
2439 *) Convert sprintf()/BIO_puts() to BIO_printf().
2440 [Ben Laurie (CHATS)]
2442 *) buffer_gets() could terminate with the buffer only half
2444 [Ben Laurie (CHATS)]
2446 *) Add assertions to prevent user-supplied crypto functions from
2447 overflowing internal buffers by having large block sizes, etc.
2448 [Ben Laurie (CHATS)]
2450 *) New OPENSSL_assert() macro (similar to assert(), but enabled
2452 [Ben Laurie (CHATS)]
2454 *) Eliminate unused copy of key in RC4.
2455 [Ben Laurie (CHATS)]
2457 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
2458 [Ben Laurie (CHATS)]
2460 *) Fix off-by-one error in EGD path.
2461 [Ben Laurie (CHATS)]
2463 *) If RANDFILE path is too long, ignore instead of truncating.
2464 [Ben Laurie (CHATS)]
2466 *) Eliminate unused and incorrectly sized X.509 structure
2468 [Ben Laurie (CHATS)]
2470 *) Eliminate unused and dangerous function knumber().
2471 [Ben Laurie (CHATS)]
2473 *) Eliminate unused and dangerous structure, KSSL_ERR.
2474 [Ben Laurie (CHATS)]
2476 *) Protect against overlong session ID context length in an encoded
2477 session object. Since these are local, this does not appear to be
2479 [Ben Laurie (CHATS)]
2481 *) Change from security patch (see 0.9.6e below) that did not affect
2482 the 0.9.6 release series:
2484 Remote buffer overflow in SSL3 protocol - an attacker could
2485 supply an oversized master key in Kerberos-enabled versions.
2487 [Ben Laurie (CHATS)]
2489 *) Change the SSL kerb5 codes to match RFC 2712.
2492 *) Make -nameopt work fully for req and add -reqopt switch.
2493 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
2495 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
2496 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
2498 *) Make sure tests can be performed even if the corresponding algorithms
2499 have been removed entirely. This was also the last step to make
2500 OpenSSL compilable with DJGPP under all reasonable conditions.
2501 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
2503 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
2504 to allow version independent disabling of normally unselected ciphers,
2505 which may be activated as a side-effect of selecting a single cipher.
2507 (E.g., cipher list string "RSA" enables ciphersuites that are left
2508 out of "ALL" because they do not provide symmetric encryption.
2509 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
2510 [Lutz Jaenicke, Bodo Moeller]
2512 *) Add appropriate support for separate platform-dependent build
2513 directories. The recommended way to make a platform-dependent
2514 build directory is the following (tested on Linux), maybe with
2517 # Place yourself outside of the OpenSSL source tree. In
2518 # this example, the environment variable OPENSSL_SOURCE
2519 # is assumed to contain the absolute OpenSSL source directory.
2520 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
2521 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
2522 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
2523 mkdir -p `dirname $F`
2524 ln -s $OPENSSL_SOURCE/$F $F
2527 To be absolutely sure not to disturb the source tree, a "make clean"
2528 is a good thing. If it isn't successfull, don't worry about it,
2529 it probably means the source directory is very clean.
2532 *) Make sure any ENGINE control commands make local copies of string
2533 pointers passed to them whenever necessary. Otherwise it is possible
2534 the caller may have overwritten (or deallocated) the original string
2535 data when a later ENGINE operation tries to use the stored values.
2536 [Götz Babin-Ebell <babinebell@trustcenter.de>]
2538 *) Improve diagnostics in file reading and command-line digests.
2539 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
2541 *) Add AES modes CFB and OFB to the object database. Correct an
2542 error in AES-CFB decryption.
2545 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
2546 allows existing EVP_CIPHER_CTX structures to be reused after
2547 calling EVP_*Final(). This behaviour is used by encryption
2548 BIOs and some applications. This has the side effect that
2549 applications must explicitly clean up cipher contexts with
2550 EVP_CIPHER_CTX_cleanup() or they will leak memory.
2553 *) Check the values of dna and dnb in bn_mul_recursive before calling
2554 bn_mul_comba (a non zero value means the a or b arrays do not contain
2555 n2 elements) and fallback to bn_mul_normal if either is not zero.
2558 *) Fix escaping of non-ASCII characters when using the -subj option
2559 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
2562 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
2563 form for "surname", serialNumber has no short form.
2564 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
2565 therefore remove "mail" short name for "internet 7".
2566 The OID for unique identifiers in X509 certificates is
2567 x500UniqueIdentifier, not uniqueIdentifier.
2568 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
2571 *) Add an "init" command to the ENGINE config module and auto initialize
2572 ENGINEs. Without any "init" command the ENGINE will be initialized
2573 after all ctrl commands have been executed on it. If init=1 the
2574 ENGINE is initailized at that point (ctrls before that point are run
2575 on the uninitialized ENGINE and after on the initialized one). If
2576 init=0 then the ENGINE will not be iniatialized at all.
2579 *) Fix the 'app_verify_callback' interface so that the user-defined
2580 argument is actually passed to the callback: In the
2581 SSL_CTX_set_cert_verify_callback() prototype, the callback
2582 declaration has been changed from
2585 int (*cb)(X509_STORE_CTX *,void *);
2586 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
2587 i=s->ctx->app_verify_callback(&ctx)
2588 has been changed into
2589 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
2591 To update applications using SSL_CTX_set_cert_verify_callback(),
2592 a dummy argument can be added to their callback functions.
2593 [D. K. Smetters <smetters@parc.xerox.com>]
2595 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
2596 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
2598 *) Add and OPENSSL_LOAD_CONF define which will cause
2599 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
2600 This allows older applications to transparently support certain
2601 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
2602 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
2603 load the config file and OPENSSL_add_all_algorithms_conf() which will
2604 always load it have also been added.
2607 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
2608 Adjust NIDs and EVP layer.
2609 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2611 *) Config modules support in openssl utility.
2613 Most commands now load modules from the config file,
2614 though in a few (such as version) this isn't done
2615 because it couldn't be used for anything.
2617 In the case of ca and req the config file used is
2618 the same as the utility itself: that is the -config
2619 command line option can be used to specify an
2623 *) Move default behaviour from OPENSSL_config(). If appname is NULL
2624 use "openssl_conf" if filename is NULL use default openssl config file.
2627 *) Add an argument to OPENSSL_config() to allow the use of an alternative
2628 config section name. Add a new flag to tolerate a missing config file
2629 and move code to CONF_modules_load_file().
2632 *) Support for crypto accelerator cards from Accelerated Encryption
2633 Processing, www.aep.ie. (Use engine 'aep')
2634 The support was copied from 0.9.6c [engine] and adapted/corrected
2635 to work with the new engine framework.
2636 [AEP Inc. and Richard Levitte]
2638 *) Support for SureWare crypto accelerator cards from Baltimore
2639 Technologies. (Use engine 'sureware')
2640 The support was copied from 0.9.6c [engine] and adapted
2641 to work with the new engine framework.
2644 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
2645 make the newer ENGINE framework commands for the CHIL engine work.
2646 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
2648 *) Make it possible to produce shared libraries on ReliantUNIX.
2649 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
2651 *) Add the configuration target debug-linux-ppro.
2652 Make 'openssl rsa' use the general key loading routines
2653 implemented in apps.c, and make those routines able to
2654 handle the key format FORMAT_NETSCAPE and the variant
2656 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2658 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2659 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2661 *) Add -keyform to rsautl, and document -engine.
2662 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
2664 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
2665 BIO_R_NO_SUCH_FILE error code rather than the generic
2666 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
2669 *) Add new functions
2671 ERR_peek_last_error_line
2672 ERR_peek_last_error_line_data.
2673 These are similar to
2676 ERR_peek_error_line_data,
2677 but report on the latest error recorded rather than the first one
2678 still in the error queue.
2679 [Ben Laurie, Bodo Moeller]
2681 *) default_algorithms option in ENGINE config module. This allows things
2683 default_algorithms = ALL
2684 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
2687 *) Prelminary ENGINE config module.
2690 *) New experimental application configuration code.
2693 *) Change the AES code to follow the same name structure as all other
2694 symmetric ciphers, and behave the same way. Move everything to
2695 the directory crypto/aes, thereby obsoleting crypto/rijndael.
2696 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2698 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
2699 [Ben Laurie and Theo de Raadt]
2701 *) Add option to output public keys in req command.
2702 [Massimiliano Pala madwolf@openca.org]
2704 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
2705 (up to about 10% better than before for P-192 and P-224).
2708 *) New functions/macros
2710 SSL_CTX_set_msg_callback(ctx, cb)
2711 SSL_CTX_set_msg_callback_arg(ctx, arg)
2712 SSL_set_msg_callback(ssl, cb)
2713 SSL_set_msg_callback_arg(ssl, arg)
2715 to request calling a callback function
2717 void cb(int write_p, int version, int content_type,
2718 const void *buf, size_t len, SSL *ssl, void *arg)
2720 whenever a protocol message has been completely received
2721 (write_p == 0) or sent (write_p == 1). Here 'version' is the
2722 protocol version according to which the SSL library interprets
2723 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
2724 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
2725 the content type as defined in the SSL 3.0/TLS 1.0 protocol
2726 specification (change_cipher_spec(20), alert(21), handshake(22)).
2727 'buf' and 'len' point to the actual message, 'ssl' to the
2728 SSL object, and 'arg' is the application-defined value set by
2729 SSL[_CTX]_set_msg_callback_arg().
2731 'openssl s_client' and 'openssl s_server' have new '-msg' options
2732 to enable a callback that displays all protocol messages.
2735 *) Change the shared library support so shared libraries are built as
2736 soon as the corresponding static library is finished, and thereby get
2737 openssl and the test programs linked against the shared library.
2738 This still only happens when the keyword "shard" has been given to
2739 the configuration scripts.
2741 NOTE: shared library support is still an experimental thing, and
2742 backward binary compatibility is still not guaranteed.
2743 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
2745 *) Add support for Subject Information Access extension.
2746 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2748 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
2749 additional bytes when new memory had to be allocated, not just
2750 when reusing an existing buffer.
2753 *) New command line and configuration option 'utf8' for the req command.
2754 This allows field values to be specified as UTF8 strings.
2757 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
2758 runs for the former and machine-readable output for the latter.
2761 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
2762 of the e-mail address in the DN (i.e., it will go into a certificate
2763 extension only). The new configuration file option 'email_in_dn = no'
2764 has the same effect.
2765 [Massimiliano Pala madwolf@openca.org]
2767 *) Change all functions with names starting with des_ to be starting
2768 with DES_ instead. Add wrappers that are compatible with libdes,
2769 but are named _ossl_old_des_*. Finally, add macros that map the
2770 des_* symbols to the corresponding _ossl_old_des_* if libdes
2771 compatibility is desired. If OpenSSL 0.9.6c compatibility is
2772 desired, the des_* symbols will be mapped to DES_*, with one
2775 Since we provide two compatibility mappings, the user needs to
2776 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
2777 compatibility is desired. The default (i.e., when that macro
2778 isn't defined) is OpenSSL 0.9.6c compatibility.
2780 There are also macros that enable and disable the support of old
2781 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
2782 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
2783 are defined, the default will apply: to support the old des routines.
2785 In either case, one must include openssl/des.h to get the correct
2786 definitions. Do not try to just include openssl/des_old.h, that
2789 NOTE: This is a major break of an old API into a new one. Software
2790 authors are encouraged to switch to the DES_ style functions. Some
2791 time in the future, des_old.h and the libdes compatibility functions
2792 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
2793 default), and then completely removed.
2796 *) Test for certificates which contain unsupported critical extensions.
2797 If such a certificate is found during a verify operation it is
2798 rejected by default: this behaviour can be overridden by either
2799 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
2800 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
2801 X509_supported_extension() has also been added which returns 1 if a
2802 particular extension is supported.
2805 *) Modify the behaviour of EVP cipher functions in similar way to digests
2806 to retain compatibility with existing code.
2809 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
2810 compatibility with existing code. In particular the 'ctx' parameter does
2811 not have to be to be initialized before the call to EVP_DigestInit() and
2812 it is tidied up after a call to EVP_DigestFinal(). New function
2813 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
2814 EVP_MD_CTX_copy() changed to not require the destination to be
2815 initialized valid and new function EVP_MD_CTX_copy_ex() added which
2816 requires the destination to be valid.
2818 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
2819 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
2822 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
2823 so that complete 'Handshake' protocol structures are kept in memory
2824 instead of overwriting 'msg_type' and 'length' with 'body' data.
2827 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
2828 [Massimo Santin via Richard Levitte]
2830 *) Major restructuring to the underlying ENGINE code. This includes
2831 reduction of linker bloat, separation of pure "ENGINE" manipulation
2832 (initialisation, etc) from functionality dealing with implementations
2833 of specific crypto iterfaces. This change also introduces integrated
2834 support for symmetric ciphers and digest implementations - so ENGINEs
2835 can now accelerate these by providing EVP_CIPHER and EVP_MD
2836 implementations of their own. This is detailed in crypto/engine/README
2837 as it couldn't be adequately described here. However, there are a few
2838 API changes worth noting - some RSA, DSA, DH, and RAND functions that
2839 were changed in the original introduction of ENGINE code have now
2840 reverted back - the hooking from this code to ENGINE is now a good
2841 deal more passive and at run-time, operations deal directly with
2842 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
2843 dereferencing through an ENGINE pointer any more. Also, the ENGINE
2844 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
2845 they were not being used by the framework as there is no concept of a
2846 BIGNUM_METHOD and they could not be generalised to the new
2847 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
2848 ENGINE_cpy() has been removed as it cannot be consistently defined in
2852 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
2855 *) Change mkdef.pl to sort symbols that get the same entry number,
2856 and make sure the automatically generated functions ERR_load_*
2857 become part of libeay.num as well.
2860 *) New function SSL_renegotiate_pending(). This returns true once
2861 renegotiation has been requested (either SSL_renegotiate() call
2862 or HelloRequest/ClientHello receveived from the peer) and becomes
2863 false once a handshake has been completed.
2864 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
2865 sends a HelloRequest, but does not ensure that a handshake takes
2866 place. SSL_renegotiate_pending() is useful for checking if the
2867 client has followed the request.)
2870 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
2871 By default, clients may request session resumption even during
2872 renegotiation (if session ID contexts permit); with this option,
2873 session resumption is possible only in the first handshake.
2875 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
2876 more bits available for options that should not be part of
2877 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
2880 *) Add some demos for certificate and certificate request creation.
2883 *) Make maximum certificate chain size accepted from the peer application
2884 settable (SSL*_get/set_max_cert_list()), as proposed by
2885 "Douglas E. Engert" <deengert@anl.gov>.
2888 *) Add support for shared libraries for Unixware-7
2889 (Boyd Lynn Gerber <gerberb@zenez.com>).
2892 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
2893 be done prior to destruction. Use this to unload error strings from
2894 ENGINEs that load their own error strings. NB: This adds two new API
2895 functions to "get" and "set" this destroy handler in an ENGINE.
2898 *) Alter all existing ENGINE implementations (except "openssl" and
2899 "openbsd") to dynamically instantiate their own error strings. This
2900 makes them more flexible to be built both as statically-linked ENGINEs
2901 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
2902 Also, add stub code to each that makes building them as self-contained
2903 shared-libraries easier (see README.ENGINE).
2906 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
2907 implementations into applications that are completely implemented in
2908 self-contained shared-libraries. The "dynamic" ENGINE exposes control
2909 commands that can be used to configure what shared-library to load and
2910 to control aspects of the way it is handled. Also, made an update to
2911 the README.ENGINE file that brings its information up-to-date and
2912 provides some information and instructions on the "dynamic" ENGINE
2913 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
2916 *) Make it possible to unload ranges of ERR strings with a new
2917 "ERR_unload_strings" function.
2920 *) Add a copy() function to EVP_MD.
2923 *) Make EVP_MD routines take a context pointer instead of just the
2924 md_data void pointer.
2927 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
2928 that the digest can only process a single chunk of data
2929 (typically because it is provided by a piece of
2930 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
2931 is only going to provide a single chunk of data, and hence the
2932 framework needn't accumulate the data for oneshot drivers.
2935 *) As with "ERR", make it possible to replace the underlying "ex_data"
2936 functions. This change also alters the storage and management of global
2937 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
2938 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
2939 index counters. The API functions that use this state have been changed
2940 to take a "class_index" rather than pointers to the class's local STACK
2941 and counter, and there is now an API function to dynamically create new
2942 classes. This centralisation allows us to (a) plug a lot of the
2943 thread-safety problems that existed, and (b) makes it possible to clean
2944 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
2945 such data would previously have always leaked in application code and
2946 workarounds were in place to make the memory debugging turn a blind eye
2947 to it. Application code that doesn't use this new function will still
2948 leak as before, but their memory debugging output will announce it now
2949 rather than letting it slide.
2951 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
2952 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
2953 has a return value to indicate success or failure.
2956 *) Make it possible to replace the underlying "ERR" functions such that the
2957 global state (2 LHASH tables and 2 locks) is only used by the "default"
2958 implementation. This change also adds two functions to "get" and "set"
2959 the implementation prior to it being automatically set the first time
2960 any other ERR function takes place. Ie. an application can call "get",
2961 pass the return value to a module it has just loaded, and that module
2962 can call its own "set" function using that value. This means the
2963 module's "ERR" operations will use (and modify) the error state in the
2964 application and not in its own statically linked copy of OpenSSL code.
2967 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
2968 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
2969 the operation, and provides a more encapsulated way for external code
2970 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
2971 to use these functions rather than manually incrementing the counts.
2973 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
2976 *) Add EVP test program.
2979 *) Add symmetric cipher support to ENGINE. Expect the API to change!
2982 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
2983 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
2984 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
2985 These allow a CRL to be built without having to access X509_CRL fields
2986 directly. Modify 'ca' application to use new functions.
2989 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
2990 bug workarounds. Rollback attack detection is a security feature.
2991 The problem will only arise on OpenSSL servers when TLSv1 is not
2992 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
2993 Software authors not wanting to support TLSv1 will have special reasons
2994 for their choice and can explicitly enable this option.
2995 [Bodo Moeller, Lutz Jaenicke]
2997 *) Rationalise EVP so it can be extended: don't include a union of
2998 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
2999 (similar to those existing for EVP_CIPHER_CTX).
3004 EVP_MD_CTX_init(&md); /* new function call */
3005 EVP_DigestInit(&md, EVP_sha1());
3006 EVP_DigestUpdate(&md, in, len);
3007 EVP_DigestFinal(&md, out, NULL);
3008 EVP_MD_CTX_cleanup(&md); /* new function call */
3012 *) Make DES key schedule conform to the usual scheme, as well as
3013 correcting its structure. This means that calls to DES functions
3014 now have to pass a pointer to a des_key_schedule instead of a
3015 plain des_key_schedule (which was actually always a pointer
3018 des_key_schedule ks;
3020 des_set_key_checked(..., &ks);
3021 des_ncbc_encrypt(..., &ks, ...);
3023 (Note that a later change renames 'des_...' into 'DES_...'.)
3026 *) Initial reduction of linker bloat: the use of some functions, such as
3027 PEM causes large amounts of unused functions to be linked in due to
3028 poor organisation. For example pem_all.c contains every PEM function
3029 which has a knock on effect of linking in large amounts of (unused)
3030 ASN1 code. Grouping together similar functions and splitting unrelated
3031 functions prevents this.
3034 *) Cleanup of EVP macros.
3037 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
3038 correct _ecb suffix.
3041 *) Add initial OCSP responder support to ocsp application. The
3042 revocation information is handled using the text based index
3043 use by the ca application. The responder can either handle
3044 requests generated internally, supplied in files (for example
3045 via a CGI script) or using an internal minimal server.
3048 *) Add configuration choices to get zlib compression for TLS.
3051 *) Changes to Kerberos SSL for RFC 2712 compliance:
3052 1. Implemented real KerberosWrapper, instead of just using
3053 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
3054 2. Implemented optional authenticator field of KerberosWrapper.
3056 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
3057 and authenticator structs; see crypto/krb5/.
3059 Generalized Kerberos calls to support multiple Kerberos libraries.
3060 [Vern Staats <staatsvr@asc.hpc.mil>,
3061 Jeffrey Altman <jaltman@columbia.edu>
3062 via Richard Levitte]
3064 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
3065 already does with RSA. testdsa.h now has 'priv_key/pub_key'
3066 values for each of the key sizes rather than having just
3067 parameters (and 'speed' generating keys each time).
3070 *) Speed up EVP routines.
3073 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
3074 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
3075 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
3076 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
3078 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
3079 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
3080 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
3083 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
3085 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
3088 *) Added the OS2-EMX target.
3089 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
3091 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
3092 to support NCONF routines in extension code. New function CONF_set_nconf()
3093 to allow functions which take an NCONF to also handle the old LHASH
3094 structure: this means that the old CONF compatible routines can be
3095 retained (in particular wrt extensions) without having to duplicate the
3096 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
3099 *) Enhance the general user interface with mechanisms for inner control
3100 and with possibilities to have yes/no kind of prompts.
3103 *) Change all calls to low level digest routines in the library and
3104 applications to use EVP. Add missing calls to HMAC_cleanup() and
3105 don't assume HMAC_CTX can be copied using memcpy().
3106 [Verdon Walker <VWalker@novell.com>, Steve Henson]
3108 *) Add the possibility to control engines through control names but with
3109 arbitrary arguments instead of just a string.
3110 Change the key loaders to take a UI_METHOD instead of a callback
3111 function pointer. NOTE: this breaks binary compatibility with earlier
3112 versions of OpenSSL [engine].
3113 Adapt the nCipher code for these new conditions and add a card insertion
3117 *) Enhance the general user interface with mechanisms to better support
3118 dialog box interfaces, application-defined prompts, the possibility
3119 to use defaults (for example default passwords from somewhere else)
3120 and interrupts/cancellations.
3123 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
3124 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
3127 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
3128 tidy up some unnecessarily weird code in 'sk_new()').
3129 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
3131 *) Change the key loading routines for ENGINEs to use the same kind
3132 callback (pem_password_cb) as all other routines that need this
3136 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
3137 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
3138 than this minimum value is recommended.
3141 *) New random seeder for OpenVMS, using the system process statistics
3142 that are easily reachable.
3145 *) Windows apparently can't transparently handle global
3146 variables defined in DLLs. Initialisations such as:
3148 const ASN1_ITEM *it = &ASN1_INTEGER_it;
3150 wont compile. This is used by the any applications that need to
3151 declare their own ASN1 modules. This was fixed by adding the option
3152 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
3153 needed for static libraries under Win32.
3156 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
3157 setting of purpose and trust fields. New X509_STORE trust and
3158 purpose functions and tidy up setting in other SSL functions.
3161 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
3162 structure. These are inherited by X509_STORE_CTX when it is
3163 initialised. This allows various defaults to be set in the
3164 X509_STORE structure (such as flags for CRL checking and custom
3165 purpose or trust settings) for functions which only use X509_STORE_CTX
3166 internally such as S/MIME.
3168 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
3169 trust settings if they are not set in X509_STORE. This allows X509_STORE
3170 purposes and trust (in S/MIME for example) to override any set by default.
3172 Add command line options for CRL checking to smime, s_client and s_server
3176 *) Initial CRL based revocation checking. If the CRL checking flag(s)
3177 are set then the CRL is looked up in the X509_STORE structure and
3178 its validity and signature checked, then if the certificate is found
3179 in the CRL the verify fails with a revoked error.
