5 Changes between 0.9.8w and 0.9.8x [xx XXX xxxx]
7 *) Sanity check record length before skipping explicit IV in DTLS
10 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
11 fuzzing as a service testing platform.
15 *) Initialise tkeylen properly when encrypting CMS messages.
16 Thanks to Solar Designer of Openwall for reporting this issue.
19 Changes between 0.9.8v and 0.9.8w [23 Apr 2012]
21 *) The fix for CVE-2012-2110 did not take into account that the
22 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an
23 int in OpenSSL 0.9.8, making it still vulnerable. Fix by
24 rejecting negative len parameter. (CVE-2012-2131)
25 [Tomas Hoger <thoger@redhat.com>]
27 Changes between 0.9.8u and 0.9.8v [19 Apr 2012]
29 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
30 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
31 in CRYPTO_realloc_clean.
33 Thanks to Tavis Ormandy, Google Security Team, for discovering this
34 issue and to Adam Langley <agl@chromium.org> for fixing it.
36 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
38 Changes between 0.9.8t and 0.9.8u [12 Mar 2012]
40 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
41 in CMS and PKCS7 code. When RSA decryption fails use a random key for
42 content decryption and always return the same error. Note: this attack
43 needs on average 2^20 messages so it only affects automated senders. The
44 old behaviour can be reenabled in the CMS code by setting the
45 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
46 an MMA defence is not necessary.
47 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
48 this issue. (CVE-2012-0884)
51 *) Fix CVE-2011-4619: make sure we really are receiving a
52 client hello before rejecting multiple SGC restarts. Thanks to
53 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
56 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
58 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
59 Thanks to Antonio Martin, Enterprise Secure Access Research and
60 Development, Cisco Systems, Inc. for discovering this bug and
61 preparing a fix. (CVE-2012-0050)
64 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
66 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
67 of the Vaudenay padding oracle attack on CBC mode encryption
68 which enables an efficient plaintext recovery attack against
69 the OpenSSL implementation of DTLS. Their attack exploits timing
70 differences arising during decryption processing. A research
71 paper describing this attack can be found at:
72 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
73 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
74 Security Group at Royal Holloway, University of London
75 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
76 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
77 for preparing the fix. (CVE-2011-4108)
78 [Robin Seggelmann, Michael Tuexen]
80 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
81 [Ben Laurie, Kasper <ekasper@google.com>]
83 *) Clear bytes used for block padding of SSL 3.0 records.
85 [Adam Langley (Google)]
87 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
88 Kadianakis <desnacked@gmail.com> for discovering this issue and
89 Adam Langley for preparing the fix. (CVE-2011-4619)
90 [Adam Langley (Google)]
92 *) Prevent malformed RFC3779 data triggering an assertion failure.
93 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
94 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
95 [Rob Austein <sra@hactrn.net>]
97 *) Fix ssl_ciph.c set-up race.
98 [Adam Langley (Google)]
100 *) Fix spurious failures in ecdsatest.c.
101 [Emilia Käsper (Google)]
103 *) Fix the BIO_f_buffer() implementation (which was mixing different
104 interpretations of the '..._len' fields).
105 [Adam Langley (Google)]
107 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
108 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
109 threads won't reuse the same blinding coefficients.
111 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
112 lock to call BN_BLINDING_invert_ex, and avoids one use of
113 BN_BLINDING_update for each BN_BLINDING structure (previously,
114 the last update always remained unused).
115 [Emilia Käsper (Google)]
117 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
118 for multi-threaded use of ECDH.
119 [Adam Langley (Google)]
121 *) Fix x509_name_ex_d2i memory leak on bad inputs.
124 *) Add protection against ECDSA timing attacks as mentioned in the paper
125 by Billy Bob Brumley and Nicola Tuveri, see:
127 http://eprint.iacr.org/2011/232.pdf
129 [Billy Bob Brumley and Nicola Tuveri]
131 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
133 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
134 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
136 *) Fix bug in string printing code: if *any* escaping is enabled we must
137 escape the escape character (backslash) or the resulting string is
141 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
143 *) Disable code workaround for ancient and obsolete Netscape browsers
144 and servers: an attacker can use it in a ciphersuite downgrade attack.
145 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
148 *) Fixed J-PAKE implementation error, originally discovered by
149 Sebastien Martini, further info and confirmation from Stefan
150 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
153 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
155 *) Fix extension code to avoid race conditions which can result in a buffer
156 overrun vulnerability: resumed sessions must not be modified as they can
157 be shared by multiple threads. CVE-2010-3864
160 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
163 *) Don't reencode certificate when calculating signature: cache and use
164 the original encoding instead. This makes signature verification of
165 some broken encodings work correctly.
168 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
169 is also one of the inputs.
170 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
172 *) Don't repeatedly append PBE algorithms to table if they already exist.
173 Sort table on each new add. This effectively makes the table read only
174 after all algorithms are added and subsequent calls to PKCS12_pbe_add
178 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
180 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
183 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
184 access or freeing data twice (CVE-2010-0742)
185 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
187 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
188 common in certificates and some applications which only call
189 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
193 Reduce copying into .apps and .test in makevms.com
194 Don't try to use blank CA certificate in CA.com
195 Allow use of C files from original directories in maketests.com
196 [Steven M. Schweda" <sms@antinode.info>]
198 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
200 *) When rejecting SSL/TLS records due to an incorrect version number, never
201 update s->server with a new major version number. As of
202 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
203 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
204 the previous behavior could result in a read attempt at NULL when
205 receiving specific incorrect SSL/TLS records once record payload
206 protection is active. (CVE-2010-0740)
207 [Bodo Moeller, Adam Langley <agl@chromium.org>]
209 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
210 could be crashed if the relevant tables were not present (e.g. chrooted).
211 [Tomas Hoger <thoger@redhat.com>]
213 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
215 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
216 [Martin Olsson, Neel Mehta]
218 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
219 accommodate for stack sorting, always a write lock!).
222 *) On some versions of WIN32 Heap32Next is very slow. This can cause
223 excessive delays in the RAND_poll(): over a minute. As a workaround
224 include a time check in the inner Heap32Next loop too.
227 *) The code that handled flushing of data in SSL/TLS originally used the
228 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
229 the problem outlined in PR#1949. The fix suggested there however can
230 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
231 of Apache). So instead simplify the code to flush unconditionally.
232 This should be fine since flushing with no data to flush is a no op.
235 *) Handle TLS versions 2.0 and later properly and correctly use the
236 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
237 off ancient servers have a habit of sticking around for a while...