3181 Various new CRL related callbacks added to X509_STORE_CTX structure.
3183 Command line options added to 'verify' application to support this.
3185 This needs some additional work, such as being able to handle multiple
3186 CRLs with different times, extension based lookup (rather than just
3187 by subject name) and ultimately more complete V2 CRL extension
3191 *) Add a general user interface API (crypto/ui/). This is designed
3192 to replace things like des_read_password and friends (backward
3193 compatibility functions using this new API are provided).
3194 The purpose is to remove prompting functions from the DES code
3195 section as well as provide for prompting through dialog boxes in
3196 a window system and the like.
3199 *) Add "ex_data" support to ENGINE so implementations can add state at a
3200 per-structure level rather than having to store it globally.
3203 *) Make it possible for ENGINE structures to be copied when retrieved by
3204 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
3205 This causes the "original" ENGINE structure to act like a template,
3206 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
3207 operational state can be localised to each ENGINE structure, despite the
3208 fact they all share the same "methods". New ENGINE structures returned in
3209 this case have no functional references and the return value is the single
3210 structural reference. This matches the single structural reference returned
3211 by ENGINE_by_id() normally, when it is incremented on the pre-existing
3215 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
3216 needs to match any other type at all we need to manually clear the
3220 *) Changes to the "openssl engine" utility to include;
3221 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
3222 about an ENGINE's available control commands.
3223 - executing control commands from command line arguments using the
3224 '-pre' and '-post' switches. '-post' is only used if '-t' is
3225 specified and the ENGINE is successfully initialised. The syntax for
3226 the individual commands are colon-separated, for example;
3227 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
3230 *) New dynamic control command support for ENGINEs. ENGINEs can now
3231 declare their own commands (numbers), names (strings), descriptions,
3232 and input types for run-time discovery by calling applications. A
3233 subset of these commands are implicitly classed as "executable"
3234 depending on their input type, and only these can be invoked through
3235 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
3236 can be based on user input, config files, etc). The distinction is
3237 that "executable" commands cannot return anything other than a boolean
3238 result and can only support numeric or string input, whereas some
3239 discoverable commands may only be for direct use through
3240 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
3241 pointers, or other custom uses. The "executable" commands are to
3242 support parameterisations of ENGINE behaviour that can be
3243 unambiguously defined by ENGINEs and used consistently across any
3244 OpenSSL-based application. Commands have been added to all the
3245 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
3246 control over shared-library paths without source code alterations.
3249 *) Changed all ENGINE implementations to dynamically allocate their
3250 ENGINEs rather than declaring them statically. Apart from this being
3251 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
3252 this also allows the implementations to compile without using the
3253 internal engine_int.h header.
3256 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
3257 'const' value. Any code that should be able to modify a RAND_METHOD
3258 should already have non-const pointers to it (ie. they should only
3259 modify their own ones).
3262 *) Made a variety of little tweaks to the ENGINE code.
3263 - "atalla" and "ubsec" string definitions were moved from header files
3264 to C code. "nuron" string definitions were placed in variables
3265 rather than hard-coded - allowing parameterisation of these values
3266 later on via ctrl() commands.
3267 - Removed unused "#if 0"'d code.
3268 - Fixed engine list iteration code so it uses ENGINE_free() to release
3269 structural references.
3270 - Constified the RAND_METHOD element of ENGINE structures.
3271 - Constified various get/set functions as appropriate and added
3272 missing functions (including a catch-all ENGINE_cpy that duplicates
3273 all ENGINE values onto a new ENGINE except reference counts/state).
3274 - Removed NULL parameter checks in get/set functions. Setting a method
3275 or function to NULL is a way of cancelling out a previously set
3276 value. Passing a NULL ENGINE parameter is just plain stupid anyway
3277 and doesn't justify the extra error symbols and code.
3278 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
3279 flags from engine_int.h to engine.h.
3280 - Changed prototypes for ENGINE handler functions (init(), finish(),
3281 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
3284 *) Implement binary inversion algorithm for BN_mod_inverse in addition
3285 to the algorithm using long division. The binary algorithm can be
3286 used only if the modulus is odd. On 32-bit systems, it is faster
3287 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
3288 roughly 5-15% for 256-bit moduli), so we use it only for moduli
3289 up to 450 bits. In 64-bit environments, the binary algorithm
3290 appears to be advantageous for much longer moduli; here we use it
3291 for moduli up to 2048 bits.
3294 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
3295 could not support the combine flag in choice fields.
3298 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
3299 extensions from a certificate request to the certificate.
3302 *) Allow multiple 'certopt' and 'nameopt' options to be separated
3303 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
3304 file: this allows the display of the certificate about to be
3305 signed to be customised, to allow certain fields to be included
3306 or excluded and extension details. The old system didn't display
3307 multicharacter strings properly, omitted fields not in the policy
3308 and couldn't display additional details such as extensions.
3311 *) Function EC_POINTs_mul for multiple scalar multiplication
3312 of an arbitrary number of elliptic curve points
3313 \sum scalars[i]*points[i],
3314 optionally including the generator defined for the EC_GROUP:
3315 scalar*generator + \sum scalars[i]*points[i].
3317 EC_POINT_mul is a simple wrapper function for the typical case
3318 that the point list has just one item (besides the optional
3322 *) First EC_METHODs for curves over GF(p):
3324 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
3325 operations and provides various method functions that can also
3326 operate with faster implementations of modular arithmetic.
3328 EC_GFp_mont_method() reuses most functions that are part of
3329 EC_GFp_simple_method, but uses Montgomery arithmetic.
3331 [Bodo Moeller; point addition and point doubling
3332 implementation directly derived from source code provided by
3333 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
3335 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
3336 crypto/ec/ec_lib.c):
3338 Curves are EC_GROUP objects (with an optional group generator)
3339 based on EC_METHODs that are built into the library.
3341 Points are EC_POINT objects based on EC_GROUP objects.
3343 Most of the framework would be able to handle curves over arbitrary
3344 finite fields, but as there are no obvious types for fields other
3345 than GF(p), some functions are limited to that for now.
3348 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
3349 that the file contains a complete HTTP response.
3352 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
3353 change the def and num file printf format specifier from "%-40sXXX"
3354 to "%-39s XXX". The latter will always guarantee a space after the
3355 field while the former will cause them to run together if the field
3356 is 40 of more characters long.
3359 *) Constify the cipher and digest 'method' functions and structures
3360 and modify related functions to take constant EVP_MD and EVP_CIPHER
3364 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
3365 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
3368 *) Modify EVP_Digest*() routines so they now return values. Although the
3369 internal software routines can never fail additional hardware versions
3373 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
3375 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
3376 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
3379 ERR_R_NESTED_ASN1_ERROR
3381 ERR_R_MISSING_ASN1_EOS
3382 were 4 .. 9, conflicting with
3383 ERR_LIB_RSA (= ERR_R_RSA_LIB)
3385 ERR_LIB_PEM (= ERR_R_PEM_LIB).
3386 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
3388 Add new error code 'ERR_R_INTERNAL_ERROR'.
3391 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
3395 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
3396 sets the subject name for a new request or supersedes the
3397 subject name in a given request. Formats that can be parsed are
3398 'CN=Some Name, OU=myOU, C=IT'
3400 'CN=Some Name/OU=myOU/C=IT'.
3402 Add options '-batch' and '-verbose' to 'openssl req'.
3403 [Massimiliano Pala <madwolf@hackmasters.net>]
3405 *) Introduce the possibility to access global variables through
3406 functions on platform were that's the best way to handle exporting
3407 global variables in shared libraries. To enable this functionality,
3408 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
3409 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
3410 is normally done by Configure or something similar).
3412 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
3413 in the source file (foo.c) like this:
3415 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
3416 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
3418 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
3419 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
3421 OPENSSL_DECLARE_GLOBAL(int,foo);
3422 #define foo OPENSSL_GLOBAL_REF(foo)
3423 OPENSSL_DECLARE_GLOBAL(double,bar);
3424 #define bar OPENSSL_GLOBAL_REF(bar)
3426 The #defines are very important, and therefore so is including the
3427 header file everywhere where the defined globals are used.
3429 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
3430 of ASN.1 items, but that structure is a bit different.
3432 The largest change is in util/mkdef.pl which has been enhanced with
3433 better and easier to understand logic to choose which symbols should
3434 go into the Windows .def files as well as a number of fixes and code
3435 cleanup (among others, algorithm keywords are now sorted
3436 lexicographically to avoid constant rewrites).
3439 *) In BN_div() keep a copy of the sign of 'num' before writing the
3440 result to 'rm' because if rm==num the value will be overwritten
3441 and produce the wrong result if 'num' is negative: this caused
3442 problems with BN_mod() and BN_nnmod().
3445 *) Function OCSP_request_verify(). This checks the signature on an
3446 OCSP request and verifies the signer certificate. The signer
3447 certificate is just checked for a generic purpose and OCSP request
3451 *) Add OCSP_check_validity() function to check the validity of OCSP
3452 responses. OCSP responses are prepared in real time and may only
3453 be a few seconds old. Simply checking that the current time lies
3454 between thisUpdate and nextUpdate max reject otherwise valid responses
3455 caused by either OCSP responder or client clock inaccuracy. Instead
3456 we allow thisUpdate and nextUpdate to fall within a certain period of
3457 the current time. The age of the response can also optionally be
3458 checked. Two new options -validity_period and -status_age added to
3462 *) If signature or public key algorithm is unrecognized print out its
3463 OID rather that just UNKNOWN.
3466 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
3467 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
3468 ID to be generated from the issuer certificate alone which can then be
3469 passed to OCSP_id_issuer_cmp().
3472 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
3473 ASN1 modules to export functions returning ASN1_ITEM pointers
3474 instead of the ASN1_ITEM structures themselves. This adds several
3475 new macros which allow the underlying ASN1 function/structure to
3476 be accessed transparently. As a result code should not use ASN1_ITEM
3477 references directly (such as &X509_it) but instead use the relevant
3478 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
3479 use of the new ASN1 code on platforms where exporting structures
3480 is problematical (for example in shared libraries) but exporting
3481 functions returning pointers to structures is not.
3484 *) Add support for overriding the generation of SSL/TLS session IDs.
3485 These callbacks can be registered either in an SSL_CTX or per SSL.
3486 The purpose of this is to allow applications to control, if they wish,
3487 the arbitrary values chosen for use as session IDs, particularly as it
3488 can be useful for session caching in multiple-server environments. A
3489 command-line switch for testing this (and any client code that wishes
3490 to use such a feature) has been added to "s_server".
3491 [Geoff Thorpe, Lutz Jaenicke]
3493 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
3494 of the form '#if defined(...) || defined(...) || ...' and
3495 '#if !defined(...) && !defined(...) && ...'. This also avoids
3496 the growing number of special cases it was previously handling.
3499 *) Make all configuration macros available for application by making
3500 sure they are available in opensslconf.h, by giving them names starting
3501 with "OPENSSL_" to avoid conflicts with other packages and by making
3502 sure e_os2.h will cover all platform-specific cases together with
3504 Additionally, it is now possible to define configuration/platform-
3505 specific names (called "system identities"). In the C code, these
3506 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
3507 macro with the name beginning with "OPENSSL_SYS_", which is determined
3508 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
3512 *) New option -set_serial to 'req' and 'x509' this allows the serial
3513 number to use to be specified on the command line. Previously self
3514 signed certificates were hard coded with serial number 0 and the
3515 CA options of 'x509' had to use a serial number in a file which was
3519 *) New options to 'ca' utility to support V2 CRL entry extensions.
3520 Currently CRL reason, invalidity date and hold instruction are
3521 supported. Add new CRL extensions to V3 code and some new objects.
3524 *) New function EVP_CIPHER_CTX_set_padding() this is used to
3525 disable standard block padding (aka PKCS#5 padding) in the EVP
3526 API, which was previously mandatory. This means that the data is
3527 not padded in any way and so the total length much be a multiple
3528 of the block size, otherwise an error occurs.
3531 *) Initial (incomplete) OCSP SSL support.
3534 *) New function OCSP_parse_url(). This splits up a URL into its host,
3535 port and path components: primarily to parse OCSP URLs. New -url
3536 option to ocsp utility.
3539 *) New nonce behavior. The return value of OCSP_check_nonce() now
3540 reflects the various checks performed. Applications can decide
3541 whether to tolerate certain situations such as an absent nonce
3542 in a response when one was present in a request: the ocsp application
3543 just prints out a warning. New function OCSP_add1_basic_nonce()
3544 this is to allow responders to include a nonce in a response even if
3545 the request is nonce-less.
3548 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
3549 skipped when using openssl x509 multiple times on a single input file,
3550 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
3553 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
3554 set string type: to handle setting ASN1_TIME structures. Fix ca
3555 utility to correctly initialize revocation date of CRLs.
3558 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
3559 the clients preferred ciphersuites and rather use its own preferences.
3560 Should help to work around M$ SGC (Server Gated Cryptography) bug in
3561 Internet Explorer by ensuring unchanged hash method during stepup.
3562 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
3565 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
3566 to aes and add a new 'exist' option to print out symbols that don't
3570 *) Additional options to ocsp utility to allow flags to be set and
3571 additional certificates supplied.
3574 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
3575 OCSP client a number of certificate to only verify the response
3579 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
3580 handle the new API. Currently only ECB, CBC modes supported. Add new
3583 Add TLS AES ciphersuites as described in RFC3268, "Advanced
3584 Encryption Standard (AES) Ciphersuites for Transport Layer
3585 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
3586 not enabled by default and were not part of the "ALL" ciphersuite
3587 alias because they were not yet official; they could be
3588 explicitly requested by specifying the "AESdraft" ciphersuite
3589 group alias. In the final release of OpenSSL 0.9.7, the group
3590 alias is called "AES" and is part of "ALL".)
3591 [Ben Laurie, Steve Henson, Bodo Moeller]
3593 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
3594 request to response.
3597 *) Functions for OCSP responders. OCSP_request_onereq_count(),
3598 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
3599 extract information from a certificate request. OCSP_response_create()
3600 creates a response and optionally adds a basic response structure.
3601 OCSP_basic_add1_status() adds a complete single response to a basic
3602 response and returns the OCSP_SINGLERESP structure just added (to allow
3603 extensions to be included for example). OCSP_basic_add1_cert() adds a
3604 certificate to a basic response and OCSP_basic_sign() signs a basic
3605 response with various flags. New helper functions ASN1_TIME_check()
3606 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
3607 (converts ASN1_TIME to GeneralizedTime).
3610 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
3611 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
3612 structure from a certificate. X509_pubkey_digest() digests the public_key
3613 contents: this is used in various key identifiers.
3616 *) Make sk_sort() tolerate a NULL argument.
3617 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
3619 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
3620 passed by the function are trusted implicitly. If any of them signed the
3621 response then it is assumed to be valid and is not verified.
3624 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
3625 to data. This was previously part of the PKCS7 ASN1 code. This
3626 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
3627 [Steve Henson, reported by Kenneth R. Robinette
3628 <support@securenetterm.com>]
3630 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
3631 routines: without these tracing memory leaks is very painful.
3632 Fix leaks in PKCS12 and PKCS7 routines.
3635 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
3636 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
3637 effectively meant GeneralizedTime would never be used. Now it
3638 is initialised to -1 but X509_time_adj() now has to check the value
3639 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
3640 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
3641 [Steve Henson, reported by Kenneth R. Robinette
3642 <support@securenetterm.com>]
3644 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
3645 result in a zero length in the ASN1_INTEGER structure which was
3646 not consistent with the structure when d2i_ASN1_INTEGER() was used
3647 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
3648 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
3649 where it did not print out a minus for negative ASN1_INTEGER.
3652 *) Add summary printout to ocsp utility. The various functions which
3653 convert status values to strings have been renamed to:
3654 OCSP_response_status_str(), OCSP_cert_status_str() and
3655 OCSP_crl_reason_str() and are no longer static. New options
3656 to verify nonce values and to disable verification. OCSP response
3657 printout format cleaned up.
3660 *) Add additional OCSP certificate checks. These are those specified
3661 in RFC2560. This consists of two separate checks: the CA of the
3662 certificate being checked must either be the OCSP signer certificate
3663 or the issuer of the OCSP signer certificate. In the latter case the
3664 OCSP signer certificate must contain the OCSP signing extended key
3665 usage. This check is performed by attempting to match the OCSP
3666 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
3667 in the OCSP_CERTID structures of the response.
3670 *) Initial OCSP certificate verification added to OCSP_basic_verify()
3671 and related routines. This uses the standard OpenSSL certificate
3672 verify routines to perform initial checks (just CA validity) and
3673 to obtain the certificate chain. Then additional checks will be
3674 performed on the chain. Currently the root CA is checked to see
3675 if it is explicitly trusted for OCSP signing. This is used to set
3676 a root CA as a global signing root: that is any certificate that
3677 chains to that CA is an acceptable OCSP signing certificate.
3680 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
3681 extensions from a separate configuration file.
3682 As when reading extensions from the main configuration file,
3683 the '-extensions ...' option may be used for specifying the
3685 [Massimiliano Pala <madwolf@comune.modena.it>]
3687 *) New OCSP utility. Allows OCSP requests to be generated or
3688 read. The request can be sent to a responder and the output
3689 parsed, outputed or printed in text form. Not complete yet:
3690 still needs to check the OCSP response validity.
3693 *) New subcommands for 'openssl ca':
3694 'openssl ca -status <serial>' prints the status of the cert with
3695 the given serial number (according to the index file).
3696 'openssl ca -updatedb' updates the expiry status of certificates
3698 [Massimiliano Pala <madwolf@comune.modena.it>]
3700 *) New '-newreq-nodes' command option to CA.pl. This is like
3701 '-newreq', but calls 'openssl req' with the '-nodes' option
3702 so that the resulting key is not encrypted.
3703 [Damien Miller <djm@mindrot.org>]
3705 *) New configuration for the GNU Hurd.
3706 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
3708 *) Initial code to implement OCSP basic response verify. This
3709 is currently incomplete. Currently just finds the signer's
3710 certificate and verifies the signature on the response.
3713 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
3714 value of OPENSSLDIR. This is available via the new '-d' option
3715 to 'openssl version', and is also included in 'openssl version -a'.
3718 *) Allowing defining memory allocation callbacks that will be given
3719 file name and line number information in additional arguments
3720 (a const char* and an int). The basic functionality remains, as
3721 well as the original possibility to just replace malloc(),
3722 realloc() and free() by functions that do not know about these
3723 additional arguments. To register and find out the current
3724 settings for extended allocation functions, the following
3725 functions are provided:
3727 CRYPTO_set_mem_ex_functions
3728 CRYPTO_set_locked_mem_ex_functions
3729 CRYPTO_get_mem_ex_functions
3730 CRYPTO_get_locked_mem_ex_functions
3732 These work the same way as CRYPTO_set_mem_functions and friends.
3733 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
3734 extended allocation function is enabled.
3735 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
3736 a conventional allocation function is enabled.
3737 [Richard Levitte, Bodo Moeller]
3739 *) Finish off removing the remaining LHASH function pointer casts.
3740 There should no longer be any prototype-casting required when using
3741 the LHASH abstraction, and any casts that remain are "bugs". See
3742 the callback types and macros at the head of lhash.h for details
3743 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
3746 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
3747 If /dev/[u]random devices are not available or do not return enough
3748 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
3750 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
3751 /etc/entropy will be queried once each in this sequence, quering stops
3752 when enough entropy was collected without querying more sockets.
3755 *) Change the Unix RAND_poll() variant to be able to poll several
3756 random devices, as specified by DEVRANDOM, until a sufficient amount
3757 of data has been collected. We spend at most 10 ms on each file
3758 (select timeout) and read in non-blocking mode. DEVRANDOM now
3759 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
3760 (previously it was just the string "/dev/urandom"), so on typical
3761 platforms the 10 ms delay will never occur.
3762 Also separate out the Unix variant to its own file, rand_unix.c.
3763 For VMS, there's a currently-empty rand_vms.c.
3766 *) Move OCSP client related routines to ocsp_cl.c. These
3767 provide utility functions which an application needing
3768 to issue a request to an OCSP responder and analyse the
3769 response will typically need: as opposed to those which an
3770 OCSP responder itself would need which will be added later.
3772 OCSP_request_sign() signs an OCSP request with an API similar
3773 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
3774 response. OCSP_response_get1_basic() extracts basic response
3775 from response. OCSP_resp_find_status(): finds and extracts status
3776 information from an OCSP_CERTID structure (which will be created
3777 when the request structure is built). These are built from lower
3778 level functions which work on OCSP_SINGLERESP structures but
3779 wont normally be used unless the application wishes to examine
3780 extensions in the OCSP response for example.
3782 Replace nonce routines with a pair of functions.
3783 OCSP_request_add1_nonce() adds a nonce value and optionally
3784 generates a random value. OCSP_check_nonce() checks the
3785 validity of the nonce in an OCSP response.
3788 *) Change function OCSP_request_add() to OCSP_request_add0_id().
3789 This doesn't copy the supplied OCSP_CERTID and avoids the
3790 need to free up the newly created id. Change return type
3791 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
3792 This can then be used to add extensions to the request.
3793 Deleted OCSP_request_new(), since most of its functionality
3794 is now in OCSP_REQUEST_new() (and the case insensitive name
3795 clash) apart from the ability to set the request name which
3796 will be added elsewhere.
3799 *) Update OCSP API. Remove obsolete extensions argument from
3800 various functions. Extensions are now handled using the new
3801 OCSP extension code. New simple OCSP HTTP function which
3802 can be used to send requests and parse the response.
3805 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
3806 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
3807 uses the special reorder version of SET OF to sort the attributes
3808 and reorder them to match the encoded order. This resolves a long
3809 standing problem: a verify on a PKCS7 structure just after signing
3810 it used to fail because the attribute order did not match the
3811 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
3812 it uses the received order. This is necessary to tolerate some broken
3813 software that does not order SET OF. This is handled by encoding
3814 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
3815 to produce the required SET OF.
3818 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
3819 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
3820 files to get correct declarations of the ASN.1 item variables.
3823 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
3824 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
3825 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
3826 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
3827 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
3828 ASN1_ITEM and no wrapper functions.
3831 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
3832 replace the old function pointer based I/O routines. Change most of
3833 the *_d2i_bio() and *_d2i_fp() functions to use these.
3836 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
3837 lines, recognice more "algorithms" that can be deselected, and make
3838 it complain about algorithm deselection that isn't recognised.
3841 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
3842 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
3843 to use new functions. Add NO_ASN1_OLD which can be set to remove
3844 some old style ASN1 functions: this can be used to determine if old
3845 code will still work when these eventually go away.
3848 *) New extension functions for OCSP structures, these follow the
3849 same conventions as certificates and CRLs.
3852 *) New function X509V3_add1_i2d(). This automatically encodes and
3853 adds an extension. Its behaviour can be customised with various
3854 flags to append, replace or delete. Various wrappers added for
3855 certifcates and CRLs.
3858 *) Fix to avoid calling the underlying ASN1 print routine when
3859 an extension cannot be parsed. Correct a typo in the
3860 OCSP_SERVICELOC extension. Tidy up print OCSP format.
3863 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
3864 entries for variables.
3867 *) Add functionality to apps/openssl.c for detecting locking
3868 problems: As the program is single-threaded, all we have
3869 to do is register a locking callback using an array for
3870 storing which locks are currently held by the program.
3873 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
3874 SSL_get_ex_data_X509_STORE_idx(), which is used in
3875 ssl_verify_cert_chain() and thus can be called at any time
3876 during TLS/SSL handshakes so that thread-safety is essential.
3877 Unfortunately, the ex_data design is not at all suited
3878 for multi-threaded use, so it probably should be abolished.
3881 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
3882 [Broadcom, tweaked and integrated by Geoff Thorpe]
3884 *) Move common extension printing code to new function
3885 X509V3_print_extensions(). Reorganise OCSP print routines and
3886 implement some needed OCSP ASN1 functions. Add OCSP extensions.
3889 *) New function X509_signature_print() to remove duplication in some
3893 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
3894 set (this was treated exactly the same as SET OF previously). This
3895 is used to reorder the STACK representing the structure to match the
3896 encoding. This will be used to get round a problem where a PKCS7
3897 structure which was signed could not be verified because the STACK
3898 order did not reflect the encoded order.
3901 *) Reimplement the OCSP ASN1 module using the new code.
3904 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
3905 for its ASN1 operations. The old style function pointers still exist
3906 for now but they will eventually go away.
3909 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
3910 completely replaces the old ASN1 functionality with a table driven
3911 encoder and decoder which interprets an ASN1_ITEM structure describing
3912 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
3913 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
3914 has also been converted to the new form.
3917 *) Change BN_mod_exp_recp so that negative moduli are tolerated
3918 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
3919 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
3920 for negative moduli.
3923 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
3924 of not touching the result's sign bit.
3927 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
3931 *) Changed the LHASH code to use prototypes for callbacks, and created
3932 macros to declare and implement thin (optionally static) functions
3933 that provide type-safety and avoid function pointer casting for the
3934 type-specific callbacks.
3937 *) Added Kerberos Cipher Suites to be used with TLS, as written in
3939 [Veers Staats <staatsvr@asc.hpc.mil>,
3940 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
3942 *) Reformat the FAQ so the different questions and answers can be divided
3943 in sections depending on the subject.
3946 *) Have the zlib compression code load ZLIB.DLL dynamically under
3950 *) New function BN_mod_sqrt for computing square roots modulo a prime
3951 (using the probabilistic Tonelli-Shanks algorithm unless
3952 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
3953 be handled deterministically).
3954 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3956 *) Make BN_mod_inverse faster by explicitly handling small quotients
3957 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
3958 512 bits], about 30% for larger ones [1024 or 2048 bits].)
3961 *) New function BN_kronecker.
3964 *) Fix BN_gcd so that it works on negative inputs; the result is
3965 positive unless both parameters are zero.
3966 Previously something reasonably close to an infinite loop was
3967 possible because numbers could be growing instead of shrinking
3968 in the implementation of Euclid's algorithm.
3971 *) Fix BN_is_word() and BN_is_one() macros to take into account the
3972 sign of the number in question.
3974 Fix BN_is_word(a,w) to work correctly for w == 0.
3976 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
3977 because its test if the absolute value of 'a' equals 'w'.
3978 Note that BN_abs_is_word does *not* handle w == 0 reliably;
3979 it exists mostly for use in the implementations of BN_is_zero(),
3980 BN_is_one(), and BN_is_word().
3983 *) New function BN_swap.
3986 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
3987 the exponentiation functions are more likely to produce reasonable
3988 results on negative inputs.
3991 *) Change BN_mod_mul so that the result is always non-negative.
3992 Previously, it could be negative if one of the factors was negative;
3993 I don't think anyone really wanted that behaviour.
3996 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
3997 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
3998 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
3999 and add new functions:
4008 BN_mod_lshift1_quick
4012 These functions always generate non-negative results.
4014 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
4015 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
4017 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
4018 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
4019 be reduced modulo m.
4020 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
4023 The following entry accidentily appeared in the CHANGES file
4024 distributed with OpenSSL 0.9.7. The modifications described in
4025 it do *not* apply to OpenSSL 0.9.7.
4027 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4028 was actually never needed) and in BN_mul(). The removal in BN_mul()
4029 required a small change in bn_mul_part_recursive() and the addition
4030 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4031 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4032 bn_sub_words() and bn_add_words() except they take arrays with
4037 *) In 'openssl passwd', verify passwords read from the terminal
4038 unless the '-salt' option is used (which usually means that
4039 verification would just waste user's time since the resulting
4040 hash is going to be compared with some given password hash)
4041 or the new '-noverify' option is used.
4043 This is an incompatible change, but it does not affect
4044 non-interactive use of 'openssl passwd' (passwords on the command
4045 line, '-stdin' option, '-in ...' option) and thus should not
4049 *) Remove all references to RSAref, since there's no more need for it.
4052 *) Make DSO load along a path given through an environment variable
4053 (SHLIB_PATH) with shl_load().
4056 *) Constify the ENGINE code as a result of BIGNUM constification.
4057 Also constify the RSA code and most things related to it. In a
4058 few places, most notable in the depth of the ASN.1 code, ugly
4059 casts back to non-const were required (to be solved at a later
4063 *) Make it so the openssl application has all engines loaded by default.
4066 *) Constify the BIGNUM routines a little more.
4069 *) Add the following functions:
4071 ENGINE_load_cswift()
4073 ENGINE_load_atalla()
4075 ENGINE_load_builtin_engines()
4077 That way, an application can itself choose if external engines that
4078 are built-in in OpenSSL shall ever be used or not. The benefit is
4079 that applications won't have to be linked with libdl or other dso
4080 libraries unless it's really needed.
4082 Changed 'openssl engine' to load all engines on demand.
4083 Changed the engine header files to avoid the duplication of some
4084 declarations (they differed!).
4087 *) 'openssl engine' can now list capabilities.
4090 *) Better error reporting in 'openssl engine'.
4093 *) Never call load_dh_param(NULL) in s_server.
4096 *) Add engine application. It can currently list engines by name and
4097 identity, and test if they are actually available.
4100 *) Improve RPM specification file by forcing symbolic linking and making
4101 sure the installed documentation is also owned by root.root.
4102 [Damien Miller <djm@mindrot.org>]
4104 *) Give the OpenSSL applications more possibilities to make use of
4105 keys (public as well as private) handled by engines.
4108 *) Add OCSP code that comes from CertCo.
4111 *) Add VMS support for the Rijndael code.
4114 *) Added untested support for Nuron crypto accelerator.
4117 *) Add support for external cryptographic devices. This code was
4118 previously distributed separately as the "engine" branch.
4119 [Geoff Thorpe, Richard Levitte]
4121 *) Rework the filename-translation in the DSO code. It is now possible to
4122 have far greater control over how a "name" is turned into a filename
4123 depending on the operating environment and any oddities about the
4124 different shared library filenames on each system.
4127 *) Support threads on FreeBSD-elf in Configure.
4130 *) Fix for SHA1 assembly problem with MASM: it produces
4131 warnings about corrupt line number information when assembling
4132 with debugging information. This is caused by the overlapping
4134 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
4137 NCONF_get_number() has no error checking at all. As a replacement,
4138 NCONF_get_number_e() is defined (_e for "error checking") and is
4139 promoted strongly. The old NCONF_get_number is kept around for
4140 binary backward compatibility.
4141 Make it possible for methods to load from something other than a BIO,
4142 by providing a function pointer that is given a name instead of a BIO.
4143 For example, this could be used to load configuration data from an
4147 *) Fix for non blocking accept BIOs. Added new I/O special reason
4148 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
4149 with non blocking I/O was not possible because no retry code was
4150 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
4154 *) Added the beginnings of Rijndael support.
4157 *) Fix for bug in DirectoryString mask setting. Add support for
4158 X509_NAME_print_ex() in 'req' and X509_print_ex() function
4159 to allow certificate printing to more controllable, additional
4160 'certopt' option to 'x509' to allow new printing options to be
4164 *) Clean old EAY MD5 hack from e_os.h.
4167 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
4169 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4170 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4171 [Joe Orton, Steve Henson]
4173 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
4175 *) Fix additional bug revealed by the NISCC test suite:
4177 Stop bug triggering large recursion when presented with
4178 certain ASN.1 tags (CVE-2003-0851)
4181 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
4183 *) Fix various bugs revealed by running the NISCC test suite:
4185 Stop out of bounds reads in the ASN1 code when presented with
4186 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4188 If verify callback ignores invalid public key errors don't try to check
4189 certificate signature with the NULL public key.
4193 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4194 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4198 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4199 extra data after the compression methods not only for TLS 1.0
4200 but also for SSL 3.0 (as required by the specification).
4201 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4203 *) Change X509_certificate_type() to mark the key as exported/exportable
4204 when it's 512 *bits* long, not 512 bytes.
4207 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
4209 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4210 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4211 a protocol version number mismatch like a decryption error
4212 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4215 *) Turn on RSA blinding by default in the default implementation
4216 to avoid a timing attack. Applications that don't want it can call
4217 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4218 They would be ill-advised to do so in most cases.
4219 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4221 *) Change RSA blinding code so that it works when the PRNG is not
4222 seeded (in this case, the secret RSA exponent is abused as
4223 an unpredictable seed -- if it is not unpredictable, there
4224 is no point in blinding anyway). Make RSA blinding thread-safe
4225 by remembering the creator's thread ID in rsa->blinding and
4226 having all other threads use local one-time blinding factors
4227 (this requires more computation than sharing rsa->blinding, but
4228 avoids excessive locking; and if an RSA object is not shared
4229 between threads, blinding will still be very fast).
4232 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
4234 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4235 via timing by performing a MAC computation even if incorrrect
4236 block cipher padding has been found. This is a countermeasure
4237 against active attacks where the attacker has to distinguish
4238 between bad padding and a MAC verification error. (CVE-2003-0078)
4240 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4241 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4242 Martin Vuagnoux (EPFL, Ilion)]
4244 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
4246 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
4247 memory from it's contents. This is done with a counter that will
4248 place alternating values in each byte. This can be used to solve
4249 two issues: 1) the removal of calls to memset() by highly optimizing
4250 compilers, and 2) cleansing with other values than 0, since those can
4251 be read through on certain media, for example a swap space on disk.
4254 *) Bugfix: client side session caching did not work with external caching,
4255 because the session->cipher setting was not restored when reloading
4256 from the external cache. This problem was masked, when
4257 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
4258 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
4261 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
4262 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
4263 [Zeev Lieber <zeev-l@yahoo.com>]
4265 *) Undo an undocumented change introduced in 0.9.6e which caused
4266 repeated calls to OpenSSL_add_all_ciphers() and
4267 OpenSSL_add_all_digests() to be ignored, even after calling
4271 *) Change the default configuration reader to deal with last line not
4272 being properly terminated.
4275 *) Change X509_NAME_cmp() so it applies the special rules on handling
4276 DN values that are of type PrintableString, as well as RDNs of type
4277 emailAddress where the value has the type ia5String.
4278 [stefank@valicert.com via Richard Levitte]
4280 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
4281 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
4282 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
4283 the bitwise-OR of the two for use by the majority of applications
4284 wanting this behaviour, and update the docs. The documented
4285 behaviour and actual behaviour were inconsistent and had been
4286 changing anyway, so this is more a bug-fix than a behavioural
4288 [Geoff Thorpe, diagnosed by Nadav Har'El]
4290 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
4291 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
4294 *) Fix initialization code race conditions in
4295 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
4296 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
4297 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
4298 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
4299 ssl2_get_cipher_by_char(),
4300 ssl3_get_cipher_by_char().
4301 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
4303 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
4304 the cached sessions are flushed, as the remove_cb() might use ex_data
4305 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
4306 (see [openssl.org #212]).
4307 [Geoff Thorpe, Lutz Jaenicke]
4309 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
4310 length, instead of the encoding length to d2i_ASN1_OBJECT.
4313 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
4315 *) [In 0.9.6g-engine release:]
4316 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
4317 [Lynn Gazis <lgazis@rainbow.com>]
4319 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
4321 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
4322 and get fix the header length calculation.
4323 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
4324 Alon Kantor <alonk@checkpoint.com> (and others),
4327 *) Use proper error handling instead of 'assertions' in buffer
4328 overflow checks added in 0.9.6e. This prevents DoS (the
4329 assertions could call abort()).
4330 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
4332 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
4334 *) Add various sanity checks to asn1_get_length() to reject
4335 the ASN1 length bytes if they exceed sizeof(long), will appear
4336 negative or the content length exceeds the length of the
4338 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4340 *) Fix cipher selection routines: ciphers without encryption had no flags
4341 for the cipher strength set and where therefore not handled correctly
4342 by the selection routines (PR #130).
4345 *) Fix EVP_dsa_sha macro.
4349 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
4350 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
4351 that was added in OpenSSL 0.9.6d.
4353 As the countermeasure turned out to be incompatible with some
4354 broken SSL implementations, the new option is part of SSL_OP_ALL.
4355 SSL_OP_ALL is usually employed when compatibility with weird SSL
4356 implementations is desired (e.g. '-bugs' option to 's_client' and
4357 's_server'), so the new option is automatically set in many
4361 *) Changes in security patch:
4363 Changes marked "(CHATS)" were sponsored by the Defense Advanced
4364 Research Projects Agency (DARPA) and Air Force Research Laboratory,
4365 Air Force Materiel Command, USAF, under agreement number
4368 *) Add various sanity checks to asn1_get_length() to reject
4369 the ASN1 length bytes if they exceed sizeof(long), will appear
4370 negative or the content length exceeds the length of the
4371 supplied buffer. (CVE-2002-0659)
4372 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4374 *) Assertions for various potential buffer overflows, not known to
4376 [Ben Laurie (CHATS)]
4378 *) Various temporary buffers to hold ASCII versions of integers were
4379 too small for 64 bit platforms. (CVE-2002-0655)
4380 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
4382 *) Remote buffer overflow in SSL3 protocol - an attacker could
4383 supply an oversized session ID to a client. (CVE-2002-0656)
4384 [Ben Laurie (CHATS)]
4386 *) Remote buffer overflow in SSL2 protocol - an attacker could
4387 supply an oversized client master key. (CVE-2002-0656)
4388 [Ben Laurie (CHATS)]
4390 Changes between 0.9.6c and 0.9.6d [9 May 2002]
4392 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
4393 encoded as NULL) with id-dsa-with-sha1.
4394 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
4396 *) Check various X509_...() return values in apps/req.c.
4397 [Nils Larsch <nla@trustcenter.de>]
4399 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
4400 an end-of-file condition would erronously be flagged, when the CRLF
4401 was just at the end of a processed block. The bug was discovered when
4402 processing data through a buffering memory BIO handing the data to a
4403 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
4404 <ptsekov@syntrex.com> and Nedelcho Stanev.
4407 *) Implement a countermeasure against a vulnerability recently found
4408 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
4409 before application data chunks to avoid the use of known IVs
4410 with data potentially chosen by the attacker.
4413 *) Fix length checks in ssl3_get_client_hello().
4416 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
4417 to prevent ssl3_read_internal() from incorrectly assuming that
4418 ssl3_read_bytes() found application data while handshake
4419 processing was enabled when in fact s->s3->in_read_app_data was
4420 merely automatically cleared during the initial handshake.
4421 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
4423 *) Fix object definitions for Private and Enterprise: they were not
4424 recognized in their shortname (=lowercase) representation. Extend
4425 obj_dat.pl to issue an error when using undefined keywords instead
4426 of silently ignoring the problem (Svenning Sorensen
4427 <sss@sss.dnsalias.net>).
4430 *) Fix DH_generate_parameters() so that it works for 'non-standard'
4431 generators, i.e. generators other than 2 and 5. (Previously, the
4432 code did not properly initialise the 'add' and 'rem' values to
4433 BN_generate_prime().)
4435 In the new general case, we do not insist that 'generator' is
4436 actually a primitive root: This requirement is rather pointless;
4437 a generator of the order-q subgroup is just as good, if not
4441 *) Map new X509 verification errors to alerts. Discovered and submitted by
4442 Tom Wu <tom@arcot.com>.
4445 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
4446 returning non-zero before the data has been completely received
4447 when using non-blocking I/O.
4448 [Bodo Moeller; problem pointed out by John Hughes]
4450 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
4451 [Ben Laurie, Lutz Jaenicke]
4453 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
4454 Yoram Zahavi <YoramZ@gilian.com>).
4457 *) Add information about CygWin 1.3 and on, and preserve proper
4458 configuration for the versions before that.
4459 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4461 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
4462 check whether we deal with a copy of a session and do not delete from
4463 the cache in this case. Problem reported by "Izhar Shoshani Levi"
4464 <izhar@checkpoint.com>.
4467 *) Do not store session data into the internal session cache, if it
4468 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
4469 flag is set). Proposed by Aslam <aslam@funk.com>.
4472 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
4476 *) [In 0.9.6d-engine release:]
4477 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4478 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4480 *) Add the configuration target linux-s390x.
4481 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
4483 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
4484 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
4485 variable as an indication that a ClientHello message has been
4486 received. As the flag value will be lost between multiple
4487 invocations of ssl3_accept when using non-blocking I/O, the
4488 function may not be aware that a handshake has actually taken
4489 place, thus preventing a new session from being added to the
4492 To avoid this problem, we now set s->new_session to 2 instead of
4493 using a local variable.
4494 [Lutz Jaenicke, Bodo Moeller]
4496 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
4497 if the SSL_R_LENGTH_MISMATCH error is detected.
4498 [Geoff Thorpe, Bodo Moeller]
4500 *) New 'shared_ldflag' column in Configure platform table.
4503 *) Fix EVP_CIPHER_mode macro.
4504 ["Dan S. Camper" <dan@bti.net>]
4506 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
4507 type, we must throw them away by setting rr->length to 0.
4508 [D P Chang <dpc@qualys.com>]
4510 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
4512 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
4513 <Dominikus.Scherkl@biodata.com>. (The previous implementation
4514 worked incorrectly for those cases where range = 10..._2 and
4515 3*range is two bits longer than range.)
4518 *) Only add signing time to PKCS7 structures if it is not already
4522 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
4523 OBJ_ld_ce should be OBJ_id_ce.
4524 Also some ip-pda OIDs in crypto/objects/objects.txt were
4525 incorrect (cf. RFC 3039).
4526 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
4528 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
4529 returns early because it has nothing to do.
4530 [Andy Schneider <andy.schneider@bjss.co.uk>]
4532 *) [In 0.9.6c-engine release:]
4533 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
4534 [Andy Schneider <andy.schneider@bjss.co.uk>]
4536 *) [In 0.9.6c-engine release:]
4537 Add support for Cryptographic Appliance's keyserver technology.
4538 (Use engine 'keyclient')
4539 [Cryptographic Appliances and Geoff Thorpe]
4541 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
4542 is called via tools/c89.sh because arguments have to be
4543 rearranged (all '-L' options must appear before the first object
4545 [Richard Shapiro <rshapiro@abinitio.com>]
4547 *) [In 0.9.6c-engine release:]
4548 Add support for Broadcom crypto accelerator cards, backported
4550 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
4552 *) [In 0.9.6c-engine release:]
4553 Add support for SureWare crypto accelerator cards from
4554 Baltimore Technologies. (Use engine 'sureware')
4555 [Baltimore Technologies and Mark Cox]
4557 *) [In 0.9.6c-engine release:]
4558 Add support for crypto accelerator cards from Accelerated
4559 Encryption Processing, www.aep.ie. (Use engine 'aep')
4560 [AEP Inc. and Mark Cox]
4562 *) Add a configuration entry for gcc on UnixWare.
4563 [Gary Benson <gbenson@redhat.com>]
4565 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
4566 messages are stored in a single piece (fixed-length part and
4567 variable-length part combined) and fix various bugs found on the way.
4570 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
4571 instead. BIO_gethostbyname() does not know what timeouts are
4572 appropriate, so entries would stay in cache even when they have
4574 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
4576 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
4577 faced with a pathologically small ClientHello fragment that does
4578 not contain client_version: Instead of aborting with an error,
4579 simply choose the highest available protocol version (i.e.,
4580 TLS 1.0 unless it is disabled). In practice, ClientHello
4581 messages are never sent like this, but this change gives us
4582 strictly correct behaviour at least for TLS.
4585 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
4586 never resets s->method to s->ctx->method when called from within
4587 one of the SSL handshake functions.
4588 [Bodo Moeller; problem pointed out by Niko Baric]
4590 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
4591 (sent using the client's version number) if client_version is
4592 smaller than the protocol version in use. Also change
4593 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
4594 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
4595 the client will at least see that alert.
4598 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
4602 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
4603 client receives HelloRequest while in a handshake.
4604 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
4606 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
4607 should end in 'break', not 'goto end' which circuments various
4608 cleanups done in state SSL_ST_OK. But session related stuff
4609 must be disabled for SSL_ST_OK in the case that we just sent a
4612 Also avoid some overhead by not calling ssl_init_wbio_buffer()
4613 before just sending a HelloRequest.
4614 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
4616 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
4617 reveal whether illegal block cipher padding was found or a MAC
4618 verification error occured. (Neither SSLerr() codes nor alerts
4619 are directly visible to potential attackers, but the information
4620 may leak via logfiles.)
4622 Similar changes are not required for the SSL 2.0 implementation
4623 because the number of padding bytes is sent in clear for SSL 2.0,
4624 and the extra bytes are just ignored. However ssl/s2_pkt.c
4625 failed to verify that the purported number of padding bytes is in
4629 *) Add OpenUNIX-8 support including shared libraries
4630 (Boyd Lynn Gerber <gerberb@zenez.com>).
4633 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
4634 'wristwatch attack' using huge encoding parameters (cf.
4635 James H. Manger's CRYPTO 2001 paper). Note that the
4636 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
4637 encoding parameters and hence was not vulnerable.
4640 *) BN_sqr() bug fix.
4641 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
4643 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
4644 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
4645 followed by modular reduction.
4646 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
4648 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
4649 equivalent based on BN_pseudo_rand() instead of BN_rand().
4652 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
4653 This function was broken, as the check for a new client hello message
4654 to handle SGC did not allow these large messages.
4655 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
4658 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
4661 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
4662 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
4665 *) Rework the configuration and shared library support for Tru64 Unix.
4666 The configuration part makes use of modern compiler features and
4667 still retains old compiler behavior for those that run older versions
4668 of the OS. The shared library support part includes a variant that
4669 uses the RPATH feature, and is available through the special
4670 configuration target "alpha-cc-rpath", which will never be selected
4672 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
4674 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
4675 with the same message size as in ssl3_get_certificate_request().
4676 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
4677 messages might inadvertently be reject as too long.
4678 [Petr Lampa <lampa@fee.vutbr.cz>]
4680 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
4683 *) Modified SSL library such that the verify_callback that has been set
4684 specificly for an SSL object with SSL_set_verify() is actually being
4685 used. Before the change, a verify_callback set with this function was
4686 ignored and the verify_callback() set in the SSL_CTX at the time of
4687 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
4688 to allow the necessary settings.
4691 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
4692 explicitly to NULL, as at least on Solaris 8 this seems not always to be
4693 done automatically (in contradiction to the requirements of the C
4694 standard). This made problems when used from OpenSSH.
4697 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
4698 dh->length and always used
4700 BN_rand_range(priv_key, dh->p).
4702 BN_rand_range() is not necessary for Diffie-Hellman, and this
4703 specific range makes Diffie-Hellman unnecessarily inefficient if
4704 dh->length (recommended exponent length) is much smaller than the
4705 length of dh->p. We could use BN_rand_range() if the order of
4706 the subgroup was stored in the DH structure, but we only have
4711 BN_rand(priv_key, l, ...)
4713 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
4719 RSA_eay_public_encrypt
4720 RSA_eay_private_decrypt
4721 RSA_eay_private_encrypt (signing)
4722 RSA_eay_public_decrypt (signature verification)
4724 (default implementations for RSA_public_encrypt,
4725 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
4726 always reject numbers >= n.