240 *) Modify compression code so it frees up structures without using the
241 ex_data callbacks. This works around a problem where some applications
242 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
243 restarting) then use compression (e.g. SSL with compression) later.
244 This results in significant per-connection memory leaks and
245 has caused some security issues including CVE-2008-1678 and
249 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
250 change when encrypting or decrypting.
253 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
254 connect and renegotiate with servers which do not support RI.
255 Until RI is more widely deployed this option is enabled by default.
258 *) Add "missing" ssl ctrls to clear options and mode.
261 *) If client attempts to renegotiate and doesn't support RI respond with
262 a no_renegotiation alert as required by RFC5746. Some renegotiating
263 TLS clients will continue a connection gracefully when they receive
264 the alert. Unfortunately OpenSSL mishandled this alert and would hang
265 waiting for a server hello which it will never receive. Now we treat a
266 received no_renegotiation alert as a fatal error. This is because
267 applications requesting a renegotiation might well expect it to succeed
268 and would have no code in place to handle the server denying it so the
269 only safe thing to do is to terminate the connection.
272 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
273 peer supports secure renegotiation and 0 otherwise. Print out peer
274 renegotiation support in s_client/s_server.
277 *) Replace the highly broken and deprecated SPKAC certification method with
278 the updated NID creation version. This should correctly handle UTF8.
281 *) Implement RFC5746. Re-enable renegotiation but require the extension
282 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
283 turns out to be a bad idea. It has been replaced by
284 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
285 SSL_CTX_set_options(). This is really not recommended unless you
286 know what you are doing.
287 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
289 *) Fixes to stateless session resumption handling. Use initial_ctx when
290 issuing and attempting to decrypt tickets in case it has changed during
291 servername handling. Use a non-zero length session ID when attempting
292 stateless session resumption: this makes it possible to determine if
293 a resumption has occurred immediately after receiving server hello
294 (several places in OpenSSL subtly assume this) instead of later in
298 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
299 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
300 fixes for a few places where the return code is not checked
302 [Julia Lawall <julia@diku.dk>]
304 *) Add --strict-warnings option to Configure script to include devteam
305 warnings in other configurations.
308 *) Add support for --libdir option and LIBDIR variable in makefiles. This
309 makes it possible to install openssl libraries in locations which
310 have names other than "lib", for example "/usr/lib64" which some
312 [Steve Henson, based on patch from Jeremy Utley]
314 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
315 X690 8.9.12 and can produce some misleading textual output of OIDs.
316 [Steve Henson, reported by Dan Kaminsky]
318 *) Delete MD2 from algorithm tables. This follows the recommendation in
319 several standards that it is not used in new applications due to
320 several cryptographic weaknesses. For binary compatibility reasons
321 the MD2 API is still compiled in by default.
324 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
328 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
329 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
331 [Guenter <lists@gknw.net>]
333 *) Fix the server certificate chain building code to use X509_verify_cert(),
334 it used to have an ad-hoc builder which was unable to cope with anything
335 other than a simple chain.
336 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
338 *) Don't check self signed certificate signatures in X509_verify_cert()
339 by default (a flag can override this): it just wastes time without
340 adding any security. As a useful side effect self signed root CAs
341 with non-FIPS digests are now usable in FIPS mode.
344 *) In dtls1_process_out_of_seq_message() the check if the current message
345 is already buffered was missing. For every new message was memory
346 allocated, allowing an attacker to perform an denial of service attack
347 with sending out of seq handshake messages until there is no memory
348 left. Additionally every future messege was buffered, even if the
349 sequence number made no sense and would be part of another handshake.
350 So only messages with sequence numbers less than 10 in advance will be
351 buffered. (CVE-2009-1378)
352 [Robin Seggelmann, discovered by Daniel Mentz]
354 *) Records are buffered if they arrive with a future epoch to be
355 processed after finishing the corresponding handshake. There is
356 currently no limitation to this buffer allowing an attacker to perform
357 a DOS attack with sending records with future epochs until there is no
358 memory left. This patch adds the pqueue_size() function to detemine
359 the size of a buffer and limits the record buffer to 100 entries.
361 [Robin Seggelmann, discovered by Daniel Mentz]
363 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
364 parent structure is freed. (CVE-2009-1379)
367 *) Handle non-blocking I/O properly in SSL_shutdown() call.
368 [Darryl Miles <darryl-mailinglists@netbauds.net>]
371 [Ilya O. <vrghost@gmail.com>]
373 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
375 *) Disable renegotiation completely - this fixes a severe security
376 problem (CVE-2009-3555) at the cost of breaking all
377 renegotiation. Renegotiation can be re-enabled by setting
378 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
379 run-time. This is really not recommended unless you know what
383 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
385 *) Don't set val to NULL when freeing up structures, it is freed up by
386 underlying code. If sizeof(void *) > sizeof(long) this can result in
387 zeroing past the valid field. (CVE-2009-0789)
388 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
390 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
391 checked correctly. This would allow some invalid signed attributes to
392 appear to verify correctly. (CVE-2009-0591)
393 [Ivan Nestlerode <inestlerode@us.ibm.com>]
395 *) Reject UniversalString and BMPString types with invalid lengths. This
396 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
397 a legal length. (CVE-2009-0590)
400 *) Set S/MIME signing as the default purpose rather than setting it
401 unconditionally. This allows applications to override it at the store
405 *) Permit restricted recursion of ASN1 strings. This is needed in practice
406 to handle some structures.
409 *) Improve efficiency of mem_gets: don't search whole buffer each time
411 [Jeremy Shapiro <jnshapir@us.ibm.com>]
413 *) New -hex option for openssl rand.
416 *) Print out UTF8String and NumericString when parsing ASN1.
419 *) Support NumericString type for name components.
422 *) Allow CC in the environment to override the automatically chosen
423 compiler. Note that nothing is done to ensure flags work with the
427 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
429 *) Properly check EVP_VerifyFinal() and similar return values
431 [Ben Laurie, Bodo Moeller, Google Security Team]
433 *) Enable TLS extensions by default.
436 *) Allow the CHIL engine to be loaded, whether the application is
437 multithreaded or not. (This does not release the developer from the
438 obligation to set up the dynamic locking callbacks.)
439 [Sander Temme <sander@temme.net>]
441 *) Use correct exit code if there is an error in dgst command.
442 [Steve Henson; problem pointed out by Roland Dirlewanger]
444 *) Tweak Configure so that you need to say "experimental-jpake" to enable
445 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
448 *) Add experimental JPAKE support, including demo authentication in
449 s_client and s_server.