4729 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
4730 to synchronize access to 'locking_thread'. This is necessary on
4731 systems where access to 'locking_thread' (an 'unsigned long'
4732 variable) is not atomic.
4735 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
4736 *before* setting the 'crypto_lock_rand' flag. The previous code had
4737 a race condition if 0 is a valid thread ID.
4738 [Travis Vitek <vitek@roguewave.com>]
4740 *) Add support for shared libraries under Irix.
4741 [Albert Chin-A-Young <china@thewrittenword.com>]
4743 *) Add configuration option to build on Linux on both big-endian and
4745 [Ralf Baechle <ralf@uni-koblenz.de>]
4747 *) Add the possibility to create shared libraries on HP-UX.
4750 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
4752 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
4753 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
4754 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
4755 PRNG state recovery was possible based on the output of
4756 one PRNG request appropriately sized to gain knowledge on
4757 'md' followed by enough consecutive 1-byte PRNG requests
4758 to traverse all of 'state'.
4760 1. When updating 'md_local' (the current thread's copy of 'md')
4761 during PRNG output generation, hash all of the previous
4762 'md_local' value, not just the half used for PRNG output.
4764 2. Make the number of bytes from 'state' included into the hash
4765 independent from the number of PRNG bytes requested.
4767 The first measure alone would be sufficient to avoid
4768 Markku-Juhani's attack. (Actually it had never occurred
4769 to me that the half of 'md_local' used for chaining was the
4770 half from which PRNG output bytes were taken -- I had always
4771 assumed that the secret half would be used.) The second
4772 measure makes sure that additional data from 'state' is never
4773 mixed into 'md_local' in small portions; this heuristically
4774 further strengthens the PRNG.
4777 *) Fix crypto/bn/asm/mips3.s.
4780 *) When only the key is given to "enc", the IV is undefined. Print out
4781 an error message in this case.
4784 *) Handle special case when X509_NAME is empty in X509 printing routines.
4787 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
4788 positive and less than q.
4791 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
4792 used: it isn't thread safe and the add_lock_callback should handle
4794 [Paul Rose <Paul.Rose@bridge.com>]
4796 *) Verify that incoming data obeys the block size in
4797 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
4801 [Ulf Möller, Bodo Möller]
4803 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
4804 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
4805 when fixing the server behaviour for backwards-compatible 'client
4806 hello' messages. (Note that the attack is impractical against
4807 SSL 3.0 and TLS 1.0 anyway because length and version checking
4808 means that the probability of guessing a valid ciphertext is
4809 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
4812 Before 0.9.5, the countermeasure (hide the error by generating a
4813 random 'decryption result') did not work properly because
4814 ERR_clear_error() was missing, meaning that SSL_get_error() would
4815 detect the supposedly ignored error.
4817 Both problems are now fixed.
4820 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
4821 (previously it was 1024).
4824 *) Fix for compatibility mode trust settings: ignore trust settings
4825 unless some valid trust or reject settings are present.
4828 *) Fix for blowfish EVP: its a variable length cipher.
4831 *) Fix various bugs related to DSA S/MIME verification. Handle missing
4832 parameters in DSA public key structures and return an error in the
4833 DSA routines if parameters are absent.
4836 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
4837 in the current directory if neither $RANDFILE nor $HOME was set.
4838 RAND_file_name() in 0.9.6a returned NULL in this case. This has
4839 caused some confusion to Windows users who haven't defined $HOME.
4840 Thus RAND_file_name() is changed again: e_os.h can define a
4841 DEFAULT_HOME, which will be used if $HOME is not set.
4842 For Windows, we use "C:"; on other platforms, we still require
4843 environment variables.
4845 *) Move 'if (!initialized) RAND_poll()' into regions protected by
4846 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
4847 having multiple threads call RAND_poll() concurrently.
4850 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
4851 combination of a flag and a thread ID variable.
4852 Otherwise while one thread is in ssleay_rand_bytes (which sets the
4853 flag), *other* threads can enter ssleay_add_bytes without obeying
4854 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
4855 that they do not hold after the first thread unsets add_do_not_lock).
4858 *) Change bctest again: '-x' expressions are not available in all
4862 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
4864 *) Fix a couple of memory leaks in PKCS7_dataDecode()
4865 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
4867 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
4868 the default extension for executables, if any. Also, make the perl
4869 scripts that use symlink() to test if it really exists and use "cp"
4870 if it doesn't. All this made OpenSSL compilable and installable in
4874 *) Fix for asn1_GetSequence() for indefinite length constructed data.
4875 If SEQUENCE is length is indefinite just set c->slen to the total
4876 amount of data available.
4877 [Steve Henson, reported by shige@FreeBSD.org]
4878 [This change does not apply to 0.9.7.]
4880 *) Change bctest to avoid here-documents inside command substitution
4881 (workaround for FreeBSD /bin/sh bug).
4882 For compatibility with Ultrix, avoid shell functions (introduced
4883 in the bctest version that searches along $PATH).
4886 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
4887 with des_encrypt() defined on some operating systems, like Solaris
4891 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
4892 On the Importance of Eliminating Errors in Cryptographic
4893 Computations, J. Cryptology 14 (2001) 2, 101-119,
4894 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
4897 *) MIPS assembler BIGNUM division bug fix.
4900 *) Disabled incorrect Alpha assembler code.
4903 *) Fix PKCS#7 decode routines so they correctly update the length
4904 after reading an EOC for the EXPLICIT tag.
4906 [This change does not apply to 0.9.7.]
4908 *) Fix bug in PKCS#12 key generation routines. This was triggered
4909 if a 3DES key was generated with a 0 initial byte. Include
4910 PKCS12_BROKEN_KEYGEN compilation option to retain the old
4911 (but broken) behaviour.
4914 *) Enhance bctest to search for a working bc along $PATH and print
4916 [Tim Rice <tim@multitalents.net> via Richard Levitte]
4918 *) Fix memory leaks in err.c: free err_data string if necessary;
4919 don't write to the wrong index in ERR_set_error_data.
4922 *) Implement ssl23_peek (analogous to ssl23_read), which previously
4926 *) Replace rdtsc with _emit statements for VC++ version 5.
4927 [Jeremy Cooper <jeremy@baymoo.org>]
4929 *) Make it possible to reuse SSLv2 sessions.
4932 *) In copy_email() check for >= 0 as a return value for
4933 X509_NAME_get_index_by_NID() since 0 is a valid index.
4934 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
4936 *) Avoid coredump with unsupported or invalid public keys by checking if
4937 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
4938 PKCS7_verify() fails with non detached data.
4941 *) Don't use getenv in library functions when run as setuid/setgid.
4942 New function OPENSSL_issetugid().
4945 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
4946 due to incorrect handling of multi-threading:
4948 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
4950 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
4952 3. Count how many times MemCheck_off() has been called so that
4953 nested use can be treated correctly. This also avoids
4954 inband-signalling in the previous code (which relied on the
4955 assumption that thread ID 0 is impossible).
4958 *) Add "-rand" option also to s_client and s_server.
4961 *) Fix CPU detection on Irix 6.x.
4962 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
4963 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4965 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
4968 [This change does not apply to 0.9.7.]
4970 *) Use the cached encoding of an X509_NAME structure rather than
4971 copying it. This is apparently the reason for the libsafe "errors"
4972 but the code is actually correct.
4975 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
4976 Bleichenbacher's DSA attack.
4977 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
4978 to be set and top=0 forces the highest bit to be set; top=-1 is new
4979 and leaves the highest bit random.
4980 [Ulf Moeller, Bodo Moeller]
4982 *) In the NCONF_...-based implementations for CONF_... queries
4983 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
4984 a temporary CONF structure with the data component set to NULL
4985 (which gives segmentation faults in lh_retrieve).
4986 Instead, use NULL for the CONF pointer in CONF_get_string and
4987 CONF_get_number (which may use environment variables) and directly
4988 return NULL from CONF_get_section.
4991 *) Fix potential buffer overrun for EBCDIC.
4994 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
4995 keyUsage if basicConstraints absent for a CA.
4998 *) Make SMIME_write_PKCS7() write mail header values with a format that
4999 is more generally accepted (no spaces before the semicolon), since
5000 some programs can't parse those values properly otherwise. Also make
5001 sure BIO's that break lines after each write do not create invalid
5005 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
5006 macros previously used would not encode an empty SEQUENCE OF
5007 and break the signature.
5009 [This change does not apply to 0.9.7.]
5011 *) Zero the premaster secret after deriving the master secret in
5015 *) Add some EVP_add_digest_alias registrations (as found in
5016 OpenSSL_add_all_digests()) to SSL_library_init()
5017 aka OpenSSL_add_ssl_algorithms(). This provides improved
5018 compatibility with peers using X.509 certificates
5019 with unconventional AlgorithmIdentifier OIDs.
5022 *) Fix for Irix with NO_ASM.
5023 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
5025 *) ./config script fixes.
5026 [Ulf Moeller, Richard Levitte]
5028 *) Fix 'openssl passwd -1'.
5031 *) Change PKCS12_key_gen_asc() so it can cope with non null
5032 terminated strings whose length is passed in the passlen
5033 parameter, for example from PEM callbacks. This was done
5034 by adding an extra length parameter to asc2uni().
5035 [Steve Henson, reported by <oddissey@samsung.co.kr>]
5037 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
5038 call failed, free the DSA structure.
5041 *) Fix to uni2asc() to cope with zero length Unicode strings.
5042 These are present in some PKCS#12 files.
5045 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
5046 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
5047 when writing a 32767 byte record.
5048 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
5050 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
5051 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
5053 (RSA objects have a reference count access to which is protected
5054 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
5055 so they are meant to be shared between threads.)
5056 [Bodo Moeller, Geoff Thorpe; original patch submitted by
5057 "Reddie, Steven" <Steven.Reddie@ca.com>]
5059 *) Fix a deadlock in CRYPTO_mem_leaks().
5062 *) Use better test patterns in bntest.
5065 *) rand_win.c fix for Borland C.
5068 *) BN_rshift bugfix for n == 0.
5071 *) Add a 'bctest' script that checks for some known 'bc' bugs
5072 so that 'make test' does not abort just because 'bc' is broken.
5075 *) Store verify_result within SSL_SESSION also for client side to
5076 avoid potential security hole. (Re-used sessions on the client side
5077 always resulted in verify_result==X509_V_OK, not using the original
5078 result of the server certificate verification.)
5081 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
5082 SSL3_RT_APPLICATION_DATA, return 0.
5083 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
5087 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
5088 releases, have been re-implemented by renaming the previous
5089 implementations of ssl2_read and ssl3_read to ssl2_read_internal
5090 and ssl3_read_internal, respectively, and adding 'peek' parameters
5091 to them. The new ssl[23]_{read,peek} functions are calls to
5092 ssl[23]_read_internal with the 'peek' flag set appropriately.
5093 A 'peek' parameter has also been added to ssl3_read_bytes, which
5094 does the actual work for ssl3_read_internal.
5097 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
5098 the method-specific "init()" handler. Also clean up ex_data after
5099 calling the method-specific "finish()" handler. Previously, this was
5100 happening the other way round.
5103 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
5104 The previous value, 12, was not always sufficient for BN_mod_exp().
5107 *) Make sure that shared libraries get the internal name engine with
5108 the full version number and not just 0. This should mark the
5109 shared libraries as not backward compatible. Of course, this should
5110 be changed again when we can guarantee backward binary compatibility.
5113 *) Fix typo in get_cert_by_subject() in by_dir.c
5114 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
5116 *) Rework the system to generate shared libraries:
5118 - Make note of the expected extension for the shared libraries and
5119 if there is a need for symbolic links from for example libcrypto.so.0
5120 to libcrypto.so.0.9.7. There is extended info in Configure for
5123 - Make as few rebuilds of the shared libraries as possible.
5125 - Still avoid linking the OpenSSL programs with the shared libraries.
5127 - When installing, install the shared libraries separately from the
5131 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
5133 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
5134 and not in SSL_clear because the latter is also used by the
5135 accept/connect functions; previously, the settings made by
5136 SSL_set_read_ahead would be lost during the handshake.
5137 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
5139 *) Correct util/mkdef.pl to be selective about disabled algorithms.
5140 Previously, it would create entries for disableed algorithms no
5144 *) Added several new manual pages for SSL_* function.
5147 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
5149 *) In ssl23_get_client_hello, generate an error message when faced
5150 with an initial SSL 3.0/TLS record that is too small to contain the
5151 first two bytes of the ClientHello message, i.e. client_version.
5152 (Note that this is a pathologic case that probably has never happened
5153 in real life.) The previous approach was to use the version number
5154 from the record header as a substitute; but our protocol choice
5155 should not depend on that one because it is not authenticated
5156 by the Finished messages.
5159 *) More robust randomness gathering functions for Windows.
5160 [Jeffrey Altman <jaltman@columbia.edu>]
5162 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
5163 not set then we don't setup the error code for issuer check errors
5164 to avoid possibly overwriting other errors which the callback does
5165 handle. If an application does set the flag then we assume it knows
5166 what it is doing and can handle the new informational codes
5170 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
5171 a general "ANY" type, as such it should be able to decode anything
5172 including tagged types. However it didn't check the class so it would
5173 wrongly interpret tagged types in the same way as their universal
5174 counterpart and unknown types were just rejected. Changed so that the
5175 tagged and unknown types are handled in the same way as a SEQUENCE:
5176 that is the encoding is stored intact. There is also a new type
5177 "V_ASN1_OTHER" which is used when the class is not universal, in this
5178 case we have no idea what the actual type is so we just lump them all
5182 *) On VMS, stdout may very well lead to a file that is written to
5183 in a record-oriented fashion. That means that every write() will
5184 write a separate record, which will be read separately by the
5185 programs trying to read from it. This can be very confusing.
5187 The solution is to put a BIO filter in the way that will buffer
5188 text until a linefeed is reached, and then write everything a
5189 line at a time, so every record written will be an actual line,
5190 not chunks of lines and not (usually doesn't happen, but I've
5191 seen it once) several lines in one record. BIO_f_linebuffer() is
5194 Currently, it's a VMS-only method, because that's where it has
5195 been tested well enough.
5198 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
5199 it can return incorrect results.
5200 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
5201 but it was in 0.9.6-beta[12].)
5204 *) Disable the check for content being present when verifying detached
5205 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
5206 include zero length content when signing messages.
5209 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
5210 BIO_ctrl (for BIO pairs).
5213 *) Add DSO method for VMS.
5216 *) Bug fix: Montgomery multiplication could produce results with the
5220 *) Add RPM specification openssl.spec and modify it to build three
5221 packages. The default package contains applications, application
5222 documentation and run-time libraries. The devel package contains
5223 include files, static libraries and function documentation. The
5224 doc package contains the contents of the doc directory. The original
5225 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
5228 *) Add a large number of documentation files for many SSL routines.
5229 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5231 *) Add a configuration entry for Sony News 4.
5232 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
5234 *) Don't set the two most significant bits to one when generating a
5235 random number < q in the DSA library.
5238 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
5239 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
5240 the underlying transport is blocking) if a handshake took place.
5241 (The default behaviour is needed by applications such as s_client
5242 and s_server that use select() to determine when to use SSL_read;
5243 but for applications that know in advance when to expect data, it
5244 just makes things more complicated.)
5247 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
5251 *) Add a few more EBCDIC conditionals that make `req' and `x509'
5252 work better on such systems.
5253 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5255 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
5256 Update PKCS12_parse() so it copies the friendlyName and the
5257 keyid to the certificates aux info.
5260 *) Fix bug in PKCS7_verify() which caused an infinite loop
5261 if there was more than one signature.
5262 [Sven Uszpelkat <su@celocom.de>]
5264 *) Major change in util/mkdef.pl to include extra information
5265 about each symbol, as well as presentig variables as well
5266 as functions. This change means that there's n more need
5267 to rebuild the .num files when some algorithms are excluded.
5270 *) Allow the verify time to be set by an application,
5271 rather than always using the current time.
5274 *) Phase 2 verify code reorganisation. The certificate
5275 verify code now looks up an issuer certificate by a
5276 number of criteria: subject name, authority key id
5277 and key usage. It also verifies self signed certificates
5278 by the same criteria. The main comparison function is
5279 X509_check_issued() which performs these checks.
5281 Lot of changes were necessary in order to support this
5282 without completely rewriting the lookup code.
5284 Authority and subject key identifier are now cached.
5286 The LHASH 'certs' is X509_STORE has now been replaced
5287 by a STACK_OF(X509_OBJECT). This is mainly because an
5288 LHASH can't store or retrieve multiple objects with
5289 the same hash value.
5291 As a result various functions (which were all internal
5292 use only) have changed to handle the new X509_STORE
5293 structure. This will break anything that messed round
5294 with X509_STORE internally.
5296 The functions X509_STORE_add_cert() now checks for an
5297 exact match, rather than just subject name.
5299 The X509_STORE API doesn't directly support the retrieval
5300 of multiple certificates matching a given criteria, however
5301 this can be worked round by performing a lookup first
5302 (which will fill the cache with candidate certificates)
5303 and then examining the cache for matches. This is probably
5304 the best we can do without throwing out X509_LOOKUP
5305 entirely (maybe later...).
5307 The X509_VERIFY_CTX structure has been enhanced considerably.
5309 All certificate lookup operations now go via a get_issuer()
5310 callback. Although this currently uses an X509_STORE it
5311 can be replaced by custom lookups. This is a simple way
5312 to bypass the X509_STORE hackery necessary to make this
5313 work and makes it possible to use more efficient techniques
5314 in future. A very simple version which uses a simple
5315 STACK for its trusted certificate store is also provided
5316 using X509_STORE_CTX_trusted_stack().
5318 The verify_cb() and verify() callbacks now have equivalents
5319 in the X509_STORE_CTX structure.
5321 X509_STORE_CTX also has a 'flags' field which can be used
5322 to customise the verify behaviour.
5325 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
5326 excludes S/MIME capabilities.
5329 *) When a certificate request is read in keep a copy of the
5330 original encoding of the signed data and use it when outputing
5331 again. Signatures then use the original encoding rather than
5332 a decoded, encoded version which may cause problems if the
5333 request is improperly encoded.
5336 *) For consistency with other BIO_puts implementations, call
5337 buffer_write(b, ...) directly in buffer_puts instead of calling
5340 In BIO_puts, increment b->num_write as in BIO_write.
5341 [Peter.Sylvester@EdelWeb.fr]
5343 *) Fix BN_mul_word for the case where the word is 0. (We have to use
5344 BN_zero, we may not return a BIGNUM with an array consisting of
5348 *) Avoid calling abort() from within the library when problems are
5349 detected, except if preprocessor symbols have been defined
5350 (such as REF_CHECK, BN_DEBUG etc.).
5353 *) New openssl application 'rsautl'. This utility can be
5354 used for low level RSA operations. DER public key
5355 BIO/fp routines also added.
5358 *) New Configure entry and patches for compiling on QNX 4.
5359 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
5361 *) A demo state-machine implementation was sponsored by
5362 Nuron (http://www.nuron.com/) and is now available in
5363 demos/state_machine.
5366 *) New options added to the 'dgst' utility for signature
5367 generation and verification.
5370 *) Unrecognized PKCS#7 content types are now handled via a
5371 catch all ASN1_TYPE structure. This allows unsupported
5372 types to be stored as a "blob" and an application can
5373 encode and decode it manually.
5376 *) Fix various signed/unsigned issues to make a_strex.c
5378 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
5380 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
5381 length if passed a buffer. ASN1_INTEGER_to_BN failed
5382 if passed a NULL BN and its argument was negative.
5383 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
5385 *) Modification to PKCS#7 encoding routines to output definite
5386 length encoding. Since currently the whole structures are in
5387 memory there's not real point in using indefinite length
5388 constructed encoding. However if OpenSSL is compiled with
5389 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
5392 *) Added BIO_vprintf() and BIO_vsnprintf().
5395 *) Added more prefixes to parse for in the the strings written
5396 through a logging bio, to cover all the levels that are available
5397 through syslog. The prefixes are now:
5399 PANIC, EMERG, EMR => LOG_EMERG
5400 ALERT, ALR => LOG_ALERT
5401 CRIT, CRI => LOG_CRIT
5402 ERROR, ERR => LOG_ERR
5403 WARNING, WARN, WAR => LOG_WARNING
5404 NOTICE, NOTE, NOT => LOG_NOTICE
5405 INFO, INF => LOG_INFO
5406 DEBUG, DBG => LOG_DEBUG
5408 and as before, if none of those prefixes are present at the
5409 beginning of the string, LOG_ERR is chosen.
5411 On Win32, the LOG_* levels are mapped according to this:
5413 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
5414 LOG_WARNING => EVENTLOG_WARNING_TYPE
5415 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
5419 *) Made it possible to reconfigure with just the configuration
5420 argument "reconf" or "reconfigure". The command line arguments
5421 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
5422 and are retrieved from there when reconfiguring.
5426 [Assar Westerlund <assar@sics.se>, Richard Levitte]
5428 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
5431 *) The obj_dat.pl script was messing up the sorting of object
5432 names. The reason was that it compared the quoted version
5433 of strings as a result "OCSP" > "OCSP Signing" because
5434 " > SPACE. Changed script to store unquoted versions of
5435 names and add quotes on output. It was also omitting some
5436 names from the lookup table if they were given a default
5437 value (that is if SN is missing it is given the same
5438 value as LN and vice versa), these are now added on the
5439 grounds that if an object has a name we should be able to
5440 look it up. Finally added warning output when duplicate
5441 short or long names are found.
5444 *) Changes needed for Tandem NSK.
5445 [Scott Uroff <scott@xypro.com>]
5447 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
5448 RSA_padding_check_SSLv23(), special padding was never detected
5449 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
5450 version rollback attacks was not effective.
5452 In s23_clnt.c, don't use special rollback-attack detection padding
5453 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
5454 client; similarly, in s23_srvr.c, don't do the rollback check if
5455 SSL 2.0 is the only protocol enabled in the server.
5458 *) Make it possible to get hexdumps of unprintable data with 'openssl
5459 asn1parse'. By implication, the functions ASN1_parse_dump() and
5460 BIO_dump_indent() are added.
5463 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
5464 these print out strings and name structures based on various
5465 flags including RFC2253 support and proper handling of
5466 multibyte characters. Added options to the 'x509' utility
5467 to allow the various flags to be set.
5470 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
5471 Also change the functions X509_cmp_current_time() and
5472 X509_gmtime_adj() work with an ASN1_TIME structure,
5473 this will enable certificates using GeneralizedTime in validity
5474 dates to be checked.
5477 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
5478 negative public key encodings) on by default,
5479 NO_NEG_PUBKEY_BUG can be set to disable it.
5482 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
5483 content octets. An i2c_ASN1_OBJECT is unnecessary because
5484 the encoding can be trivially obtained from the structure.
5487 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
5488 not read locks (CRYPTO_r_[un]lock).
5491 *) A first attempt at creating official support for shared
5492 libraries through configuration. I've kept it so the
5493 default is static libraries only, and the OpenSSL programs
5494 are always statically linked for now, but there are
5495 preparations for dynamic linking in place.
5496 This has been tested on Linux and Tru64.
5499 *) Randomness polling function for Win9x, as described in:
5500 Peter Gutmann, Software Generation of Practically Strong
5504 *) Fix so PRNG is seeded in req if using an already existing
5508 *) New options to smime application. -inform and -outform
5509 allow alternative formats for the S/MIME message including
5510 PEM and DER. The -content option allows the content to be
5511 specified separately. This should allow things like Netscape
5512 form signing output easier to verify.