452 *) Set the comparison function in v3_addr_canonize().
453 [Rob Austein <sra@hactrn.net>]
455 *) Add support for XMPP STARTTLS in s_client.
456 [Philip Paeps <philip@freebsd.org>]
458 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
459 to ensure that even with this option, only ciphersuites in the
460 server's preference list will be accepted. (Note that the option
461 applies only when resuming a session, so the earlier behavior was
462 just about the algorithm choice for symmetric cryptography.)
465 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
467 *) Fix NULL pointer dereference if a DTLS server received
468 ChangeCipherSpec as first record (CVE-2009-1386).
471 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
472 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
475 *) The fix in 0.9.8c that supposedly got rid of unsafe
476 double-checked locking was incomplete for RSA blinding,
477 addressing just one layer of what turns out to have been
478 doubly unsafe triple-checked locking.
480 So now fix this for real by retiring the MONT_HELPER macro
481 in crypto/rsa/rsa_eay.c.
483 [Bodo Moeller; problem pointed out by Marius Schilder]
485 *) Various precautionary measures:
487 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
489 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
490 (NB: This would require knowledge of the secret session ticket key
491 to exploit, in which case you'd be SOL either way.)
493 - Change bn_nist.c so that it will properly handle input BIGNUMs
494 outside the expected range.
496 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
499 [Neel Mehta, Bodo Moeller]
501 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
502 the load fails. Useful for distros.
503 [Ben Laurie and the FreeBSD team]
505 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
508 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
511 *) Expand ENGINE to support engine supplied SSL client certificate functions.
513 This work was sponsored by Logica.
516 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
517 keystores. Support for SSL/TLS client authentication too.
518 Not compiled unless enable-capieng specified to Configure.
520 This work was sponsored by Logica.
523 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
524 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
525 attribute creation routines such as certifcate requests and PKCS#12
529 Changes between 0.9.8g and 0.9.8h [28 May 2008]
531 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
532 handshake which could lead to a cilent crash as found using the
533 Codenomicon TLS test suite (CVE-2008-1672)
534 [Steve Henson, Mark Cox]
536 *) Fix double free in TLS server name extensions which could lead to
537 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
540 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
542 Clear the error queue to ensure that error entries left from
543 older function calls do not interfere with the correct operation.
544 [Lutz Jaenicke, Erik de Castro Lopo]
546 *) Remove root CA certificates of commercial CAs:
548 The OpenSSL project does not recommend any specific CA and does not
549 have any policy with respect to including or excluding any CA.
550 Therefore it does not make any sense to ship an arbitrary selection
551 of root CA certificates with the OpenSSL software.
554 *) RSA OAEP patches to fix two separate invalid memory reads.
555 The first one involves inputs when 'lzero' is greater than
556 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
557 before the beginning of from). The second one involves inputs where
558 the 'db' section contains nothing but zeroes (there is a one-byte
559 invalid read after the end of 'db').
560 [Ivan Nestlerode <inestlerode@us.ibm.com>]
562 *) Partial backport from 0.9.9-dev:
564 Introduce bn_mul_mont (dedicated Montgomery multiplication
565 procedure) as a candidate for BIGNUM assembler implementation.
566 While 0.9.9-dev uses assembler for various architectures, only
567 x86_64 is available by default here in the 0.9.8 branch, and
568 32-bit x86 is available through a compile-time setting.
570 To try the 32-bit x86 assembler implementation, use Configure
571 option "enable-montasm" (which exists only for this backport).
573 As "enable-montasm" for 32-bit x86 disclaims code stability
574 anyway, in this constellation we activate additional code
575 backported from 0.9.9-dev for further performance improvements,
576 namely BN_from_montgomery_word. (To enable this otherwise,
577 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
579 [Andy Polyakov (backport partially by Bodo Moeller)]
581 *) Add TLS session ticket callback. This allows an application to set
582 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
583 values. This is useful for key rollover for example where several key
584 sets may exist with different names.
587 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
588 This was broken until now in 0.9.8 releases, such that the only way
589 a registered ENGINE could be used (assuming it initialises
590 successfully on the host) was to explicitly set it as the default
591 for the relevant algorithms. This is in contradiction with 0.9.7
592 behaviour and the documentation. With this fix, when an ENGINE is
593 registered into a given algorithm's table of implementations, the
594 'uptodate' flag is reset so that auto-discovery will be used next
595 time a new context for that algorithm attempts to select an
597 [Ian Lister (tweaked by Geoff Thorpe)]
599 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
600 implemention in the following ways:
602 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
605 Lack of BER streaming support means one pass streaming processing is
606 only supported if data is detached: setting the streaming flag is
607 ignored for embedded content.
609 CMS support is disabled by default and must be explicitly enabled
610 with the enable-cms configuration option.
613 *) Update the GMP engine glue to do direct copies between BIGNUM and
614 mpz_t when openssl and GMP use the same limb size. Otherwise the
615 existing "conversion via a text string export" trick is still used.
616 [Paul Sheer <paulsheer@gmail.com>]
618 *) Zlib compression BIO. This is a filter BIO which compressed and
619 uncompresses any data passed through it.
622 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
623 RFC3394 compatible AES key wrapping.
626 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
627 sets string data without copying. X509_ALGOR_set0() and
628 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
629 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
630 from an X509_ATTRIBUTE structure optionally checking it occurs only
631 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
635 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
636 to get the expected BN_FLG_CONSTTIME behavior.
637 [Bodo Moeller (Google)]
641 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
642 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
643 - added some more tests to do_tests.pl
644 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
645 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
646 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
647 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
648 - various changes to netware.pl to enable gcc-cross builds on Win32
650 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
651 - various changes to fix missing prototype warnings
652 - fixed x86nasm.pl to create correct asm files for NASM COFF output
653 - added AES, WHIRLPOOL and CPUID assembler code to build files
654 - added missing AES assembler make rules to mk1mf.pl
655 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
656 [Guenter Knauf <eflash@gmx.net>]
658 *) Implement certificate status request TLS extension defined in RFC3546.
659 A client can set the appropriate parameters and receive the encoded
660 OCSP response via a callback. A server can query the supplied parameters
661 and set the encoded OCSP response in the callback. Add simplified examples
662 to s_client and s_server.
665 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
668 + Binary incompatibility of ssl_ctx_st structure
669 + DTLS interoperation with non-compliant servers
670 + Don't call get_session_cb() without proposed session
671 + Fix ia64 assembler code
672 [Andy Polyakov, Steve Henson]
674 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
676 *) DTLS Handshake overhaul. There were longstanding issues with
677 OpenSSL DTLS implementation, which were making it impossible for
678 RFC 4347 compliant client to communicate with OpenSSL server.
679 Unfortunately just fixing these incompatibilities would "cut off"
680 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
681 server keeps tolerating non RFC compliant syntax. The opposite is
682 not true, 0.9.8f client can not communicate with earlier server.
683 This update even addresses CVE-2007-4995.
686 *) Changes to avoid need for function casts in OpenSSL: some compilers
687 (gcc 4.2 and later) reject their use.
688 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
691 *) Add RFC4507 support to OpenSSL. This includes the corrections in
692 RFC4507bis. The encrypted ticket format is an encrypted encoded
693 SSL_SESSION structure, that way new session features are automatically
696 If a client application caches session in an SSL_SESSION structure
697 support is transparent because tickets are now stored in the encoded
700 The SSL_CTX structure automatically generates keys for ticket
701 protection in servers so again support should be possible
702 with no application modification.
704 If a client or server wishes to disable RFC4507 support then the option
705 SSL_OP_NO_TICKET can be set.
707 Add a TLS extension debugging callback to allow the contents of any client
708 or server extensions to be examined.
710 This work was sponsored by Google.
713 *) Add initial support for TLS extensions, specifically for the server_name
714 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
715 have new members for a host name. The SSL data structure has an
716 additional member SSL_CTX *initial_ctx so that new sessions can be
717 stored in that context to allow for session resumption, even after the
718 SSL has been switched to a new SSL_CTX in reaction to a client's
719 server_name extension.
721 New functions (subject to change):
724 SSL_get_servername_type()
727 New CTRL codes and macros (subject to change):
729 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
730 - SSL_CTX_set_tlsext_servername_callback()
731 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
732 - SSL_CTX_set_tlsext_servername_arg()
733 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
735 openssl s_client has a new '-servername ...' option.
737 openssl s_server has new options '-servername_host ...', '-cert2 ...',
738 '-key2 ...', '-servername_fatal' (subject to change). This allows
739 testing the HostName extension for a specific single host name ('-cert'
740 and '-key' remain fallbacks for handshakes without HostName
741 negotiation). If the unrecogninzed_name alert has to be sent, this by
742 default is a warning; it becomes fatal with the '-servername_fatal'
745 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
747 *) Add AES and SSE2 assembly language support to VC++ build.
750 *) Mitigate attack on final subtraction in Montgomery reduction.
753 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
754 (which previously caused an internal error).
757 *) Squeeze another 10% out of IGE mode when in != out.
760 *) AES IGE mode speedup.
761 [Dean Gaudet (Google)]
763 *) Add the Korean symmetric 128-bit cipher SEED (see
764 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
765 add SEED ciphersuites from RFC 4162:
767 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
768 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
769 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
770 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
772 To minimize changes between patchlevels in the OpenSSL 0.9.8
773 series, SEED remains excluded from compilation unless OpenSSL
774 is configured with 'enable-seed'.
777 *) Mitigate branch prediction attacks, which can be practical if a
778 single processor is shared, allowing a spy process to extract
779 information. For detailed background information, see
780 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
781 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
782 and Necessary Software Countermeasures"). The core of the change
783 are new versions BN_div_no_branch() and
784 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
785 respectively, which are slower, but avoid the security-relevant
786 conditional branches. These are automatically called by BN_div()
787 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
788 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
789 remove a conditional branch.
791 BN_FLG_CONSTTIME is the new name for the previous
792 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
793 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
794 in the exponent causes BN_mod_exp_mont() to use the alternative
795 implementation in BN_mod_exp_mont_consttime().) The old name
796 remains as a deprecated alias.
798 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
799 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
800 constant-time implementations for more than just exponentiation.
801 Here too the old name is kept as a deprecated alias.
803 BN_BLINDING_new() will now use BN_dup() for the modulus so that
804 the BN_BLINDING structure gets an independent copy of the
805 modulus. This means that the previous "BIGNUM *m" argument to
806 BN_BLINDING_new() and to BN_BLINDING_create_param() now
807 essentially becomes "const BIGNUM *m", although we can't actually
808 change this in the header file before 0.9.9. It allows
809 RSA_setup_blinding() to use BN_with_flags() on the modulus to
810 enable BN_FLG_CONSTTIME.
812 [Matthew D Wood (Intel Corp)]
814 *) In the SSL/TLS server implementation, be strict about session ID
815 context matching (which matters if an application uses a single
816 external cache for different purposes). Previously,
817 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
818 set. This did ensure strict client verification, but meant that,
819 with applications using a single external cache for quite
820 different requirements, clients could circumvent ciphersuite
821 restrictions for a given session ID context by starting a session
822 in a different context.
825 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
826 a ciphersuite string such as "DEFAULT:RSA" cannot enable
827 authentication-only ciphersuites.
830 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
831 not complete and could lead to a possible single byte overflow
832 (CVE-2007-5135) [Ben Laurie]
834 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
836 *) Since AES128 and AES256 (and similarly Camellia128 and
837 Camellia256) share a single mask bit in the logic of
838 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
839 kludge to work properly if AES128 is available and AES256 isn't
840 (or if Camellia128 is available and Camellia256 isn't).
843 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
844 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
845 When a point or a seed is encoded in a BIT STRING, we need to
846 prevent the removal of trailing zero bits to get the proper DER
847 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
848 of a NamedBitList, for which trailing 0 bits need to be removed.)
851 *) Have SSL/TLS server implementation tolerate "mismatched" record
852 protocol version while receiving ClientHello even if the
853 ClientHello is fragmented. (The server can't insist on the
854 particular protocol version it has chosen before the ServerHello
855 message has informed the client about his choice.)
858 *) Add RFC 3779 support.
859 [Rob Austein for ARIN, Ben Laurie]
861 *) Load error codes if they are not already present instead of using a
862 static variable. This allows them to be cleanly unloaded and reloaded.
863 Improve header file function name parsing.
866 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
867 or CAPABILITY handshake as required by RFCs.
870 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
872 *) Introduce limits to prevent malicious keys being able to
873 cause a denial of service. (CVE-2006-2940)
874 [Steve Henson, Bodo Moeller]
876 *) Fix ASN.1 parsing of certain invalid structures that can result
877 in a denial of service. (CVE-2006-2937) [Steve Henson]
879 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
880 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
882 *) Fix SSL client code which could crash if connecting to a
883 malicious SSLv2 server. (CVE-2006-4343)
884 [Tavis Ormandy and Will Drewry, Google Security Team]
886 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
887 match only those. Before that, "AES256-SHA" would be interpreted
888 as a pattern and match "AES128-SHA" too (since AES128-SHA got
889 the same strength classification in 0.9.7h) as we currently only
890 have a single AES bit in the ciphersuite description bitmap.