5515 *) Fix the ASN1 encoding of tags using the 'long form'.
5518 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
5519 STRING types. These convert content octets to and from the
5520 underlying type. The actual tag and length octets are
5521 already assumed to have been read in and checked. These
5522 are needed because all other string types have virtually
5523 identical handling apart from the tag. By having versions
5524 of the ASN1 functions that just operate on content octets
5525 IMPLICIT tagging can be handled properly. It also allows
5526 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
5527 and ASN1_INTEGER are identical apart from the tag.
5530 *) Change the handling of OID objects as follows:
5532 - New object identifiers are inserted in objects.txt, following
5533 the syntax given in objects.README.
5534 - objects.pl is used to process obj_mac.num and create a new
5536 - obj_dat.pl is used to create a new obj_dat.h, using the data in
5539 This is currently kind of a hack, and the perl code in objects.pl
5540 isn't very elegant, but it works as I intended. The simplest way
5541 to check that it worked correctly is to look in obj_dat.h and
5542 check the array nid_objs and make sure the objects haven't moved
5543 around (this is important!). Additions are OK, as well as
5544 consistent name changes.
5547 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
5550 *) Addition of the command line parameter '-rand file' to 'openssl req'.
5551 The given file adds to whatever has already been seeded into the
5552 random pool through the RANDFILE configuration file option or
5553 environment variable, or the default random state file.
5556 *) mkstack.pl now sorts each macro group into lexical order.
5557 Previously the output order depended on the order the files
5558 appeared in the directory, resulting in needless rewriting
5562 *) Patches to make OpenSSL compile under Win32 again. Mostly
5563 work arounds for the VC++ problem that it treats func() as
5564 func(void). Also stripped out the parts of mkdef.pl that
5565 added extra typesafe functions: these no longer exist.
5568 *) Reorganisation of the stack code. The macros are now all
5569 collected in safestack.h . Each macro is defined in terms of
5570 a "stack macro" of the form SKM_<name>(type, a, b). The
5571 DEBUG_SAFESTACK is now handled in terms of function casts,
5572 this has the advantage of retaining type safety without the
5573 use of additional functions. If DEBUG_SAFESTACK is not defined
5574 then the non typesafe macros are used instead. Also modified the
5575 mkstack.pl script to handle the new form. Needs testing to see
5576 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
5577 the default if no major problems. Similar behaviour for ASN1_SET_OF
5578 and PKCS12_STACK_OF.
5581 *) When some versions of IIS use the 'NET' form of private key the
5582 key derivation algorithm is different. Normally MD5(password) is
5583 used as a 128 bit RC4 key. In the modified case
5584 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
5585 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
5586 as the old Netscape_RSA functions except they have an additional
5587 'sgckey' parameter which uses the modified algorithm. Also added
5588 an -sgckey command line option to the rsa utility. Thanks to
5589 Adrian Peck <bertie@ncipher.com> for posting details of the modified
5590 algorithm to openssl-dev.
5593 *) The evp_local.h macros were using 'c.##kname' which resulted in
5594 invalid expansion on some systems (SCO 5.0.5 for example).
5595 Corrected to 'c.kname'.
5596 [Phillip Porch <root@theporch.com>]
5598 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
5599 a STACK of email addresses from a certificate or request, these look
5600 in the subject name and the subject alternative name extensions and
5601 omit any duplicate addresses.
5604 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
5605 This makes DSA verification about 2 % faster.
5608 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
5609 (meaning that now 2^5 values will be precomputed, which is only 4 KB
5610 plus overhead for 1024 bit moduli).
5611 This makes exponentiations about 0.5 % faster for 1024 bit
5612 exponents (as measured by "openssl speed rsa2048").
5615 *) Rename memory handling macros to avoid conflicts with other
5617 Malloc => OPENSSL_malloc
5618 Malloc_locked => OPENSSL_malloc_locked
5619 Realloc => OPENSSL_realloc
5620 Free => OPENSSL_free
5623 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
5624 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
5627 *) CygWin32 support.
5628 [John Jarvie <jjarvie@newsguy.com>]
5630 *) The type-safe stack code has been rejigged. It is now only compiled
5631 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
5632 by default all type-specific stack functions are "#define"d back to
5633 standard stack functions. This results in more streamlined output
5634 but retains the type-safety checking possibilities of the original
5638 *) The STACK code has been cleaned up, and certain type declarations
5639 that didn't make a lot of sense have been brought in line. This has
5640 also involved a cleanup of sorts in safestack.h to more correctly
5641 map type-safe stack functions onto their plain stack counterparts.
5642 This work has also resulted in a variety of "const"ifications of
5643 lots of the code, especially "_cmp" operations which should normally
5644 be prototyped with "const" parameters anyway.
5647 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
5648 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
5649 (The PRNG state consists of two parts, the large pool 'state' and 'md',
5650 where all of 'md' is used each time the PRNG is used, but 'state'
5651 is used only indexed by a cyclic counter. As entropy may not be
5652 well distributed from the beginning, 'md' is important as a
5653 chaining variable. However, the output function chains only half
5654 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
5655 all of 'md', and seeding with STATE_SIZE dummy bytes will result
5656 in all of 'state' being rewritten, with the new values depending
5657 on virtually all of 'md'. This overcomes the 80 bit limitation.)
5660 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
5661 the handshake is continued after ssl_verify_cert_chain();
5662 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
5663 can lead to 'unexplainable' connection aborts later.
5664 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
5666 *) Major EVP API cipher revision.
5667 Add hooks for extra EVP features. This allows various cipher
5668 parameters to be set in the EVP interface. Support added for variable
5669 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
5670 setting of RC2 and RC5 parameters.
5672 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
5675 Remove lots of duplicated code from the EVP library. For example *every*
5676 cipher init() function handles the 'iv' in the same way according to the
5677 cipher mode. They also all do nothing if the 'key' parameter is NULL and
5678 for CFB and OFB modes they zero ctx->num.
5680 New functionality allows removal of S/MIME code RC2 hack.
5682 Most of the routines have the same form and so can be declared in terms
5685 By shifting this to the top level EVP_CipherInit() it can be removed from
5686 all individual ciphers. If the cipher wants to handle IVs or keys
5687 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
5690 Change lots of functions like EVP_EncryptUpdate() to now return a
5691 value: although software versions of the algorithms cannot fail
5692 any installed hardware versions can.
5695 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
5696 this option is set, tolerate broken clients that send the negotiated
5697 protocol version number instead of the requested protocol version
5701 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
5702 i.e. non-zero for export ciphersuites, zero otherwise.
5703 Previous versions had this flag inverted, inconsistent with
5704 rsa_tmp_cb (..._TMP_RSA_CB).
5705 [Bodo Moeller; problem reported by Amit Chopra]
5707 *) Add missing DSA library text string. Work around for some IIS
5708 key files with invalid SEQUENCE encoding.
5711 *) Add a document (doc/standards.txt) that list all kinds of standards
5712 and so on that are implemented in OpenSSL.
5715 *) Enhance c_rehash script. Old version would mishandle certificates
5716 with the same subject name hash and wouldn't handle CRLs at all.
5717 Added -fingerprint option to crl utility, to support new c_rehash
5721 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
5724 *) Fix for SSL server purpose checking. Server checking was
5725 rejecting certificates which had extended key usage present
5726 but no ssl client purpose.
5727 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
5729 *) Make PKCS#12 code work with no password. The PKCS#12 spec
5730 is a little unclear about how a blank password is handled.
5731 Since the password in encoded as a BMPString with terminating
5732 double NULL a zero length password would end up as just the
5733 double NULL. However no password at all is different and is
5734 handled differently in the PKCS#12 key generation code. NS
5735 treats a blank password as zero length. MSIE treats it as no
5736 password on export: but it will try both on import. We now do
5737 the same: PKCS12_parse() tries zero length and no password if
5738 the password is set to "" or NULL (NULL is now a valid password:
5739 it wasn't before) as does the pkcs12 application.
5742 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
5743 perror when PEM_read_bio_X509_REQ fails, the error message must
5744 be obtained from the error queue.
5747 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
5748 it in ERR_remove_state if appropriate, and change ERR_get_state
5749 accordingly to avoid race conditions (this is necessary because
5750 thread_hash is no longer constant once set).
5753 *) Bugfix for linux-elf makefile.one.
5756 *) RSA_get_default_method() will now cause a default
5757 RSA_METHOD to be chosen if one doesn't exist already.
5758 Previously this was only set during a call to RSA_new()
5759 or RSA_new_method(NULL) meaning it was possible for
5760 RSA_get_default_method() to return NULL.
5763 *) Added native name translation to the existing DSO code
5764 that will convert (if the flag to do so is set) filenames
5765 that are sufficiently small and have no path information
5766 into a canonical native form. Eg. "blah" converted to
5767 "libblah.so" or "blah.dll" etc.
5770 *) New function ERR_error_string_n(e, buf, len) which is like
5771 ERR_error_string(e, buf), but writes at most 'len' bytes
5772 including the 0 terminator. For ERR_error_string_n, 'buf'
5774 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
5776 *) CONF library reworked to become more general. A new CONF
5777 configuration file reader "class" is implemented as well as a
5778 new functions (NCONF_*, for "New CONF") to handle it. The now
5779 old CONF_* functions are still there, but are reimplemented to
5780 work in terms of the new functions. Also, a set of functions
5781 to handle the internal storage of the configuration data is
5782 provided to make it easier to write new configuration file
5783 reader "classes" (I can definitely see something reading a
5784 configuration file in XML format, for example), called _CONF_*,
5785 or "the configuration storage API"...
5787 The new configuration file reading functions are:
5789 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
5790 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
5792 NCONF_default, NCONF_WIN32
5794 NCONF_dump_fp, NCONF_dump_bio
5796 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
5797 NCONF_new creates a new CONF object. This works in the same way
5798 as other interfaces in OpenSSL, like the BIO interface.
5799 NCONF_dump_* dump the internal storage of the configuration file,
5800 which is useful for debugging. All other functions take the same
5801 arguments as the old CONF_* functions wth the exception of the
5802 first that must be a `CONF *' instead of a `LHASH *'.
5804 To make it easer to use the new classes with the old CONF_* functions,
5805 the function CONF_set_default_method is provided.
5808 *) Add '-tls1' option to 'openssl ciphers', which was already
5809 mentioned in the documentation but had not been implemented.
5810 (This option is not yet really useful because even the additional
5811 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
5814 *) Initial DSO code added into libcrypto for letting OpenSSL (and
5815 OpenSSL-based applications) load shared libraries and bind to
5816 them in a portable way.
5817 [Geoff Thorpe, with contributions from Richard Levitte]
5819 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
5821 *) Make sure _lrotl and _lrotr are only used with MSVC.
5823 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
5824 (the default implementation of RAND_status).
5826 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
5827 to '-clrext' (= clear extensions), as intended and documented.
5828 [Bodo Moeller; inconsistency pointed out by Michael Attili
5831 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
5832 was larger than the MD block size.
5833 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
5835 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
5836 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
5837 using the passed key: if the passed key was a private key the result
5838 of X509_print(), for example, would be to print out all the private key
5842 *) des_quad_cksum() byte order bug fix.
5843 [Ulf Möller, using the problem description in krb4-0.9.7, where
5844 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
5846 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
5848 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
5850 *) For easily testing in shell scripts whether some command
5851 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
5852 returns with exit code 0 iff no command of the given name is available.
5853 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
5854 the output goes to stdout and nothing is printed to stderr.
5855 Additional arguments are always ignored.
5857 Since for each cipher there is a command of the same name,
5858 the 'no-cipher' compilation switches can be tested this way.
5860 ('openssl no-XXX' is not able to detect pseudo-commands such
5861 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
5864 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
5867 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
5868 is set; it will be thrown away anyway because each handshake creates
5870 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
5871 to parameters -- in previous versions (since OpenSSL 0.9.3) the
5872 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
5873 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
5876 *) New s_client option -ign_eof: EOF at stdin is ignored, and
5877 'Q' and 'R' lose their special meanings (quit/renegotiate).
5878 This is part of what -quiet does; unlike -quiet, -ign_eof
5879 does not suppress any output.
5882 *) Add compatibility options to the purpose and trust code. The
5883 purpose X509_PURPOSE_ANY is "any purpose" which automatically
5884 accepts a certificate or CA, this was the previous behaviour,
5885 with all the associated security issues.
5887 X509_TRUST_COMPAT is the old trust behaviour: only and
5888 automatically trust self signed roots in certificate store. A
5889 new trust setting X509_TRUST_DEFAULT is used to specify that
5890 a purpose has no associated trust setting and it should instead
5891 use the value in the default purpose.
5894 *) Fix the PKCS#8 DSA private key code so it decodes keys again
5895 and fix a memory leak.
5898 *) In util/mkerr.pl (which implements 'make errors'), preserve
5899 reason strings from the previous version of the .c file, as
5900 the default to have only downcase letters (and digits) in
5901 automatically generated reasons codes is not always appropriate.
5904 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
5905 using strerror. Previously, ERR_reason_error_string() returned
5906 library names as reason strings for SYSerr; but SYSerr is a special
5907 case where small numbers are errno values, not library numbers.
5910 *) Add '-dsaparam' option to 'openssl dhparam' application. This
5911 converts DSA parameters into DH parameters. (When creating parameters,
5912 DSA_generate_parameters is used.)
5915 *) Include 'length' (recommended exponent length) in C code generated
5916 by 'openssl dhparam -C'.
5919 *) The second argument to set_label in perlasm was already being used
5920 so couldn't be used as a "file scope" flag. Moved to third argument
5924 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
5925 instead of RAND_bytes for encryption IVs and salts.
5928 *) Include RAND_status() into RAND_METHOD instead of implementing
5929 it only for md_rand.c Otherwise replacing the PRNG by calling
5930 RAND_set_rand_method would be impossible.
5933 *) Don't let DSA_generate_key() enter an infinite loop if the random
5934 number generation fails.
5937 *) New 'rand' application for creating pseudo-random output.
5940 *) Added configuration support for Linux/IA64
5941 [Rolf Haberrecker <rolf@suse.de>]
5943 *) Assembler module support for Mingw32.
5946 *) Shared library support for HPUX (in shlib/).
5947 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
5949 *) Shared library support for Solaris gcc.
5950 [Lutz Behnke <behnke@trustcenter.de>]
5952 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
5954 *) PKCS7_encrypt() was adding text MIME headers twice because they
5955 were added manually and by SMIME_crlf_copy().
5958 *) In bntest.c don't call BN_rand with zero bits argument.
5959 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
5961 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
5962 case was implemented. This caused BN_div_recp() to fail occasionally.
5965 *) Add an optional second argument to the set_label() in the perl
5966 assembly language builder. If this argument exists and is set
5967 to 1 it signals that the assembler should use a symbol whose
5968 scope is the entire file, not just the current function. This
5969 is needed with MASM which uses the format label:: for this scope.
5970 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
5972 *) Change the ASN1 types so they are typedefs by default. Before
5973 almost all types were #define'd to ASN1_STRING which was causing
5974 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
5978 *) Change names of new functions to the new get1/get0 naming
5979 convention: After 'get1', the caller owns a reference count
5980 and has to call ..._free; 'get0' returns a pointer to some
5981 data structure without incrementing reference counters.
5982 (Some of the existing 'get' functions increment a reference
5983 counter, some don't.)
5984 Similarly, 'set1' and 'add1' functions increase reference
5985 counters or duplicate objects.
5988 *) Allow for the possibility of temp RSA key generation failure:
5989 the code used to assume it always worked and crashed on failure.
5992 *) Fix potential buffer overrun problem in BIO_printf().
5993 [Ulf Möller, using public domain code by Patrick Powell; problem
5994 pointed out by David Sacerdote <das33@cornell.edu>]
5996 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
5997 RAND_egd() and RAND_status(). In the command line application,
5998 the EGD socket can be specified like a seed file using RANDFILE
6002 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
6003 Some CAs (e.g. Verisign) distribute certificates in this form.
6006 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
6007 list to exclude them. This means that no special compilation option
6008 is needed to use anonymous DH: it just needs to be included in the
6012 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
6013 EVP_MD_type. The old functionality is available in a new macro called
6014 EVP_MD_md(). Change code that uses it and update docs.
6017 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
6018 where the 'void *' argument is replaced by a function pointer argument.
6019 Previously 'void *' was abused to point to functions, which works on
6020 many platforms, but is not correct. As these functions are usually
6021 called by macros defined in OpenSSL header files, most source code
6022 should work without changes.
6025 *) <openssl/opensslconf.h> (which is created by Configure) now contains
6026 sections with information on -D... compiler switches used for
6027 compiling the library so that applications can see them. To enable
6028 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
6029 must be defined. E.g.,
6030 #define OPENSSL_ALGORITHM_DEFINES
6031 #include <openssl/opensslconf.h>
6032 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
6033 [Richard Levitte, Ulf and Bodo Möller]
6035 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
6039 *) Change the 'other' type in certificate aux info to a STACK_OF
6040 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
6041 the required ASN1 format: arbitrary types determined by an OID.
6044 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
6045 argument to 'req'. This is not because the function is newer or
6046 better than others it just uses the work 'NEW' in the certificate
6047 request header lines. Some software needs this.
6050 *) Reorganise password command line arguments: now passwords can be
6051 obtained from various sources. Delete the PEM_cb function and make
6052 it the default behaviour: i.e. if the callback is NULL and the
6053 usrdata argument is not NULL interpret it as a null terminated pass
6054 phrase. If usrdata and the callback are NULL then the pass phrase
6055 is prompted for as usual.
6058 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
6059 the support is automatically enabled. The resulting binaries will
6060 autodetect the card and use it if present.
6061 [Ben Laurie and Compaq Inc.]
6063 *) Work around for Netscape hang bug. This sends certificate request
6064 and server done in one record. Since this is perfectly legal in the
6065 SSL/TLS protocol it isn't a "bug" option and is on by default. See
6066 the bugs/SSLv3 entry for more info.
6069 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
6072 *) Add -rand argument to smime and pkcs12 applications and read/write
6076 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
6079 *) Add command line password options to the remaining applications.
6082 *) Bug fix for BN_div_recp() for numerators with an even number of
6086 *) More tests in bntest.c, and changed test_bn output.
6089 *) ./config recognizes MacOS X now.
6092 *) Bug fix for BN_div() when the first words of num and divsor are
6093 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
6096 *) Add support for various broken PKCS#8 formats, and command line
6097 options to produce them.
6100 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
6101 get temporary BIGNUMs from a BN_CTX.
6104 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
6108 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
6109 include a #define from the old name to the new. The original intent
6110 was that statically linked binaries could for example just call
6111 SSLeay_add_all_ciphers() to just add ciphers to the table and not
6112 link with digests. This never worked becayse SSLeay_add_all_digests()
6113 and SSLeay_add_all_ciphers() were in the same source file so calling
6114 one would link with the other. They are now in separate source files.
6117 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
6120 *) Use a less unusual form of the Miller-Rabin primality test (it used
6121 a binary algorithm for exponentiation integrated into the Miller-Rabin
6122 loop, our standard modexp algorithms are faster).
6125 *) Support for the EBCDIC character set completed.
6126 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
6128 *) Source code cleanups: use const where appropriate, eliminate casts,
6129 use void * instead of char * in lhash.
6132 *) Bugfix: ssl3_send_server_key_exchange was not restartable
6133 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
6134 this the server could overwrite ephemeral keys that the client
6138 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
6139 using 50 iterations of the Rabin-Miller test.
6141 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
6142 iterations of the Rabin-Miller test as required by the appendix
6143 to FIPS PUB 186[-1]) instead of DSA_is_prime.
6144 As BN_is_prime_fasttest includes trial division, DSA parameter
6145 generation becomes much faster.
6147 This implies a change for the callback functions in DSA_is_prime
6148 and DSA_generate_parameters: The callback function is called once
6149 for each positive witness in the Rabin-Miller test, not just
6150 occasionally in the inner loop; and the parameters to the
6151 callback function now provide an iteration count for the outer
6152 loop rather than for the current invocation of the inner loop.
6153 DSA_generate_parameters additionally can call the callback
6154 function with an 'iteration count' of -1, meaning that a
6155 candidate has passed the trial division test (when q is generated
6156 from an application-provided seed, trial division is skipped).
6159 *) New function BN_is_prime_fasttest that optionally does trial
6160 division before starting the Rabin-Miller test and has
6161 an additional BN_CTX * argument (whereas BN_is_prime always
6162 has to allocate at least one BN_CTX).
6163 'callback(1, -1, cb_arg)' is called when a number has passed the
6164 trial division stage.
6167 *) Fix for bug in CRL encoding. The validity dates weren't being handled
6171 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
6174 *) New function BN_pseudo_rand().
6177 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
6178 bignum version of BN_from_montgomery() with the working code from
6179 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
6183 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
6184 made it impossible to use the same SSL_SESSION data structure in
6185 SSL2 clients in multiple threads.
6188 *) The return value of RAND_load_file() no longer counts bytes obtained
6189 by stat(). RAND_load_file(..., -1) is new and uses the complete file
6190 to seed the PRNG (previously an explicit byte count was required).
6191 [Ulf Möller, Bodo Möller]
6193 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
6194 used (char *) instead of (void *) and had casts all over the place.
6197 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
6200 *) Retain source code compatibility for BN_prime_checks macro:
6201 BN_is_prime(..., BN_prime_checks, ...) now uses
6202 BN_prime_checks_for_size to determine the appropriate number of
6203 Rabin-Miller iterations.
6206 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
6207 DH_CHECK_P_NOT_SAFE_PRIME.
6208 (Check if this is true? OpenPGP calls them "strong".)
6211 *) Merge the functionality of "dh" and "gendh" programs into a new program
6212 "dhparam". The old programs are retained for now but will handle DH keys
6213 (instead of parameters) in future.
6216 *) Make the ciphers, s_server and s_client programs check the return values
6217 when a new cipher list is set.
6220 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
6221 ciphers. Before when the 56bit ciphers were enabled the sorting was
6224 The syntax for the cipher sorting has been extended to support sorting by
6225 cipher-strength (using the strength_bits hard coded in the tables).
6226 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
6228 Fix a bug in the cipher-command parser: when supplying a cipher command
6229 string with an "undefined" symbol (neither command nor alphanumeric
6230 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
6231 an error is flagged.
6233 Due to the strength-sorting extension, the code of the
6234 ssl_create_cipher_list() function was completely rearranged. I hope that
6235 the readability was also increased :-)
6236 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6238 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
6239 for the first serial number and places 2 in the serial number file. This
6240 avoids problems when the root CA is created with serial number zero and
6241 the first user certificate has the same issuer name and serial number
6245 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
6246 the new code. Add documentation for this stuff.
6249 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
6250 X509_*() to X509at_*() on the grounds that they don't handle X509
6251 structures and behave in an analagous way to the X509v3 functions:
6252 they shouldn't be called directly but wrapper functions should be used
6255 So we also now have some wrapper functions that call the X509at functions
6256 when passed certificate requests. (TO DO: similar things can be done with
6257 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
6258 things. Some of these need some d2i or i2d and print functionality
6259 because they handle more complex structures.)
6262 *) Add missing #ifndefs that caused missing symbols when building libssl
6263 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
6264 NO_RSA in ssl/s2*.c.
6265 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
6267 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
6268 has a return value which indicates the quality of the random data
6269 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
6270 error queue. New function RAND_pseudo_bytes() generates output that is
6271 guaranteed to be unique but not unpredictable. RAND_add is like
6272 RAND_seed, but takes an extra argument for an entropy estimate
6273 (RAND_seed always assumes full entropy).