891 That change, however, also applied to ciphersuite strings such as
892 "RC4-MD5" that intentionally matched multiple ciphersuites --
893 namely, SSL 2.0 ciphersuites in addition to the more common ones
894 from SSL 3.0/TLS 1.0.
896 So we change the selection algorithm again: Naming an explicit
897 ciphersuite selects this one ciphersuite, and any other similar
898 ciphersuite (same bitmap) from *other* protocol versions.
899 Thus, "RC4-MD5" again will properly select both the SSL 2.0
900 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
902 Since SSL 2.0 does not have any ciphersuites for which the
903 128/256 bit distinction would be relevant, this works for now.
904 The proper fix will be to use different bits for AES128 and
905 AES256, which would have avoided the problems from the beginning;
906 however, bits are scarce, so we can only do this in a new release
907 (not just a patchlevel) when we can change the SSL_CIPHER
908 definition to split the single 'unsigned long mask' bitmap into
909 multiple values to extend the available space.
913 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
915 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
916 (CVE-2006-4339) [Ben Laurie and Google Security Team]
918 *) Add AES IGE and biIGE modes.
921 *) Change the Unix randomness entropy gathering to use poll() when
922 possible instead of select(), since the latter has some
923 undesirable limitations.
924 [Darryl Miles via Richard Levitte and Bodo Moeller]
926 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
927 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
928 cannot be implicitly activated as part of, e.g., the "AES" alias.
929 However, please upgrade to OpenSSL 0.9.9[-dev] for
930 non-experimental use of the ECC ciphersuites to get TLS extension
931 support, which is required for curve and point format negotiation
932 to avoid potential handshake problems.
935 *) Disable rogue ciphersuites:
937 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
938 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
939 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
941 The latter two were purportedly from
942 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
945 Also deactivate the remaining ciphersuites from
946 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
947 unofficial, and the ID has long expired.
950 *) Fix RSA blinding Heisenbug (problems sometimes occured on
951 dual-core machines) and other potential thread-safety issues.
954 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
955 versions), which is now available for royalty-free use
956 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
957 Also, add Camellia TLS ciphersuites from RFC 4132.
959 To minimize changes between patchlevels in the OpenSSL 0.9.8
960 series, Camellia remains excluded from compilation unless OpenSSL
961 is configured with 'enable-camellia'.
964 *) Disable the padding bug check when compression is in use. The padding
965 bug check assumes the first packet is of even length, this is not
966 necessarily true if compresssion is enabled and can result in false
967 positives causing handshake failure. The actual bug test is ancient
968 code so it is hoped that implementations will either have fixed it by
969 now or any which still have the bug do not support compression.
972 Changes between 0.9.8a and 0.9.8b [04 May 2006]
974 *) When applying a cipher rule check to see if string match is an explicit
975 cipher suite and only match that one cipher suite if it is.
978 *) Link in manifests for VC++ if needed.
979 [Austin Ziegler <halostatue@gmail.com>]
981 *) Update support for ECC-based TLS ciphersuites according to
982 draft-ietf-tls-ecc-12.txt with proposed changes (but without
983 TLS extensions, which are supported starting with the 0.9.9
984 branch, not in the OpenSSL 0.9.8 branch).
987 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
988 opaque EVP_CIPHER_CTX handling.
991 *) Fixes and enhancements to zlib compression code. We now only use
992 "zlib1.dll" and use the default __cdecl calling convention on Win32
993 to conform with the standards mentioned here:
994 http://www.zlib.net/DLL_FAQ.txt
995 Static zlib linking now works on Windows and the new --with-zlib-include
996 --with-zlib-lib options to Configure can be used to supply the location
997 of the headers and library. Gracefully handle case where zlib library
1001 *) Several fixes and enhancements to the OID generation code. The old code
1002 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
1003 handle numbers larger than ULONG_MAX, truncated printing and had a
1004 non standard OBJ_obj2txt() behaviour.
1007 *) Add support for building of engines under engine/ as shared libraries
1008 under VC++ build system.
1011 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
1012 Hopefully, we will not see any false combination of paths any more.
1015 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
1017 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1018 (part of SSL_OP_ALL). This option used to disable the
1019 countermeasure against man-in-the-middle protocol-version
1020 rollback in the SSL 2.0 server implementation, which is a bad
1021 idea. (CVE-2005-2969)
1023 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1024 for Information Security, National Institute of Advanced Industrial
1025 Science and Technology [AIST], Japan)]
1027 *) Add two function to clear and return the verify parameter flags.
1030 *) Keep cipherlists sorted in the source instead of sorting them at
1031 runtime, thus removing the need for a lock.
1034 *) Avoid some small subgroup attacks in Diffie-Hellman.
1035 [Nick Mathewson and Ben Laurie]
1037 *) Add functions for well-known primes.
1040 *) Extended Windows CE support.
1041 [Satoshi Nakamura and Andy Polyakov]
1043 *) Initialize SSL_METHOD structures at compile time instead of during
1044 runtime, thus removing the need for a lock.
1047 *) Make PKCS7_decrypt() work even if no certificate is supplied by
1048 attempting to decrypt each encrypted key in turn. Add support to
1052 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
1054 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
1057 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
1060 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
1061 key into the same file any more.
1064 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
1067 *) Add -utf8 command line and config file option to 'ca'.
1068 [Stefan <stf@udoma.org]
1070 *) Removed the macro des_crypt(), as it seems to conflict with some
1071 libraries. Use DES_crypt().
1074 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
1075 involves renaming the source and generated shared-libs for
1076 both. The engines will accept the corrected or legacy ids
1077 ('ncipher' and '4758_cca' respectively) when binding. NB,
1078 this only applies when building 'shared'.
1079 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
1081 *) Add attribute functions to EVP_PKEY structure. Modify
1082 PKCS12_create() to recognize a CSP name attribute and
1083 use it. Make -CSP option work again in pkcs12 utility.
1086 *) Add new functionality to the bn blinding code:
1087 - automatic re-creation of the BN_BLINDING parameters after
1088 a fixed number of uses (currently 32)
1089 - add new function for parameter creation
1090 - introduce flags to control the update behaviour of the
1091 BN_BLINDING parameters
1092 - hide BN_BLINDING structure
1093 Add a second BN_BLINDING slot to the RSA structure to improve
1094 performance when a single RSA object is shared among several
1098 *) Add support for DTLS.
1099 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
1101 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
1102 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
1105 *) Remove buggy and incompletet DH cert support from
1106 ssl/ssl_rsa.c and ssl/s3_both.c
1109 *) Use SHA-1 instead of MD5 as the default digest algorithm for
1110 the apps/openssl applications.
1113 *) Compile clean with "-Wall -Wmissing-prototypes
1114 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
1115 DEBUG_SAFESTACK must also be set.
1118 *) Change ./Configure so that certain algorithms can be disabled by default.
1119 The new counterpiece to "no-xxx" is "enable-xxx".
1121 The patented RC5 and MDC2 algorithms will now be disabled unless
1122 "enable-rc5" and "enable-mdc2", respectively, are specified.
1124 (IDEA remains enabled despite being patented. This is because IDEA
1125 is frequently required for interoperability, and there is no license
1126 fee for non-commercial use. As before, "no-idea" can be used to
1127 avoid this algorithm.)
1131 *) Add processing of proxy certificates (see RFC 3820). This work was
1132 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
1133 EGEE (Enabling Grids for E-science in Europe).
1136 *) RC4 performance overhaul on modern architectures/implementations, such
1137 as Intel P4, IA-64 and AMD64.
1140 *) New utility extract-section.pl. This can be used specify an alternative
1141 section number in a pod file instead of having to treat each file as
1142 a separate case in Makefile. This can be done by adding two lines to the
1145 =for comment openssl_section:XXX
1147 The blank line is mandatory.
1151 *) New arguments -certform, -keyform and -pass for s_client and s_server
1152 to allow alternative format key and certificate files and passphrase
1156 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
1157 update associated structures and add various utility functions.
1159 Add new policy related verify parameters, include policy checking in
1160 standard verify code. Enhance 'smime' application with extra parameters
1161 to support policy checking and print out.
1164 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
1165 Nehemiah processors. These extensions support AES encryption in hardware
1166 as well as RNG (though RNG support is currently disabled).
1167 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
1169 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
1172 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
1173 [Andy Polyakov and a number of other people]
1175 *) Improved PowerPC platform support. Most notably BIGNUM assembler
1176 implementation contributed by IBM.
1177 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
1179 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
1180 exponent rather than 'unsigned long'. There is a corresponding change to
1181 the new 'rsa_keygen' element of the RSA_METHOD structure.
1182 [Jelte Jansen, Geoff Thorpe]
1184 *) Functionality for creating the initial serial number file is now
1185 moved from CA.pl to the 'ca' utility with a new option -create_serial.
1187 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
1188 number file to 1, which is bound to cause problems. To avoid
1189 the problems while respecting compatibility between different 0.9.7
1190 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
1191 CA.pl for serial number initialization. With the new release 0.9.8,
1192 we can fix the problem directly in the 'ca' utility.)
1195 *) Reduced header interdepencies by declaring more opaque objects in
1196 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
1197 give fewer recursive includes, which could break lazy source code - so
1198 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
1199 developers should define this symbol when building and using openssl to
1200 ensure they track the recommended behaviour, interfaces, [etc], but
1201 backwards-compatible behaviour prevails when this isn't defined.
1204 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
1207 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
1208 This will generate a random key of the appropriate length based on the
1209 cipher context. The EVP_CIPHER can provide its own random key generation
1210 routine to support keys of a specific form. This is used in the des and
1211 3des routines to generate a key of the correct parity. Update S/MIME
1212 code to use new functions and hence generate correct parity DES keys.
1213 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
1214 valid (weak or incorrect parity).
1217 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
1218 as looking them up. This is useful when the verified structure may contain
1219 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
1220 present unless the new PKCS7_NO_CRL flag is asserted.
1223 *) Extend ASN1 oid configuration module. It now additionally accepts the
1226 shortName = some long name, 1.2.3.4
1229 *) Reimplemented the BN_CTX implementation. There is now no more static
1230 limitation on the number of variables it can handle nor the depth of the
1231 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
1232 information can now expand as required, and rather than having a single
1233 static array of bignums, BN_CTX now uses a linked-list of such arrays
1234 allowing it to expand on demand whilst maintaining the usefulness of
1235 BN_CTX's "bundling".
1238 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
1239 to allow all RSA operations to function using a single BN_CTX.
1242 *) Preliminary support for certificate policy evaluation and checking. This
1243 is initially intended to pass the tests outlined in "Conformance Testing
1244 of Relying Party Client Certificate Path Processing Logic" v1.07.
1247 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
1248 remained unused and not that useful. A variety of other little bignum
1249 tweaks and fixes have also been made continuing on from the audit (see
1253 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
1254 associated ASN1, EVP and SSL functions and old ASN1 macros.
1257 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
1258 and this should never fail. So the return value from the use of
1259 BN_set_word() (which can fail due to needless expansion) is now deprecated;
1260 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
1263 *) BN_CTX_get() should return zero-valued bignums, providing the same
1264 initialised value as BN_new().
1265 [Geoff Thorpe, suggested by Ulf Möller]
1267 *) Support for inhibitAnyPolicy certificate extension.
1270 *) An audit of the BIGNUM code is underway, for which debugging code is
1271 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
1272 is considered valid when processing BIGNUMs, and causes execution to
1273 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
1274 further steps are taken to deliberately pollute unused data in BIGNUM
1275 structures to try and expose faulty code further on. For now, openssl will
1276 (in its default mode of operation) continue to tolerate the inconsistent
1277 forms that it has tolerated in the past, but authors and packagers should
1278 consider trying openssl and their own applications when compiled with
1279 these debugging symbols defined. It will help highlight potential bugs in
1280 their own code, and will improve the test coverage for OpenSSL itself. At
1281 some point, these tighter rules will become openssl's default to improve
1282 maintainability, though the assert()s and other overheads will remain only
1283 in debugging configurations. See bn.h for more details.
1284 [Geoff Thorpe, Nils Larsch, Ulf Möller]
1286 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
1287 that can only be obtained through BN_CTX_new() (which implicitly
1288 initialises it). The presence of this function only made it possible
1289 to overwrite an existing structure (and cause memory leaks).
1292 *) Because of the callback-based approach for implementing LHASH as a
1293 template type, lh_insert() adds opaque objects to hash-tables and
1294 lh_doall() or lh_doall_arg() are typically used with a destructor callback
1295 to clean up those corresponding objects before destroying the hash table
1296 (and losing the object pointers). So some over-zealous constifications in
1297 LHASH have been relaxed so that lh_insert() does not take (nor store) the
1298 objects as "const" and the lh_doall[_arg] callback wrappers are not
1299 prototyped to have "const" restrictions on the object pointers they are
1300 given (and so aren't required to cast them away any more).