6276 *) Do more iterations of Rabin-Miller probable prime test (specifically,
6277 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
6278 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
6279 in crypto/bn/bn_prime.c for the complete table). This guarantees a
6280 false-positive rate of at most 2^-80 for random input.
6283 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
6286 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
6287 in the 0.9.5 release), this returns the chain
6288 from an X509_CTX structure with a dup of the stack and all
6289 the X509 reference counts upped: so the stack will exist
6290 after X509_CTX_cleanup() has been called. Modify pkcs12.c
6293 Also make SSL_SESSION_print() print out the verify return
6297 *) Add manpage for the pkcs12 command. Also change the default
6298 behaviour so MAC iteration counts are used unless the new
6299 -nomaciter option is used. This improves file security and
6300 only older versions of MSIE (4.0 for example) need it.
6303 *) Honor the no-xxx Configure options when creating .DEF files.
6306 *) Add PKCS#10 attributes to field table: challengePassword,
6307 unstructuredName and unstructuredAddress. These are taken from
6308 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
6309 international characters are used.
6311 More changes to X509_ATTRIBUTE code: allow the setting of types
6312 based on strings. Remove the 'loc' parameter when adding
6313 attributes because these will be a SET OF encoding which is sorted
6317 *) Initial changes to the 'req' utility to allow request generation
6318 automation. This will allow an application to just generate a template
6319 file containing all the field values and have req construct the
6322 Initial support for X509_ATTRIBUTE handling. Stacks of these are
6323 used all over the place including certificate requests and PKCS#7
6324 structures. They are currently handled manually where necessary with
6325 some primitive wrappers for PKCS#7. The new functions behave in a
6326 manner analogous to the X509 extension functions: they allow
6327 attributes to be looked up by NID and added.
6329 Later something similar to the X509V3 code would be desirable to
6330 automatically handle the encoding, decoding and printing of the
6331 more complex types. The string types like challengePassword can
6332 be handled by the string table functions.
6334 Also modified the multi byte string table handling. Now there is
6335 a 'global mask' which masks out certain types. The table itself
6336 can use the flag STABLE_NO_MASK to ignore the mask setting: this
6337 is useful when for example there is only one permissible type
6338 (as in countryName) and using the mask might result in no valid
6342 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
6343 SSL_get_peer_finished to allow applications to obtain the latest
6344 Finished messages sent to the peer or expected from the peer,
6345 respectively. (SSL_get_peer_finished is usually the Finished message
6346 actually received from the peer, otherwise the protocol will be aborted.)
6348 As the Finished message are message digests of the complete handshake
6349 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
6350 be used for external authentication procedures when the authentication
6351 provided by SSL/TLS is not desired or is not enough.
6354 *) Enhanced support for Alpha Linux is added. Now ./config checks if
6355 the host supports BWX extension and if Compaq C is present on the
6356 $PATH. Just exploiting of the BWX extension results in 20-30%
6357 performance kick for some algorithms, e.g. DES and RC4 to mention
6358 a couple. Compaq C in turn generates ~20% faster code for MD5 and
6362 *) Add support for MS "fast SGC". This is arguably a violation of the
6363 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
6364 weak crypto and after checking the certificate is SGC a second one
6365 with strong crypto. MS SGC stops the first handshake after receiving
6366 the server certificate message and sends a second client hello. Since
6367 a server will typically do all the time consuming operations before
6368 expecting any further messages from the client (server key exchange
6369 is the most expensive) there is little difference between the two.
6371 To get OpenSSL to support MS SGC we have to permit a second client
6372 hello message after we have sent server done. In addition we have to
6373 reset the MAC if we do get this second client hello.
6376 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
6377 if a DER encoded private key is RSA or DSA traditional format. Changed
6378 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
6379 format DER encoded private key. Newer code should use PKCS#8 format which
6380 has the key type encoded in the ASN1 structure. Added DER private key
6381 support to pkcs8 application.
6384 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
6385 ciphersuites has been selected (as required by the SSL 3/TLS 1
6386 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
6387 is set, we interpret this as a request to violate the specification
6388 (the worst that can happen is a handshake failure, and 'correct'
6389 behaviour would result in a handshake failure anyway).
6392 *) In SSL_CTX_add_session, take into account that there might be multiple
6393 SSL_SESSION structures with the same session ID (e.g. when two threads
6394 concurrently obtain them from an external cache).
6395 The internal cache can handle only one SSL_SESSION with a given ID,
6396 so if there's a conflict, we now throw out the old one to achieve
6400 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
6401 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
6402 some routines that use cipher OIDs: some ciphers do not have OIDs
6403 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
6407 *) Simplify the trust setting structure and code. Now we just have
6408 two sequences of OIDs for trusted and rejected settings. These will
6409 typically have values the same as the extended key usage extension
6410 and any application specific purposes.
6412 The trust checking code now has a default behaviour: it will just
6413 check for an object with the same NID as the passed id. Functions can
6414 be provided to override either the default behaviour or the behaviour
6415 for a given id. SSL client, server and email already have functions
6416 in place for compatibility: they check the NID and also return "trusted"
6417 if the certificate is self signed.
6420 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
6421 traditional format into an EVP_PKEY structure.
6424 *) Add a password callback function PEM_cb() which either prompts for
6425 a password if usr_data is NULL or otherwise assumes it is a null
6426 terminated password. Allow passwords to be passed on command line
6427 environment or config files in a few more utilities.
6430 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
6431 keys. Add some short names for PKCS#8 PBE algorithms and allow them
6432 to be specified on the command line for the pkcs8 and pkcs12 utilities.
6433 Update documentation.
6436 *) Support for ASN1 "NULL" type. This could be handled before by using
6437 ASN1_TYPE but there wasn't any function that would try to read a NULL
6438 and produce an error if it couldn't. For compatibility we also have
6439 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
6440 don't allocate anything because they don't need to.
6443 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
6445 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
6447 *) Rebuild of the memory allocation routines used by OpenSSL code and
6448 possibly others as well. The purpose is to make an interface that
6449 provide hooks so anyone can build a separate set of allocation and
6450 deallocation routines to be used by OpenSSL, for example memory
6451 pool implementations, or something else, which was previously hard
6452 since Malloc(), Realloc() and Free() were defined as macros having
6453 the values malloc, realloc and free, respectively (except for Win32
6454 compilations). The same is provided for memory debugging code.
6455 OpenSSL already comes with functionality to find memory leaks, but
6456 this gives people a chance to debug other memory problems.
6458 With these changes, a new set of functions and macros have appeared:
6460 CRYPTO_set_mem_debug_functions() [F]
6461 CRYPTO_get_mem_debug_functions() [F]
6462 CRYPTO_dbg_set_options() [F]
6463 CRYPTO_dbg_get_options() [F]
6464 CRYPTO_malloc_debug_init() [M]
6466 The memory debug functions are NULL by default, unless the library
6467 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
6468 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
6469 gives the standard debugging functions that come with OpenSSL) or
6470 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
6471 provided by the library user) must be used. When the standard
6472 debugging functions are used, CRYPTO_dbg_set_options can be used to
6473 request additional information:
6474 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
6475 the CRYPTO_MDEBUG_xxx macro when compiling the library.
6477 Also, things like CRYPTO_set_mem_functions will always give the
6478 expected result (the new set of functions is used for allocation
6479 and deallocation) at all times, regardless of platform and compiler
6482 To finish it up, some functions that were never use in any other
6483 way than through macros have a new API and new semantic:
6486 CRYPTO_dbg_realloc()
6489 All macros of value have retained their old syntax.
6490 [Richard Levitte and Bodo Moeller]
6492 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
6493 ordering of SMIMECapabilities wasn't in "strength order" and there
6494 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
6498 *) Some ASN1 types with illegal zero length encoding (INTEGER,
6499 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
6500 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
6502 *) Merge in my S/MIME library for OpenSSL. This provides a simple
6503 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
6504 functionality to handle multipart/signed properly) and a utility
6505 called 'smime' to call all this stuff. This is based on code I
6506 originally wrote for Celo who have kindly allowed it to be
6507 included in OpenSSL.
6510 *) Add variants des_set_key_checked and des_set_key_unchecked of
6511 des_set_key (aka des_key_sched). Global variable des_check_key
6512 decides which of these is called by des_set_key; this way
6513 des_check_key behaves as it always did, but applications and
6514 the library itself, which was buggy for des_check_key == 1,
6515 have a cleaner way to pick the version they need.
6518 *) New function PKCS12_newpass() which changes the password of a
6522 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
6523 dynamic mix. In both cases the ids can be used as an index into the
6524 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
6525 functions so they accept a list of the field values and the
6526 application doesn't need to directly manipulate the X509_TRUST
6530 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
6534 *) Modify the way the V3 extension code looks up extensions. This now
6535 works in a similar way to the object code: we have some "standard"
6536 extensions in a static table which is searched with OBJ_bsearch()
6537 and the application can add dynamic ones if needed. The file
6538 crypto/x509v3/ext_dat.h now has the info: this file needs to be
6539 updated whenever a new extension is added to the core code and kept
6540 in ext_nid order. There is a simple program 'tabtest.c' which checks
6541 this. New extensions are not added too often so this file can readily
6542 be maintained manually.
6544 There are two big advantages in doing things this way. The extensions
6545 can be looked up immediately and no longer need to be "added" using
6546 X509V3_add_standard_extensions(): this function now does nothing.
6547 [Side note: I get *lots* of email saying the extension code doesn't
6548 work because people forget to call this function]
6549 Also no dynamic allocation is done unless new extensions are added:
6550 so if we don't add custom extensions there is no need to call
6551 X509V3_EXT_cleanup().
6554 *) Modify enc utility's salting as follows: make salting the default. Add a
6555 magic header, so unsalted files fail gracefully instead of just decrypting
6556 to garbage. This is because not salting is a big security hole, so people
6557 should be discouraged from doing it.
6560 *) Fixes and enhancements to the 'x509' utility. It allowed a message
6561 digest to be passed on the command line but it only used this
6562 parameter when signing a certificate. Modified so all relevant
6563 operations are affected by the digest parameter including the
6564 -fingerprint and -x509toreq options. Also -x509toreq choked if a
6565 DSA key was used because it didn't fix the digest.
6568 *) Initial certificate chain verify code. Currently tests the untrusted
6569 certificates for consistency with the verify purpose (which is set
6570 when the X509_STORE_CTX structure is set up) and checks the pathlength.
6572 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
6573 this is because it will reject chains with invalid extensions whereas
6574 every previous version of OpenSSL and SSLeay made no checks at all.
6576 Trust code: checks the root CA for the relevant trust settings. Trust
6577 settings have an initial value consistent with the verify purpose: e.g.
6578 if the verify purpose is for SSL client use it expects the CA to be
6579 trusted for SSL client use. However the default value can be changed to
6580 permit custom trust settings: one example of this would be to only trust
6581 certificates from a specific "secure" set of CAs.
6583 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
6584 which should be used for version portability: especially since the
6585 verify structure is likely to change more often now.
6587 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
6588 to set them. If not set then assume SSL clients will verify SSL servers
6591 Two new options to the verify program: -untrusted allows a set of
6592 untrusted certificates to be passed in and -purpose which sets the
6593 intended purpose of the certificate. If a purpose is set then the
6594 new chain verify code is used to check extension consistency.
6597 *) Support for the authority information access extension.
6600 *) Modify RSA and DSA PEM read routines to transparently handle
6601 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
6602 public keys in a format compatible with certificate
6603 SubjectPublicKeyInfo structures. Unfortunately there were already
6604 functions called *_PublicKey_* which used various odd formats so
6605 these are retained for compatibility: however the DSA variants were
6606 never in a public release so they have been deleted. Changed dsa/rsa
6607 utilities to handle the new format: note no releases ever handled public
6608 keys so we should be OK.
6610 The primary motivation for this change is to avoid the same fiasco
6611 that dogs private keys: there are several incompatible private key
6612 formats some of which are standard and some OpenSSL specific and
6613 require various evil hacks to allow partial transparent handling and
6614 even then it doesn't work with DER formats. Given the option anything
6615 other than PKCS#8 should be dumped: but the other formats have to
6616 stay in the name of compatibility.
6618 With public keys and the benefit of hindsight one standard format
6619 is used which works with EVP_PKEY, RSA or DSA structures: though
6620 it clearly returns an error if you try to read the wrong kind of key.
6622 Added a -pubkey option to the 'x509' utility to output the public key.
6623 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
6624 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
6625 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
6626 that do the same as the EVP_PKEY_assign_*() except they up the
6627 reference count of the added key (they don't "swallow" the
6631 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
6632 CRLs would fail if the file contained no certificates or no CRLs:
6633 added a new function to read in both types and return the number
6634 read: this means that if none are read it will be an error. The
6635 DER versions of the certificate and CRL reader would always fail
6636 because it isn't possible to mix certificates and CRLs in DER format
6637 without choking one or the other routine. Changed this to just read
6638 a certificate: this is the best we can do. Also modified the code
6639 in apps/verify.c to take notice of return codes: it was previously
6640 attempting to read in certificates from NULL pointers and ignoring
6641 any errors: this is one reason why the cert and CRL reader seemed
6642 to work. It doesn't check return codes from the default certificate
6643 routines: these may well fail if the certificates aren't installed.
6646 *) Code to support otherName option in GeneralName.
6649 *) First update to verify code. Change the verify utility
6650 so it warns if it is passed a self signed certificate:
6651 for consistency with the normal behaviour. X509_verify
6652 has been modified to it will now verify a self signed
6653 certificate if *exactly* the same certificate appears
6654 in the store: it was previously impossible to trust a
6655 single self signed certificate. This means that:
6656 openssl verify ss.pem
6657 now gives a warning about a self signed certificate but
6658 openssl verify -CAfile ss.pem ss.pem
6662 *) For servers, store verify_result in SSL_SESSION data structure
6663 (and add it to external session representation).
6664 This is needed when client certificate verifications fails,
6665 but an application-provided verification callback (set by
6666 SSL_CTX_set_cert_verify_callback) allows accepting the session
6667 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
6668 but returns 1): When the session is reused, we have to set
6669 ssl->verify_result to the appropriate error code to avoid
6671 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
6673 *) Fix a bug in the new PKCS#7 code: it didn't consider the
6674 case in PKCS7_dataInit() where the signed PKCS7 structure
6675 didn't contain any existing data because it was being created.
6676 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
6678 *) Add a salt to the key derivation routines in enc.c. This
6679 forms the first 8 bytes of the encrypted file. Also add a
6680 -S option to allow a salt to be input on the command line.
6683 *) New function X509_cmp(). Oddly enough there wasn't a function
6684 to compare two certificates. We do this by working out the SHA1
6685 hash and comparing that. X509_cmp() will be needed by the trust
6689 *) SSL_get1_session() is like SSL_get_session(), but increments
6690 the reference count in the SSL_SESSION returned.
6691 [Geoff Thorpe <geoff@eu.c2.net>]
6693 *) Fix for 'req': it was adding a null to request attributes.
6694 Also change the X509_LOOKUP and X509_INFO code to handle
6695 certificate auxiliary information.
6698 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
6702 *) Add the possibility to add extra information to the memory leak
6703 detecting output, to form tracebacks, showing from where each
6704 allocation was originated: CRYPTO_push_info("constant string") adds
6705 the string plus current file name and line number to a per-thread
6706 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
6707 is like calling CYRPTO_pop_info() until the stack is empty.
6708 Also updated memory leak detection code to be multi-thread-safe.
6711 *) Add options -text and -noout to pkcs7 utility and delete the
6712 encryption options which never did anything. Update docs.
6715 *) Add options to some of the utilities to allow the pass phrase
6716 to be included on either the command line (not recommended on
6717 OSes like Unix) or read from the environment. Update the
6718 manpages and fix a few bugs.
6721 *) Add a few manpages for some of the openssl commands.
6724 *) Fix the -revoke option in ca. It was freeing up memory twice,
6725 leaking and not finding already revoked certificates.
6728 *) Extensive changes to support certificate auxiliary information.
6729 This involves the use of X509_CERT_AUX structure and X509_AUX
6730 functions. An X509_AUX function such as PEM_read_X509_AUX()
6731 can still read in a certificate file in the usual way but it
6732 will also read in any additional "auxiliary information". By
6733 doing things this way a fair degree of compatibility can be
6734 retained: existing certificates can have this information added
6735 using the new 'x509' options.
6737 Current auxiliary information includes an "alias" and some trust
6738 settings. The trust settings will ultimately be used in enhanced
6739 certificate chain verification routines: currently a certificate
6740 can only be trusted if it is self signed and then it is trusted
6744 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
6745 The problem was that one of the replacement routines had not been working
6746 since SSLeay releases. For now the offending routine has been replaced
6747 with non-optimised assembler. Even so, this now gives around 95%
6748 performance improvement for 1024 bit RSA signs.
6751 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
6752 handling. Most clients have the effective key size in bits equal to
6753 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
6754 A few however don't do this and instead use the size of the decrypted key
6755 to determine the RC2 key length and the AlgorithmIdentifier to determine
6756 the effective key length. In this case the effective key length can still
6757 be 40 bits but the key length can be 168 bits for example. This is fixed
6758 by manually forcing an RC2 key into the EVP_PKEY structure because the
6759 EVP code can't currently handle unusual RC2 key sizes: it always assumes
6760 the key length and effective key length are equal.
6763 *) Add a bunch of functions that should simplify the creation of
6764 X509_NAME structures. Now you should be able to do:
6765 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
6766 and have it automatically work out the correct field type and fill in
6767 the structures. The more adventurous can try:
6768 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
6769 and it will (hopefully) work out the correct multibyte encoding.
6772 *) Change the 'req' utility to use the new field handling and multibyte
6773 copy routines. Before the DN field creation was handled in an ad hoc
6774 way in req, ca, and x509 which was rather broken and didn't support
6775 BMPStrings or UTF8Strings. Since some software doesn't implement
6776 BMPStrings or UTF8Strings yet, they can be enabled using the config file
6777 using the dirstring_type option. See the new comment in the default
6778 openssl.cnf for more info.
6781 *) Make crypto/rand/md_rand.c more robust:
6782 - Assure unique random numbers after fork().
6783 - Make sure that concurrent threads access the global counter and
6784 md serializably so that we never lose entropy in them
6785 or use exactly the same state in multiple threads.
6786 Access to the large state is not always serializable because
6787 the additional locking could be a performance killer, and
6788 md should be large enough anyway.
6791 *) New file apps/app_rand.c with commonly needed functionality
6792 for handling the random seed file.
6794 Use the random seed file in some applications that previously did not:
6796 dsaparam -genkey (which also ignored its '-rand' option),
6799 x509 (when signing).
6800 Except on systems with /dev/urandom, it is crucial to have a random
6801 seed file at least for key creation, DSA signing, and for DH exchanges;
6802 for RSA signatures we could do without one.
6804 gendh and gendsa (unlike genrsa) used to read only the first byte
6805 of each file listed in the '-rand' option. The function as previously
6806 found in genrsa is now in app_rand.c and is used by all programs
6807 that support '-rand'.
6810 *) In RAND_write_file, use mode 0600 for creating files;
6811 don't just chmod when it may be too late.
6814 *) Report an error from X509_STORE_load_locations
6815 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
6818 *) New function ASN1_mbstring_copy() this copies a string in either
6819 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
6820 into an ASN1_STRING type. A mask of permissible types is passed
6821 and it chooses the "minimal" type to use or an error if not type
6825 *) Add function equivalents to the various macros in asn1.h. The old
6826 macros are retained with an M_ prefix. Code inside the library can
6827 use the M_ macros. External code (including the openssl utility)
6828 should *NOT* in order to be "shared library friendly".
6831 *) Add various functions that can check a certificate's extensions
6832 to see if it usable for various purposes such as SSL client,
6833 server or S/MIME and CAs of these types. This is currently
6834 VERY EXPERIMENTAL but will ultimately be used for certificate chain
6835 verification. Also added a -purpose flag to x509 utility to
6836 print out all the purposes.
6839 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
6843 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
6844 for, obtain and decode and extension and obtain its critical flag.
6845 This allows all the necessary extension code to be handled in a
6846 single function call.
6849 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
6850 platforms. See crypto/rc4/rc4_enc.c for further details.
6853 *) New -noout option to asn1parse. This causes no output to be produced
6854 its main use is when combined with -strparse and -out to extract data
6855 from a file (which may not be in ASN.1 format).
6858 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
6859 when producing the local key id.
6860 [Richard Levitte <levitte@stacken.kth.se>]
6862 *) New option -dhparam in s_server. This allows a DH parameter file to be
6863 stated explicitly. If it is not stated then it tries the first server
6864 certificate file. The previous behaviour hard coded the filename
6868 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
6869 a public key to be input or output. For example:
6870 openssl rsa -in key.pem -pubout -out pubkey.pem
6871 Also added necessary DSA public key functions to handle this.
6874 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
6875 in the message. This was handled by allowing
6876 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
6877 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
6879 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
6880 to the end of the strings whereas this didn't. This would cause problems
6881 if strings read with d2i_ASN1_bytes() were later modified.
6882 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
6884 *) Fix for base64 decode bug. When a base64 bio reads only one line of
6885 data and it contains EOF it will end up returning an error. This is
6886 caused by input 46 bytes long. The cause is due to the way base64
6887 BIOs find the start of base64 encoded data. They do this by trying a
6888 trial decode on each line until they find one that works. When they
6889 do a flag is set and it starts again knowing it can pass all the
6890 data directly through the decoder. Unfortunately it doesn't reset
6891 the context it uses. This means that if EOF is reached an attempt
6892 is made to pass two EOFs through the context and this causes the
6893 resulting error. This can also cause other problems as well. As is
6894 usual with these problems it takes *ages* to find and the fix is
6895 trivial: move one line.
6896 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
6898 *) Ugly workaround to get s_client and s_server working under Windows. The
6899 old code wouldn't work because it needed to select() on sockets and the
6900 tty (for keypresses and to see if data could be written). Win32 only
6901 supports select() on sockets so we select() with a 1s timeout on the
6902 sockets and then see if any characters are waiting to be read, if none
6903 are present then we retry, we also assume we can always write data to
6904 the tty. This isn't nice because the code then blocks until we've
6905 received a complete line of data and it is effectively polling the
6906 keyboard at 1s intervals: however it's quite a bit better than not
6907 working at all :-) A dedicated Windows application might handle this
6908 with an event loop for example.
6911 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
6912 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
6913 will be called when RSA_sign() and RSA_verify() are used. This is useful
6914 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
6915 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
6916 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
6917 This necessitated the support of an extra signature type NID_md5_sha1
6918 for SSL signatures and modifications to the SSL library to use it instead
6919 of calling RSA_public_decrypt() and RSA_private_encrypt().
6922 *) Add new -verify -CAfile and -CApath options to the crl program, these
6923 will lookup a CRL issuers certificate and verify the signature in a
6924 similar way to the verify program. Tidy up the crl program so it
6925 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
6926 less strict. It will now permit CRL extensions even if it is not
6927 a V2 CRL: this will allow it to tolerate some broken CRLs.
6930 *) Initialize all non-automatic variables each time one of the openssl
6931 sub-programs is started (this is necessary as they may be started
6932 multiple times from the "OpenSSL>" prompt).