1303 *) The tmdiff.h API was so ugly and minimal that our own timing utility
1304 (speed) prefers to use its own implementation. The two implementations
1305 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
1306 its object type properly exposed (MS_TM) instead of casting to/from "char
1307 *". This may still change yet if someone realises MS_TM and "ms_time_***"
1308 aren't necessarily the greatest nomenclatures - but this is what was used
1309 internally to the implementation so I've used that for now.
1312 *) Ensure that deprecated functions do not get compiled when
1313 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
1314 the self-tests were still using deprecated key-generation functions so
1315 these have been updated also.
1318 *) Reorganise PKCS#7 code to separate the digest location functionality
1319 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
1320 New function PKCS7_set_digest() to set the digest type for PKCS#7
1321 digestedData type. Add additional code to correctly generate the
1322 digestedData type and add support for this type in PKCS7 initialization
1326 *) New function PKCS7_set0_type_other() this initializes a PKCS7
1327 structure of type "other".
1330 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
1331 sure the loop does correctly stop and breaking ("division by zero")
1332 modulus operations are not performed. The (pre-generated) prime
1333 table crypto/bn/bn_prime.h was already correct, but it could not be
1334 re-generated on some platforms because of the "division by zero"
1335 situation in the script.
1336 [Ralf S. Engelschall]
1338 *) Update support for ECC-based TLS ciphersuites according to
1339 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
1340 SHA-1 now is only used for "small" curves (where the
1341 representation of a field element takes up to 24 bytes); for
1342 larger curves, the field element resulting from ECDH is directly
1343 used as premaster secret.
1344 [Douglas Stebila (Sun Microsystems Laboratories)]
1346 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
1347 curve secp160r1 to the tests.
1348 [Douglas Stebila (Sun Microsystems Laboratories)]
1350 *) Add the possibility to load symbols globally with DSO.
1351 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
1353 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
1354 control of the error stack.
1357 *) Add support for STORE in ENGINE.
1360 *) Add the STORE type. The intention is to provide a common interface
1361 to certificate and key stores, be they simple file-based stores, or
1362 HSM-type store, or LDAP stores, or...
1363 NOTE: The code is currently UNTESTED and isn't really used anywhere.
1366 *) Add a generic structure called OPENSSL_ITEM. This can be used to
1367 pass a list of arguments to any function as well as provide a way
1368 for a function to pass data back to the caller.
1371 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
1372 works like BUF_strdup() but can be used to duplicate a portion of
1373 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
1377 *) Add the function sk_find_ex() which works like sk_find(), but will
1378 return an index to an element even if an exact match couldn't be
1379 found. The index is guaranteed to point at the element where the
1380 searched-for key would be inserted to preserve sorting order.
1383 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
1384 takes an extra flags argument for optional functionality. Currently,
1385 the following flags are defined:
1387 OBJ_BSEARCH_VALUE_ON_NOMATCH
1388 This one gets OBJ_bsearch_ex() to return a pointer to the first
1389 element where the comparing function returns a negative or zero
1392 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
1393 This one gets OBJ_bsearch_ex() to return a pointer to the first
1394 element where the comparing function returns zero. This is useful
1395 if there are more than one element where the comparing function
1399 *) Make it possible to create self-signed certificates with 'openssl ca'
1400 in such a way that the self-signed certificate becomes part of the
1401 CA database and uses the same mechanisms for serial number generation
1402 as all other certificate signing. The new flag '-selfsign' enables
1403 this functionality. Adapt CA.sh and CA.pl.in.
1406 *) Add functionality to check the public key of a certificate request
1407 against a given private. This is useful to check that a certificate
1408 request can be signed by that key (self-signing).
1411 *) Make it possible to have multiple active certificates with the same
1412 subject in the CA index file. This is done only if the keyword
1413 'unique_subject' is set to 'no' in the main CA section (default
1414 if 'CA_default') of the configuration file. The value is saved
1415 with the database itself in a separate index attribute file,
1416 named like the index file with '.attr' appended to the name.
1419 *) Generate muti valued AVAs using '+' notation in config files for
1423 *) Support for nameConstraints certificate extension.
1426 *) Support for policyConstraints certificate extension.
1429 *) Support for policyMappings certificate extension.
1432 *) Make sure the default DSA_METHOD implementation only uses its
1433 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
1434 and change its own handlers to be NULL so as to remove unnecessary
1435 indirection. This lets alternative implementations fallback to the
1436 default implementation more easily.
1439 *) Support for directoryName in GeneralName related extensions
1443 *) Make it possible to link applications using Makefile.shared.
1444 Make that possible even when linking against static libraries!
1447 *) Support for single pass processing for S/MIME signing. This now
1448 means that S/MIME signing can be done from a pipe, in addition
1449 cleartext signing (multipart/signed type) is effectively streaming
1450 and the signed data does not need to be all held in memory.
1452 This is done with a new flag PKCS7_STREAM. When this flag is set
1453 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
1454 is done after the data is output (and digests calculated) in
1455 SMIME_write_PKCS7().
1458 *) Add full support for -rpath/-R, both in shared libraries and
1459 applications, at least on the platforms where it's known how
1463 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
1464 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
1465 will now compute a table of multiples of the generator that
1466 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
1467 faster (notably in the case of a single point multiplication,
1468 scalar * generator).
1469 [Nils Larsch, Bodo Moeller]
1471 *) IPv6 support for certificate extensions. The various extensions
1472 which use the IP:a.b.c.d can now take IPv6 addresses using the
1473 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
1477 *) Added an ENGINE that implements RSA by performing private key
1478 exponentiations with the GMP library. The conversions to and from
1479 GMP's mpz_t format aren't optimised nor are any montgomery forms
1480 cached, and on x86 it appears OpenSSL's own performance has caught up.
1481 However there are likely to be other architectures where GMP could
1482 provide a boost. This ENGINE is not built in by default, but it can be
1483 specified at Configure time and should be accompanied by the necessary
1484 linker additions, eg;
1485 ./config -DOPENSSL_USE_GMP -lgmp
1488 *) "openssl engine" will not display ENGINE/DSO load failure errors when
1489 testing availability of engines with "-t" - the old behaviour is
1490 produced by increasing the feature's verbosity with "-tt".