6933 [Lennart Bang, Bodo Moeller]
6935 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
6936 removing all other RSA functionality (this is what NO_RSA does). This
6937 is so (for example) those in the US can disable those operations covered
6938 by the RSA patent while allowing storage and parsing of RSA keys and RSA
6942 *) Non-copying interface to BIO pairs.
6943 (still largely untested)
6946 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
6947 ASCII string. This was handled independently in various places before.
6950 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
6951 UTF8 strings a character at a time.
6954 *) Use client_version from client hello to select the protocol
6955 (s23_srvr.c) and for RSA client key exchange verification
6956 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
6959 *) Add various utility functions to handle SPKACs, these were previously
6960 handled by poking round in the structure internals. Added new function
6961 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
6962 print, verify and generate SPKACs. Based on an original idea from
6963 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
6966 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
6969 *) Allow the config file extension section to be overwritten on the
6970 command line. Based on an original idea from Massimiliano Pala
6971 <madwolf@comune.modena.it>. The new option is called -extensions
6972 and can be applied to ca, req and x509. Also -reqexts to override
6973 the request extensions in req and -crlexts to override the crl extensions
6977 *) Add new feature to the SPKAC handling in ca. Now you can include
6978 the same field multiple times by preceding it by "XXXX." for example:
6981 this is the same syntax as used in the req config file.
6984 *) Allow certificate extensions to be added to certificate requests. These
6985 are specified in a 'req_extensions' option of the req section of the
6986 config file. They can be printed out with the -text option to req but
6987 are otherwise ignored at present.
6990 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
6991 data read consists of only the final block it would not decrypted because
6992 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
6993 A misplaced 'break' also meant the decrypted final block might not be
6994 copied until the next read.
6997 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
6998 a few extra parameters to the DH structure: these will be useful if
6999 for example we want the value of 'q' or implement X9.42 DH.
7002 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
7003 provides hooks that allow the default DSA functions or functions on a
7004 "per key" basis to be replaced. This allows hardware acceleration and
7005 hardware key storage to be handled without major modification to the
7006 library. Also added low level modexp hooks and CRYPTO_EX structure and
7007 associated functions.
7010 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
7011 as "read only": it can't be written to and the buffer it points to will
7012 not be freed. Reading from a read only BIO is much more efficient than
7013 a normal memory BIO. This was added because there are several times when
7014 an area of memory needs to be read from a BIO. The previous method was
7015 to create a memory BIO and write the data to it, this results in two
7016 copies of the data and an O(n^2) reading algorithm. There is a new
7017 function BIO_new_mem_buf() which creates a read only memory BIO from
7018 an area of memory. Also modified the PKCS#7 routines to use read only
7022 *) Bugfix: ssl23_get_client_hello did not work properly when called in
7023 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
7024 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
7025 but a retry condition occured while trying to read the rest.
7028 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
7029 NID_pkcs7_encrypted by default: this was wrong since this should almost
7030 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
7031 the encrypted data type: this is a more sensible place to put it and it
7032 allows the PKCS#12 code to be tidied up that duplicated this
7036 *) Changed obj_dat.pl script so it takes its input and output files on
7037 the command line. This should avoid shell escape redirection problems
7041 *) Initial support for certificate extension requests, these are included
7042 in things like Xenroll certificate requests. Included functions to allow
7043 extensions to be obtained and added.
7046 *) -crlf option to s_client and s_server for sending newlines as
7047 CRLF (as required by many protocols).
7050 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
7052 *) Install libRSAglue.a when OpenSSL is built with RSAref.
7053 [Ralf S. Engelschall]
7055 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
7056 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
7058 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
7062 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
7063 DH parameters/keys (q is lost during that conversion, but the resulting
7064 DH parameters contain its length).
7066 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
7067 much faster than DH_generate_parameters (which creates parameters
7068 where p = 2*q + 1), and also the smaller q makes DH computations
7069 much more efficient (160-bit exponentiation instead of 1024-bit
7070 exponentiation); so this provides a convenient way to support DHE
7071 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
7072 utter importance to use
7073 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7075 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7076 when such DH parameters are used, because otherwise small subgroup
7077 attacks may become possible!
7080 *) Avoid memory leak in i2d_DHparams.
7083 *) Allow the -k option to be used more than once in the enc program:
7084 this allows the same encrypted message to be read by multiple recipients.
7087 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
7088 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
7089 it will always use the numerical form of the OID, even if it has a short
7093 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
7094 method only got called if p,q,dmp1,dmq1,iqmp components were present,
7095 otherwise bn_mod_exp was called. In the case of hardware keys for example
7096 no private key components need be present and it might store extra data
7097 in the RSA structure, which cannot be accessed from bn_mod_exp.
7098 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
7099 private key operations.
7102 *) Added support for SPARC Linux.
7105 *) pem_password_cb function type incompatibly changed from
7106 typedef int pem_password_cb(char *buf, int size, int rwflag);
7108 ....(char *buf, int size, int rwflag, void *userdata);
7109 so that applications can pass data to their callbacks:
7110 The PEM[_ASN1]_{read,write}... functions and macros now take an
7111 additional void * argument, which is just handed through whenever
7112 the password callback is called.
7113 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
7115 New function SSL_CTX_set_default_passwd_cb_userdata.
7117 Compatibility note: As many C implementations push function arguments
7118 onto the stack in reverse order, the new library version is likely to
7119 interoperate with programs that have been compiled with the old
7120 pem_password_cb definition (PEM_whatever takes some data that
7121 happens to be on the stack as its last argument, and the callback
7122 just ignores this garbage); but there is no guarantee whatsoever that
7125 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
7126 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
7127 problems not only on Windows, but also on some Unix platforms.
7128 To avoid problematic command lines, these definitions are now in an
7129 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
7130 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
7133 *) MIPS III/IV assembler module is reimplemented.
7136 *) More DES library cleanups: remove references to srand/rand and
7137 delete an unused file.
7140 *) Add support for the the free Netwide assembler (NASM) under Win32,
7141 since not many people have MASM (ml) and it can be hard to obtain.
7142 This is currently experimental but it seems to work OK and pass all
7143 the tests. Check out INSTALL.W32 for info.
7146 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
7147 without temporary keys kept an extra copy of the server key,
7148 and connections with temporary keys did not free everything in case
7152 *) New function RSA_check_key and new openssl rsa option -check
7153 for verifying the consistency of RSA keys.
7154 [Ulf Moeller, Bodo Moeller]
7156 *) Various changes to make Win32 compile work:
7157 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
7158 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
7159 comparison" warnings.
7160 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
7163 *) Add a debugging option to PKCS#5 v2 key generation function: when
7164 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
7165 derived keys are printed to stderr.
7168 *) Copy the flags in ASN1_STRING_dup().
7169 [Roman E. Pavlov <pre@mo.msk.ru>]
7171 *) The x509 application mishandled signing requests containing DSA
7172 keys when the signing key was also DSA and the parameters didn't match.
7174 It was supposed to omit the parameters when they matched the signing key:
7175 the verifying software was then supposed to automatically use the CA's
7176 parameters if they were absent from the end user certificate.
7178 Omitting parameters is no longer recommended. The test was also
7179 the wrong way round! This was probably due to unusual behaviour in
7180 EVP_cmp_parameters() which returns 1 if the parameters match.
7181 This meant that parameters were omitted when they *didn't* match and
7182 the certificate was useless. Certificates signed with 'ca' didn't have
7184 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
7186 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
7187 The interface is as follows:
7188 Applications can use
7189 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
7190 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
7191 "off" is now the default.
7192 The library internally uses
7193 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
7194 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
7195 to disable memory-checking temporarily.
7197 Some inconsistent states that previously were possible (and were
7198 even the default) are now avoided.
7200 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
7201 with each memory chunk allocated; this is occasionally more helpful
7202 than just having a counter.
7204 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
7206 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
7210 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
7211 which largely parallels "options", but is for changing API behaviour,
7212 whereas "options" are about protocol behaviour.
7213 Initial "mode" flags are:
7215 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
7216 a single record has been written.
7217 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
7218 retries use the same buffer location.
7219 (But all of the contents must be
7223 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
7226 *) Fix problems with no-hmac etc.
7227 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
7229 *) New functions RSA_get_default_method(), RSA_set_method() and
7230 RSA_get_method(). These allows replacement of RSA_METHODs without having
7231 to mess around with the internals of an RSA structure.
7234 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
7235 Also really enable memory leak checks in openssl.c and in some
7237 [Chad C. Mulligan, Bodo Moeller]
7239 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
7240 up the length of negative integers. This has now been simplified to just
7241 store the length when it is first determined and use it later, rather
7242 than trying to keep track of where data is copied and updating it to
7244 [Steve Henson, reported by Brien Wheeler
7245 <bwheeler@authentica-security.com>]
7247 *) Add a new function PKCS7_signatureVerify. This allows the verification
7248 of a PKCS#7 signature but with the signing certificate passed to the
7249 function itself. This contrasts with PKCS7_dataVerify which assumes the
7250 certificate is present in the PKCS#7 structure. This isn't always the
7251 case: certificates can be omitted from a PKCS#7 structure and be
7252 distributed by "out of band" means (such as a certificate database).
7255 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
7256 function prototypes in pem.h, also change util/mkdef.pl to add the
7257 necessary function names.
7260 *) mk1mf.pl (used by Windows builds) did not properly read the
7261 options set by Configure in the top level Makefile, and Configure
7262 was not even able to write more than one option correctly.
7263 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
7266 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
7267 file to be loaded from a BIO or FILE pointer. The BIO version will
7268 for example allow memory BIOs to contain config info.
7271 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
7272 Whoever hopes to achieve shared-library compatibility across versions
7273 must use this, not the compile-time macro.
7274 (Exercise 0.9.4: Which is the minimum library version required by
7276 Note: All this applies only to multi-threaded programs, others don't
7280 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
7281 through a BIO pair triggered the default case, i.e.
7282 SSLerr(...,SSL_R_UNKNOWN_STATE).
7285 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
7286 can use the SSL library even if none of the specific BIOs is
7290 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
7291 for the encoded length.
7292 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
7294 *) Add initial documentation of the X509V3 functions.
7297 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
7298 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
7299 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
7300 secure PKCS#8 private key format with a high iteration count.
7303 *) Fix determination of Perl interpreter: A perl or perl5
7304 _directory_ in $PATH was also accepted as the interpreter.
7305 [Ralf S. Engelschall]
7307 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
7308 wrong with it but it was very old and did things like calling
7309 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
7313 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
7314 to use the new extension code.
7317 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
7318 with macros. This should make it easier to change their form, add extra
7319 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
7323 *) Add to configuration table a new entry that can specify an alternative
7324 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
7325 according to Mark Crispin <MRC@Panda.COM>.
7329 *) DES CBC did not update the IV. Weird.
7332 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
7333 Changing the behaviour of the former might break existing programs --
7334 where IV updating is needed, des_ncbc_encrypt can be used.
7337 *) When bntest is run from "make test" it drives bc to check its
7338 calculations, as well as internally checking them. If an internal check
7339 fails, it needs to cause bc to give a non-zero result or make test carries
7340 on without noticing the failure. Fixed.
7343 *) DES library cleanups.
7346 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
7347 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
7348 ciphers. NOTE: although the key derivation function has been verified
7349 against some published test vectors it has not been extensively tested
7350 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
7354 *) Instead of "mkdir -p", which is not fully portable, use new
7355 Perl script "util/mkdir-p.pl".
7358 *) Rewrite the way password based encryption (PBE) is handled. It used to
7359 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
7360 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
7361 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
7362 the 'parameter' field of the AlgorithmIdentifier is passed to the
7363 underlying key generation function so it must do its own ASN1 parsing.
7364 This has also changed the EVP_PBE_CipherInit() function which now has a
7365 'parameter' argument instead of literal salt and iteration count values
7366 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
7369 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
7370 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
7371 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
7372 KEY" because this clashed with PKCS#8 unencrypted string. Since this
7373 value was just used as a "magic string" and not used directly its
7374 value doesn't matter.
7377 *) Introduce some semblance of const correctness to BN. Shame C doesn't
7381 *) "linux-sparc64" configuration (ultrapenguin).
7382 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
7383 "linux-sparc" configuration.
7384 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
7386 *) config now generates no-xxx options for missing ciphers.
7389 *) Support the EBCDIC character set (work in progress).
7390 File ebcdic.c not yet included because it has a different license.
7391 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7393 *) Support BS2000/OSD-POSIX.
7394 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7396 *) Make callbacks for key generation use void * instead of char *.
7399 *) Make S/MIME samples compile (not yet tested).
7402 *) Additional typesafe stacks.
7405 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
7409 Changes between 0.9.3 and 0.9.3a [29 May 1999]
7411 *) New configuration variant "sco5-gcc".
7413 *) Updated some demos.
7414 [Sean O Riordain, Wade Scholine]
7416 *) Add missing BIO_free at exit of pkcs12 application.
7419 *) Fix memory leak in conf.c.
7422 *) Updates for Win32 to assembler version of MD5.
7425 *) Set #! path to perl in apps/der_chop to where we found it
7426 instead of using a fixed path.
7429 *) SHA library changes for irix64-mips4-cc.
7432 *) Improvements for VMS support.
7436 Changes between 0.9.2b and 0.9.3 [24 May 1999]
7438 *) Bignum library bug fix. IRIX 6 passes "make test" now!
7439 This also avoids the problems with SC4.2 and unpatched SC5.
7440 [Andy Polyakov <appro@fy.chalmers.se>]
7442 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
7443 These are required because of the typesafe stack would otherwise break
7444 existing code. If old code used a structure member which used to be STACK
7445 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
7446 sk_num or sk_value it would produce an error because the num, data members
7447 are not present in STACK_OF. Now it just produces a warning. sk_set
7448 replaces the old method of assigning a value to sk_value
7449 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
7450 that does this will no longer work (and should use sk_set instead) but
7451 this could be regarded as a "questionable" behaviour anyway.
7454 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
7455 correctly handle encrypted S/MIME data.
7458 *) Change type of various DES function arguments from des_cblock
7459 (which means, in function argument declarations, pointer to char)
7460 to des_cblock * (meaning pointer to array with 8 char elements),
7461 which allows the compiler to do more typechecking; it was like
7462 that back in SSLeay, but with lots of ugly casts.
7464 Introduce new type const_des_cblock.
7467 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
7468 problems: find RecipientInfo structure that matches recipient certificate
7469 and initialise the ASN1 structures properly based on passed cipher.
7472 *) Belatedly make the BN tests actually check the results.
7475 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
7476 to and from BNs: it was completely broken. New compilation option
7477 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
7478 key elements as negative integers.
7481 *) Reorganize and speed up MD5.
7482 [Andy Polyakov <appro@fy.chalmers.se>]
7485 [Richard Levitte <richard@levitte.org>]
7487 *) New option -out to asn1parse to allow the parsed structure to be
7488 output to a file. This is most useful when combined with the -strparse
7489 option to examine the output of things like OCTET STRINGS.
7492 *) Make SSL library a little more fool-proof by not requiring any longer
7493 that SSL_set_{accept,connect}_state be called before
7494 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
7495 in many applications because usually everything *appeared* to work as
7496 intended anyway -- now it really works as intended).
7499 *) Move openssl.cnf out of lib/.
7502 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
7503 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
7504 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
7505 [Ralf S. Engelschall]
7507 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
7508 handle PKCS#7 enveloped data properly.
7509 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
7511 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
7512 copying pointers. The cert_st handling is changed by this in
7513 various ways (and thus what used to be known as ctx->default_cert
7514 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
7515 any longer when s->cert does not give us what we need).
7516 ssl_cert_instantiate becomes obsolete by this change.
7517 As soon as we've got the new code right (possibly it already is?),
7518 we have solved a couple of bugs of the earlier code where s->cert
7519 was used as if it could not have been shared with other SSL structures.
7521 Note that using the SSL API in certain dirty ways now will result
7522 in different behaviour than observed with earlier library versions:
7523 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
7524 does not influence s as it used to.
7526 In order to clean up things more thoroughly, inside SSL_SESSION
7527 we don't use CERT any longer, but a new structure SESS_CERT
7528 that holds per-session data (if available); currently, this is
7529 the peer's certificate chain and, for clients, the server's certificate
7530 and temporary key. CERT holds only those values that can have
7531 meaningful defaults in an SSL_CTX.
7534 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
7535 from the internal representation. Various PKCS#7 fixes: remove some
7536 evil casts and set the enc_dig_alg field properly based on the signing
7540 *) Allow PKCS#12 password to be set from the command line or the
7541 environment. Let 'ca' get its config file name from the environment
7542 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
7546 *) Allow certificate policies extension to use an IA5STRING for the
7547 organization field. This is contrary to the PKIX definition but
7548 VeriSign uses it and IE5 only recognises this form. Document 'x509'
7552 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
7553 without disallowing inline assembler and the like for non-pedantic builds.
7556 *) Support Borland C++ builder.
7557 [Janez Jere <jj@void.si>, modified by Ulf Möller]
7562 *) SHA-1 cleanups and performance enhancements.
7563 [Andy Polyakov <appro@fy.chalmers.se>]
7565 *) Sparc v8plus assembler for the bignum library.
7566 [Andy Polyakov <appro@fy.chalmers.se>]
7568 *) Accept any -xxx and +xxx compiler options in Configure.
7571 *) Update HPUX configuration.
7574 *) Add missing sk_<type>_unshift() function to safestack.h
7575 [Ralf S. Engelschall]
7577 *) New function SSL_CTX_use_certificate_chain_file that sets the
7578 "extra_cert"s in addition to the certificate. (This makes sense
7579 only for "PEM" format files, as chains as a whole are not
7583 *) Support verify_depth from the SSL API.
7584 x509_vfy.c had what can be considered an off-by-one-error:
7585 Its depth (which was not part of the external interface)
7586 was actually counting the number of certificates in a chain;
7587 now it really counts the depth.
7590 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
7591 instead of X509err, which often resulted in confusing error
7592 messages since the error codes are not globally unique
7593 (e.g. an alleged error in ssl3_accept when a certificate
7594 didn't match the private key).
7596 *) New function SSL_CTX_set_session_id_context that allows to set a default
7597 value (so that you don't need SSL_set_session_id_context for each
7598 connection using the SSL_CTX).
7601 *) OAEP decoding bug fix.
7604 *) Support INSTALL_PREFIX for package builders, as proposed by
7608 *) New Configure options "threads" and "no-threads". For systems
7609 where the proper compiler options are known (currently Solaris
7610 and Linux), "threads" is the default.
7613 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
7616 *) Install various scripts to $(OPENSSLDIR)/misc, not to
7617 $(INSTALLTOP)/bin -- they shouldn't clutter directories
7618 such as /usr/local/bin.
7621 *) "make linux-shared" to build shared libraries.
7622 [Niels Poppe <niels@netbox.org>]
7624 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
7627 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
7628 extension adding in x509 utility.
7631 *) Remove NOPROTO sections and error code comments.
7634 *) Partial rewrite of the DEF file generator to now parse the ANSI
7638 *) New Configure options --prefix=DIR and --openssldir=DIR.
7641 *) Complete rewrite of the error code script(s). It is all now handled
7642 by one script at the top level which handles error code gathering,
7643 header rewriting and C source file generation. It should be much better
7644 than the old method: it now uses a modified version of Ulf's parser to
7645 read the ANSI prototypes in all header files (thus the old K&R definitions
7646 aren't needed for error creation any more) and do a better job of
7647 translating function codes into names. The old 'ASN1 error code imbedded
7648 in a comment' is no longer necessary and it doesn't use .err files which
7649 have now been deleted. Also the error code call doesn't have to appear all
7650 on one line (which resulted in some large lines...).
7653 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
7656 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
7657 0 (which usually indicates a closed connection), but continue reading.
7660 *) Fix some race conditions.
7663 *) Add support for CRL distribution points extension. Add Certificate
7664 Policies and CRL distribution points documentation.
7667 *) Move the autogenerated header file parts to crypto/opensslconf.h.
7670 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
7671 8 of keying material. Merlin has also confirmed interop with this fix
7672 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
7673 [Merlin Hughes <merlin@baltimore.ie>]
7675 *) Fix lots of warnings.
7676 [Richard Levitte <levitte@stacken.kth.se>]
7678 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
7679 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
7680 [Richard Levitte <levitte@stacken.kth.se>]
7682 *) Fix problems with sizeof(long) == 8.
7683 [Andy Polyakov <appro@fy.chalmers.se>]
7685 *) Change functions to ANSI C.
7688 *) Fix typos in error codes.
7689 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
7691 *) Remove defunct assembler files from Configure.
7694 *) SPARC v8 assembler BIGNUM implementation.
7695 [Andy Polyakov <appro@fy.chalmers.se>]
7697 *) Support for Certificate Policies extension: both print and set.
7698 Various additions to support the r2i method this uses.
7701 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
7702 return a const string when you are expecting an allocated buffer.
7705 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
7706 types DirectoryString and DisplayText.
7709 *) Add code to allow r2i extensions to access the configuration database,
7710 add an LHASH database driver and add several ctx helper functions.
7713 *) Fix an evil bug in bn_expand2() which caused various BN functions to
7714 fail when they extended the size of a BIGNUM.
7717 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
7718 support typesafe stack.
7721 *) Fix typo in SSL_[gs]et_options().
7722 [Nils Frostberg <nils@medcom.se>]
7724 *) Delete various functions and files that belonged to the (now obsolete)
7725 old X509V3 handling code.
7728 *) New Configure option "rsaref".
7731 *) Don't auto-generate pem.h.
7734 *) Introduce type-safe ASN.1 SETs.
7737 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
7738 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
7740 *) Introduce type-safe STACKs. This will almost certainly break lots of code
7741 that links with OpenSSL (well at least cause lots of warnings), but fear
7742 not: the conversion is trivial, and it eliminates loads of evil casts. A
7743 few STACKed things have been converted already. Feel free to convert more.
7744 In the fullness of time, I'll do away with the STACK type altogether.
7747 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
7748 specified in <certfile> by updating the entry in the index.txt file.
7749 This way one no longer has to edit the index.txt file manually for
7750 revoking a certificate. The -revoke option does the gory details now.
7751 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
7753 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
7754 `-text' option at all and this way the `-noout -text' combination was
7755 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
7756 [Ralf S. Engelschall]
7758 *) Make sure a corresponding plain text error message exists for the
7759 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
7760 verify callback function determined that a certificate was revoked.
7761 [Ralf S. Engelschall]
7763 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
7764 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
7765 all available cipers including rc5, which was forgotten until now.
7766 In order to let the testing shell script know which algorithms
7767 are available, a new (up to now undocumented) command
7768 "openssl list-cipher-commands" is used.
7771 *) Bugfix: s_client occasionally would sleep in select() when
7772 it should have checked SSL_pending() first.
7775 *) New functions DSA_do_sign and DSA_do_verify to provide access to
7776 the raw DSA values prior to ASN.1 encoding.
7779 *) Tweaks to Configure
7780 [Niels Poppe <niels@netbox.org>]
7782 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
7786 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
7789 *) New config option to avoid instructions that are illegal on the 80386.
7790 The default code is faster, but requires at least a 486.
7793 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
7794 SSL2_SERVER_VERSION (not used at all) macros, which are now the
7795 same as SSL2_VERSION anyway.
7798 *) New "-showcerts" option for s_client.
7801 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
7802 application. Various cleanups and fixes.
7805 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
7806 modify error routines to work internally. Add error codes and PBE init
7807 to library startup routines.
7810 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
7811 packing functions to asn1 and evp. Changed function names and error
7812 codes along the way.
7815 *) PKCS12 integration: and so it begins... First of several patches to
7816 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
7817 objects to objects.h
7820 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
7821 and display support for Thawte strong extranet extension.
7824 *) Add LinuxPPC support.
7825 [Jeff Dubrule <igor@pobox.org>]
7827 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
7828 bn_div_words in alpha.s.
7829 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
7831 *) Make sure the RSA OAEP test is skipped under -DRSAref because
7832 OAEP isn't supported when OpenSSL is built with RSAref.
7833 [Ulf Moeller <ulf@fitug.de>]
7835 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
7836 so they no longer are missing under -DNOPROTO.
7837 [Soren S. Jorvang <soren@t.dk>]
7840 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
7842 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
7843 doesn't work when the session is reused. Coming soon!
7846 *) Fix a security hole, that allows sessions to be reused in the wrong
7847 context thus bypassing client cert protection! All software that uses
7848 client certs and session caches in multiple contexts NEEDS PATCHING to
7849 allow session reuse! A fuller solution is in the works.
7850 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
7852 *) Some more source tree cleanups (removed obsolete files
7853 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
7854 permission on "config" script to be executable) and a fix for the INSTALL
7856 [Ulf Moeller <ulf@fitug.de>]
7858 *) Remove some legacy and erroneous uses of malloc, free instead of
7860 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
7862 *) Make rsa_oaep_test return non-zero on error.
7863 [Ulf Moeller <ulf@fitug.de>]
7865 *) Add support for native Solaris shared libraries. Configure
7866 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
7867 if someone would make that last step automatic.
7868 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
7870 *) ctx_size was not built with the right compiler during "make links". Fixed.
7873 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
7874 except NULL ciphers". This means the default cipher list will no longer
7875 enable NULL ciphers. They need to be specifically enabled e.g. with
7876 the string "DEFAULT:eNULL".
7879 *) Fix to RSA private encryption routines: if p < q then it would
7880 occasionally produce an invalid result. This will only happen with
7881 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
7884 *) Be less restrictive and allow also `perl util/perlpath.pl
7885 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
7886 because this way one can also use an interpreter named `perl5' (which is
7887 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
7888 installed as `perl').
7889 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7891 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
7892 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7894 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
7895 advapi32.lib to Win32 build and change the pem test comparision
7896 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
7897 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
7898 and crypto/des/ede_cbcm_enc.c.
7901 *) DES quad checksum was broken on big-endian architectures. Fixed.
7904 *) Comment out two functions in bio.h that aren't implemented. Fix up the
7905 Win32 test batch file so it (might) work again. The Win32 test batch file
7906 is horrible: I feel ill....
7909 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
7910 in e_os.h. Audit of header files to check ANSI and non ANSI
7911 sections: 10 functions were absent from non ANSI section and not exported
7912 from Windows DLLs. Fixed up libeay.num for new functions.
7915 *) Make `openssl version' output lines consistent.
7916 [Ralf S. Engelschall]
7918 *) Fix Win32 symbol export lists for BIO functions: Added
7919 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
7920 to ms/libeay{16,32}.def.
7921 [Ralf S. Engelschall]
7923 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
7924 fine under Unix and passes some trivial tests I've now added. But the
7925 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
7926 added to make sure no one expects that this stuff really works in the
7927 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
7928 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
7930 [Ralf S. Engelschall]
7932 *) Fix the generation of two part addresses in perl.
7933 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
7935 *) Add config entry for Linux on MIPS.
7936 [John Tobey <jtobey@channel1.com>]
7938 *) Make links whenever Configure is run, unless we are on Windoze.
7941 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
7942 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
7946 *) Add a useful kludge to allow package maintainers to specify compiler and
7947 other platforms details on the command line without having to patch the
7948 Configure script everytime: One now can use ``perl Configure
7949 <id>:<details>'', i.e. platform ids are allowed to have details appended
7950 to them (seperated by colons). This is treated as there would be a static
7951 pre-configured entry in Configure's %table under key <id> with value
7952 <details> and ``perl Configure <id>'' is called. So, when you want to
7953 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
7954 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
7955 now, which overrides the FreeBSD-elf entry on-the-fly.
7956 [Ralf S. Engelschall]
7958 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
7961 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
7962 on the `perl Configure ...' command line. This way one can compile
7963 OpenSSL libraries with Position Independent Code (PIC) which is needed
7964 for linking it into DSOs.
7965 [Ralf S. Engelschall]
7967 *) Remarkably, export ciphers were totally broken and no-one had noticed!
7971 *) Cleaned up the LICENSE document: The official contact for any license
7972 questions now is the OpenSSL core team under openssl-core@openssl.org.
7973 And add a paragraph about the dual-license situation to make sure people
7974 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
7975 to the OpenSSL toolkit.
7976 [Ralf S. Engelschall]
7978 *) General source tree makefile cleanups: Made `making xxx in yyy...'
7979 display consistent in the source tree and replaced `/bin/rm' by `rm'.
7980 Additonally cleaned up the `make links' target: Remove unnecessary
7981 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
7982 to speed processing and no longer clutter the display with confusing
7983 stuff. Instead only the actually done links are displayed.
7984 [Ralf S. Engelschall]
7986 *) Permit null encryption ciphersuites, used for authentication only. It used
7987 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
7988 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
7992 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
7993 signed attributes when verifying signatures (this would break them),
7994 the detached data encoding was wrong and public keys obtained using
7995 X509_get_pubkey() weren't freed.
7998 *) Add text documentation for the BUFFER functions. Also added a work around
7999 to a Win95 console bug. This was triggered by the password read stuff: the
8000 last character typed gets carried over to the next fread(). If you were
8001 generating a new cert request using 'req' for example then the last
8002 character of the passphrase would be CR which would then enter the first
8006 *) Added the new `Includes OpenSSL Cryptography Software' button as
8007 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
8008 button and can be used by applications based on OpenSSL to show the
8009 relationship to the OpenSSL project.
8010 [Ralf S. Engelschall]
8012 *) Remove confusing variables in function signatures in files
8013 ssl/ssl_lib.c and ssl/ssl.h.
8014 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8016 *) Don't install bss_file.c under PREFIX/include/
8017 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8019 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
8020 functions that return function pointers and has support for NT specific
8021 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
8022 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
8023 unsigned to signed types: this was killing the Win32 compile.
8026 *) Add new certificate file to stack functions,
8027 SSL_add_dir_cert_subjects_to_stack() and
8028 SSL_add_file_cert_subjects_to_stack(). These largely supplant
8029 SSL_load_client_CA_file(), and can be used to add multiple certs easily
8030 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
8031 This means that Apache-SSL and similar packages don't have to mess around
8032 to add as many CAs as they want to the preferred list.
8035 *) Experiment with doxygen documentation. Currently only partially applied to
8037 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
8038 openssl.doxy as the configuration file.
8041 *) Get rid of remaining C++-style comments which strict C compilers hate.
8042 [Ralf S. Engelschall, pointed out by Carlos Amengual]
8044 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
8045 compiled in by default: it has problems with large keys.
8048 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
8049 DH private keys and/or callback functions which directly correspond to
8050 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
8051 is needed for applications which have to configure certificates on a
8052 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
8054 For the RSA certificate situation is makes no difference, but
8055 for the DSA certificate situation this fixes the "no shared cipher"
8056 problem where the OpenSSL cipher selection procedure failed because the
8057 temporary keys were not overtaken from the context and the API provided
8058 no way to reconfigure them.
8059 The new functions now let applications reconfigure the stuff and they
8060 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
8061 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
8062 non-public-API function ssl_cert_instantiate() is used as a helper
8063 function and also to reduce code redundancy inside ssl_rsa.c.
8064 [Ralf S. Engelschall]
8066 *) Move s_server -dcert and -dkey options out of the undocumented feature
8067 area because they are useful for the DSA situation and should be
8068 recognized by the users.
8069 [Ralf S. Engelschall]
8071 *) Fix the cipher decision scheme for export ciphers: the export bits are
8072 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
8073 SSL_EXP_MASK. So, the original variable has to be used instead of the
8074 already masked variable.
8075 [Richard Levitte <levitte@stacken.kth.se>]
8077 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
8078 [Richard Levitte <levitte@stacken.kth.se>]
8080 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
8081 from `int' to `unsigned int' because it's a length and initialized by
8082 EVP_DigestFinal() which expects an `unsigned int *'.
8083 [Richard Levitte <levitte@stacken.kth.se>]
8085 *) Don't hard-code path to Perl interpreter on shebang line of Configure
8086 script. Instead use the usual Shell->Perl transition trick.
8087 [Ralf S. Engelschall]
8089 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
8090 (in addition to RSA certificates) to match the behaviour of `openssl dsa
8091 -noout -modulus' as it's already the case for `openssl rsa -noout
8092 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
8093 currently the public key is printed (a decision which was already done by
8094 `openssl dsa -modulus' in the past) which serves a similar purpose.
8095 Additionally the NO_RSA no longer completely removes the whole -modulus
8096 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
8098 [Ralf S. Engelschall]
8100 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
8101 BIO. See the source (crypto/evp/bio_ok.c) for more info.
8102 [Arne Ansper <arne@ats.cyber.ee>]
8104 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
8105 to be added. Now both 'req' and 'ca' can use new objects defined in the
8109 *) Add cool BIO that does syslog (or event log on NT).
8110 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
8112 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
8113 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
8114 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
8115 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
8118 *) Add preliminary config info for new extension code.
8121 *) Make RSA_NO_PADDING really use no padding.
8122 [Ulf Moeller <ulf@fitug.de>]
8124 *) Generate errors when private/public key check is done.
8127 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
8128 for some CRL extensions and new objects added.
8131 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
8132 key usage extension and fuller support for authority key id.
8135 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
8136 padding method for RSA, which is recommended for new applications in PKCS
8137 #1 v2.0 (RFC 2437, October 1998).
8138 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
8139 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
8140 against Bleichbacher's attack on RSA.
8141 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
8144 *) Updates to the new SSL compression code
8145 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8147 *) Fix so that the version number in the master secret, when passed
8148 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
8149 (because the server will not accept higher), that the version number
8150 is 0x03,0x01, not 0x03,0x00
8151 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8153 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
8154 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
8155 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
8158 *) Support for RAW extensions where an arbitrary extension can be
8159 created by including its DER encoding. See apps/openssl.cnf for
8163 *) Make sure latest Perl versions don't interpret some generated C array
8164 code as Perl array code in the crypto/err/err_genc.pl script.
8165 [Lars Weber <3weber@informatik.uni-hamburg.de>]
8167 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
8168 not many people have the assembler. Various Win32 compilation fixes and
8169 update to the INSTALL.W32 file with (hopefully) more accurate Win32
8173 *) Modify configure script 'Configure' to automatically create crypto/date.h
8174 file under Win32 and also build pem.h from pem.org. New script
8175 util/mkfiles.pl to create the MINFO file on environments that can't do a
8176 'make files': perl util/mkfiles.pl >MINFO should work.
8179 *) Major rework of DES function declarations, in the pursuit of correctness
8180 and purity. As a result, many evil casts evaporated, and some weirdness,
8181 too. You may find this causes warnings in your code. Zapping your evil
8182 casts will probably fix them. Mostly.
8185 *) Fix for a typo in asn1.h. Bug fix to object creation script
8186 obj_dat.pl. It considered a zero in an object definition to mean
8187 "end of object": none of the objects in objects.h have any zeros
8188 so it wasn't spotted.
8189 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
8191 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
8192 Masking (CBCM). In the absence of test vectors, the best I have been able
8193 to do is check that the decrypt undoes the encrypt, so far. Send me test
8194 vectors if you have them.
8197 *) Correct calculation of key length for export ciphers (too much space was
8198 allocated for null ciphers). This has not been tested!
8201 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
8202 message is now correct (it understands "crypto" and "ssl" on its
8203 command line). There is also now an "update" option. This will update
8204 the util/ssleay.num and util/libeay.num files with any new functions.
8206 perl util/mkdef.pl crypto ssl update
8207 it will update them.
8210 *) Overhauled the Perl interface (perl/*):
8211 - ported BN stuff to OpenSSL's different BN library
8212 - made the perl/ source tree CVS-aware
8213 - renamed the package from SSLeay to OpenSSL (the files still contain
8214 their history because I've copied them in the repository)
8215 - removed obsolete files (the test scripts will be replaced
8216 by better Test::Harness variants in the future)
8217 [Ralf S. Engelschall]
8219 *) First cut for a very conservative source tree cleanup:
8220 1. merge various obsolete readme texts into doc/ssleay.txt
8221 where we collect the old documents and readme texts.
8222 2. remove the first part of files where I'm already sure that we no
8223 longer need them because of three reasons: either they are just temporary
8224 files which were left by Eric or they are preserved original files where
8225 I've verified that the diff is also available in the CVS via "cvs diff
8226 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
8227 the crypto/md/ stuff).
8228 [Ralf S. Engelschall]
8230 *) More extension code. Incomplete support for subject and issuer alt
8231 name, issuer and authority key id. Change the i2v function parameters
8232 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
8233 what that's for :-) Fix to ASN1 macro which messed up
8234 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
8237 *) Preliminary support for ENUMERATED type. This is largely copied from the
8241 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
8242 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8244 *) Make sure `make rehash' target really finds the `openssl' program.
8245 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8247 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
8248 like to hear about it if this slows down other processors.
8251 *) Add CygWin32 platform information to Configure script.
8252 [Alan Batie <batie@aahz.jf.intel.com>]
8254 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
8255 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
8257 *) New program nseq to manipulate netscape certificate sequences
8260 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
8264 *) Fixes to BN code. Previously the default was to define BN_RECURSION
8265 but the BN code had some problems that would cause failures when
8266 doing certificate verification and some other functions.
8267 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8269 *) Add ASN1 and PEM code to support netscape certificate sequences.
8272 *) Add ASN1 and PEM code to support netscape certificate sequences.
8275 *) Add several PKIX and private extended key usage OIDs.
8278 *) Modify the 'ca' program to handle the new extension code. Modify
8279 openssl.cnf for new extension format, add comments.
8282 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
8283 and add a sample to openssl.cnf so req -x509 now adds appropriate
8287 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
8288 error code, add initial support to X509_print() and x509 application.
8291 *) Takes a deep breath and start addding X509 V3 extension support code. Add
8292 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
8293 stuff is currently isolated and isn't even compiled yet.
8296 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
8297 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
8298 Removed the versions check from X509 routines when loading extensions:
8299 this allows certain broken certificates that don't set the version
8300 properly to be processed.
8303 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
8304 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
8305 can still be regenerated with "make depend".
8308 *) Spelling mistake in C version of CAST-128.
8309 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
8311 *) Changes to the error generation code. The perl script err-code.pl
8312 now reads in the old error codes and retains the old numbers, only
8313 adding new ones if necessary. It also only changes the .err files if new
8314 codes are added. The makefiles have been modified to only insert errors
8315 when needed (to avoid needlessly modifying header files). This is done
8316 by only inserting errors if the .err file is newer than the auto generated
8317 C file. To rebuild all the error codes from scratch (the old behaviour)
8318 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
8319 or delete all the .err files.
8322 *) CAST-128 was incorrectly implemented for short keys. The C version has
8323 been fixed, but is untested. The assembler versions are also fixed, but
8324 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
8325 to regenerate it if needed.
8326 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
8327 Hagino <itojun@kame.net>]
8329 *) File was opened incorrectly in randfile.c.
8330 [Ulf Möller <ulf@fitug.de>]
8332 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
8333 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
8334 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
8335 al: it's just almost always a UTCTime. Note this patch adds new error
8336 codes so do a "make errors" if there are problems.
8339 *) Correct Linux 1 recognition in config.
8340 [Ulf Möller <ulf@fitug.de>]
8342 *) Remove pointless MD5 hash when using DSA keys in ca.
8343 [Anonymous <nobody@replay.com>]
8345 *) Generate an error if given an empty string as a cert directory. Also
8346 generate an error if handed NULL (previously returned 0 to indicate an
8347 error, but didn't set one).
8348 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
8350 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
8353 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
8354 parameters. This was causing a warning which killed off the Win32 compile.
8357 *) Remove C++ style comments from crypto/bn/bn_local.h.
8358 [Neil Costigan <neil.costigan@celocom.com>]
8360 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
8361 based on a text string, looking up short and long names and finally
8362 "dot" format. The "dot" format stuff didn't work. Added new function
8363 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
8364 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
8365 OID is not part of the table.
8368 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
8369 X509_LOOKUP_by_alias().
8372 *) Sort openssl functions by name.
8375 *) Get the gendsa program working (hopefully) and add it to app list. Remove
8376 encryption from sample DSA keys (in case anyone is interested the password
8380 *) Make _all_ *_free functions accept a NULL pointer.
8381 [Frans Heymans <fheymans@isaserver.be>]
8383 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
8385 [Anonymous <nobody@replay.com>]
8387 *) s_server should send the CAfile as acceptable CAs, not its own cert.
8388 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8390 *) Don't blow it for numeric -newkey arguments to apps/req.
8391 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8393 *) Temp key "for export" tests were wrong in s3_srvr.c.
8394 [Anonymous <nobody@replay.com>]
8396 *) Add prototype for temp key callback functions
8397 SSL_CTX_set_tmp_{rsa,dh}_callback().
8400 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
8401 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
8404 *) X509_name_add_entry() freed the wrong thing after an error.
8405 [Arne Ansper <arne@ats.cyber.ee>]
8407 *) rsa_eay.c would attempt to free a NULL context.
8408 [Arne Ansper <arne@ats.cyber.ee>]
8410 *) BIO_s_socket() had a broken should_retry() on Windoze.
8411 [Arne Ansper <arne@ats.cyber.ee>]
8413 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
8414 [Arne Ansper <arne@ats.cyber.ee>]
8416 *) Make sure the already existing X509_STORE->depth variable is initialized
8417 in X509_STORE_new(), but document the fact that this variable is still
8418 unused in the certificate verification process.
8419 [Ralf S. Engelschall]
8421 *) Fix the various library and apps files to free up pkeys obtained from
8422 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
8425 *) Fix reference counting in X509_PUBKEY_get(). This makes
8426 demos/maurice/example2.c work, amongst others, probably.
8427 [Steve Henson and Ben Laurie]
8429 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
8430 `openssl' and second, the shortcut symlinks for the `openssl <command>'
8431 are no longer created. This way we have a single and consistent command
8432 line interface `openssl <command>', similar to `cvs <command>'.
8433 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
8435 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
8436 BIT STRING wrapper always have zero unused bits.
8439 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
8442 *) Make the top-level INSTALL documentation easier to understand.
8445 *) Makefiles updated to exit if an error occurs in a sub-directory
8446 make (including if user presses ^C) [Paul Sutton]
8448 *) Make Montgomery context stuff explicit in RSA data structure.
8451 *) Fix build order of pem and err to allow for generated pem.h.
8454 *) Fix renumbering bug in X509_NAME_delete_entry().
8457 *) Enhanced the err-ins.pl script so it makes the error library number
8458 global and can add a library name. This is needed for external ASN1 and
8459 other error libraries.
8462 *) Fixed sk_insert which never worked properly.
8465 *) Fix ASN1 macros so they can handle indefinite length construted
8466 EXPLICIT tags. Some non standard certificates use these: they can now
8470 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
8471 into a single doc/ssleay.txt bundle. This way the information is still
8472 preserved but no longer messes up this directory. Now it's new room for
8473 the new set of documenation files.
8474 [Ralf S. Engelschall]
8476 *) SETs were incorrectly DER encoded. This was a major pain, because they
8477 shared code with SEQUENCEs, which aren't coded the same. This means that
8478 almost everything to do with SETs or SEQUENCEs has either changed name or
8479 number of arguments.
8480 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
8482 *) Fix test data to work with the above.
8485 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
8486 was already fixed by Eric for 0.9.1 it seems.
8487 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
8489 *) Autodetect FreeBSD3.
8492 *) Fix various bugs in Configure. This affects the following platforms:
8496 unixware-2.0-pentium
8500 *) Eliminate generated files from CVS. Reorder tests to regenerate files
8501 before they are needed.
8504 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
8508 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
8510 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
8511 changed SSLeay to OpenSSL in version strings.
8512 [Ralf S. Engelschall]
8514 *) Some fixups to the top-level documents.
8517 *) Fixed the nasty bug where rsaref.h was not found under compile-time
8518 because the symlink to include/ was missing.
8519 [Ralf S. Engelschall]
8521 *) Incorporated the popular no-RSA/DSA-only patches
8522 which allow to compile a RSA-free SSLeay.
8523 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
8525 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
8526 when "ssleay" is still not found.
8527 [Ralf S. Engelschall]
8529 *) Added more platforms to Configure: Cray T3E, HPUX 11,
8530 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
8532 *) Updated the README file.
8533 [Ralf S. Engelschall]
8535 *) Added various .cvsignore files in the CVS repository subdirs
8536 to make a "cvs update" really silent.
8537 [Ralf S. Engelschall]
8539 *) Recompiled the error-definition header files and added
8540 missing symbols to the Win32 linker tables.
8541 [Ralf S. Engelschall]
8543 *) Cleaned up the top-level documents;
8544 o new files: CHANGES and LICENSE
8545 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
8546 o merged COPYRIGHT into LICENSE
8547 o removed obsolete TODO file
8548 o renamed MICROSOFT to INSTALL.W32
8549 [Ralf S. Engelschall]
8551 *) Removed dummy files from the 0.9.1b source tree:
8552 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
8553 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
8554 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
8555 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
8556 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
8557 [Ralf S. Engelschall]
8559 *) Added various platform portability fixes.
8562 *) The Genesis of the OpenSSL rpject:
8563 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
8564 Young and Tim J. Hudson created while they were working for C2Net until
8566 [The OpenSSL Project]
8569 Changes between 0.9.0b and 0.9.1b [not released]
8571 *) Updated a few CA certificates under certs/
8574 *) Changed some BIGNUM api stuff.
8577 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
8578 DGUX x86, Linux Alpha, etc.
8581 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
8582 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
8586 *) Add -strparse option to asn1pars program which parses nested
8588 [Dr Stephen Henson <shenson@bigfoot.com>]
8590 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
8593 *) DSA fix for "ca" program.
8596 *) Added "-genkey" option to "dsaparam" program.
8599 *) Added RIPE MD160 (rmd160) message digest.
8602 *) Added -a (all) option to "ssleay version" command.
8605 *) Added PLATFORM define which is the id given to Configure.
8608 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
8611 *) Extended the ASN.1 parser routines.
8614 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
8617 *) Added a BN_CTX to the BN library.
8620 *) Fixed the weak key values in DES library
8623 *) Changed API in EVP library for cipher aliases.
8626 *) Added support for RC2/64bit cipher.
8629 *) Converted the lhash library to the crypto/mem.c functions.
8632 *) Added more recognized ASN.1 object ids.
8635 *) Added more RSA padding checks for SSL/TLS.
8638 *) Added BIO proxy/filter functionality.
8641 *) Added extra_certs to SSL_CTX which can be used
8642 send extra CA certificates to the client in the CA cert chain sending
8643 process. It can be configured with SSL_CTX_add_extra_chain_cert().
8646 *) Now Fortezza is denied in the authentication phase because
8647 this is key exchange mechanism is not supported by SSLeay at all.
8650 *) Additional PKCS1 checks.
8653 *) Support the string "TLSv1" for all TLS v1 ciphers.
8656 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
8657 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
8660 *) Fixed a few memory leaks.
8663 *) Fixed various code and comment typos.
8666 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
8667 bytes sent in the client random.
8668 [Edward Bishop <ebishop@spyglass.com>]