1493 *) ECDSA routines: under certain error conditions uninitialized BN objects
1494 could be freed. Solution: make sure initialization is performed early
1495 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
1499 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
1500 and DH_METHOD (eg. by ENGINE implementations) to override the normal
1501 software implementations. For DSA and DH, parameter generation can
1502 also be overriden by providing the appropriate method callbacks.
1505 *) Change the "progress" mechanism used in key-generation and
1506 primality testing to functions that take a new BN_GENCB pointer in
1507 place of callback/argument pairs. The new API functions have "_ex"
1508 postfixes and the older functions are reimplemented as wrappers for
1509 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
1510 declarations of the old functions to help (graceful) attempts to
1511 migrate to the new functions. Also, the new key-generation API
1512 functions operate on a caller-supplied key-structure and return
1513 success/failure rather than returning a key or NULL - this is to
1514 help make "keygen" another member function of RSA_METHOD etc.
1516 Example for using the new callback interface:
1518 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
1522 BN_GENCB_set(&my_cb, my_callback, my_arg);
1524 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
1525 /* For the meaning of a, b in calls to my_callback(), see the
1526 * documentation of the function that calls the callback.
1527 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
1528 * my_callback should return 1 if it wants BN_is_prime_ex()
1529 * to continue, or 0 to stop.
1534 *) Change the ZLIB compression method to be stateful, and make it
1535 available to TLS with the number defined in
1536 draft-ietf-tls-compression-04.txt.
1539 *) Add the ASN.1 structures and functions for CertificatePair, which
1540 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
1542 CertificatePair ::= SEQUENCE {
1543 forward [0] Certificate OPTIONAL,
1544 reverse [1] Certificate OPTIONAL,
1545 -- at least one of the pair shall be present -- }
1547 Also implement the PEM functions to read and write certificate
1548 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
1550 This needed to be defined, mostly for the sake of the LDAP
1551 attribute crossCertificatePair, but may prove useful elsewhere as
1555 *) Make it possible to inhibit symlinking of shared libraries in
1556 Makefile.shared, for Cygwin's sake.
1559 *) Extend the BIGNUM API by creating a function
1560 void BN_set_negative(BIGNUM *a, int neg);
1561 and a macro that behave like
1562 int BN_is_negative(const BIGNUM *a);
1564 to avoid the need to access 'a->neg' directly in applications.
1567 *) Implement fast modular reduction for pseudo-Mersenne primes
1568 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
1569 EC_GROUP_new_curve_GFp() will now automatically use this
1571 [Nils Larsch <nla@trustcenter.de>]
1573 *) Add new lock type (CRYPTO_LOCK_BN).
1576 *) Change the ENGINE framework to automatically load engines
1577 dynamically from specific directories unless they could be
1578 found to already be built in or loaded. Move all the
1579 current engines except for the cryptodev one to a new
1581 The engines in engines/ are built as shared libraries if
1582 the "shared" options was given to ./Configure or ./config.
1583 Otherwise, they are inserted in libcrypto.a.
1584 /usr/local/ssl/engines is the default directory for dynamic
1585 engines, but that can be overriden at configure time through
1586 the usual use of --prefix and/or --openssldir, and at run
1587 time with the environment variable OPENSSL_ENGINES.
1588 [Geoff Thorpe and Richard Levitte]
1590 *) Add Makefile.shared, a helper makefile to build shared
1591 libraries. Addapt Makefile.org.
1594 *) Add version info to Win32 DLLs.
1595 [Peter 'Luna' Runestig" <peter@runestig.com>]
1597 *) Add new 'medium level' PKCS#12 API. Certificates and keys
1598 can be added using this API to created arbitrary PKCS#12
1599 files while avoiding the low level API.
1601 New options to PKCS12_create(), key or cert can be NULL and
1602 will then be omitted from the output file. The encryption
1603 algorithm NIDs can be set to -1 for no encryption, the mac
1604 iteration count can be set to 0 to omit the mac.
1606 Enhance pkcs12 utility by making the -nokeys and -nocerts
1607 options work when creating a PKCS#12 file. New option -nomac
1608 to omit the mac, NONE can be set for an encryption algorithm.
1609 New code is modified to use the enhanced PKCS12_create()
1610 instead of the low level API.
1613 *) Extend ASN1 encoder to support indefinite length constructed
1614 encoding. This can output sequences tags and octet strings in
1615 this form. Modify pk7_asn1.c to support indefinite length
1616 encoding. This is experimental and needs additional code to
1617 be useful, such as an ASN1 bio and some enhanced streaming
1620 Extend template encode functionality so that tagging is passed
1621 down to the template encoder.
1624 *) Let 'openssl req' fail if an argument to '-newkey' is not
1625 recognized instead of using RSA as a default.
1628 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
1629 As these are not official, they are not included in "ALL";
1630 the "ECCdraft" ciphersuite group alias can be used to select them.
1631 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
1633 *) Add ECDH engine support.
1634 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
1636 *) Add ECDH in new directory crypto/ecdh/.
1637 [Douglas Stebila (Sun Microsystems Laboratories)]
1639 *) Let BN_rand_range() abort with an error after 100 iterations
1640 without success (which indicates a broken PRNG).
1643 *) Change BN_mod_sqrt() so that it verifies that the input value
1644 is really the square of the return value. (Previously,
1645 BN_mod_sqrt would show GIGO behaviour.)
1648 *) Add named elliptic curves over binary fields from X9.62, SECG,
1649 and WAP/WTLS; add OIDs that were still missing.
1651 [Sheueling Chang Shantz and Douglas Stebila
1652 (Sun Microsystems Laboratories)]
1654 *) Extend the EC library for elliptic curves over binary fields
1655 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
1658 EC_GF2m_simple_method
1662 EC_GROUP_new_curve_GF2m
1663 EC_GROUP_set_curve_GF2m
1664 EC_GROUP_get_curve_GF2m
1665 EC_POINT_set_affine_coordinates_GF2m
1666 EC_POINT_get_affine_coordinates_GF2m
1667 EC_POINT_set_compressed_coordinates_GF2m
1669 Point compression for binary fields is disabled by default for
1670 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
1673 As binary polynomials are represented as BIGNUMs, various members
1674 of the EC_GROUP and EC_POINT data structures can be shared
1675 between the implementations for prime fields and binary fields;
1676 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
1677 are essentially identical to their ..._GFp counterparts.
1678 (For simplicity, the '..._GFp' prefix has been dropped from
1679 various internal method names.)
1681 An internal 'field_div' method (similar to 'field_mul' and
1682 'field_sqr') has been added; this is used only for binary fields.
1684 [Sheueling Chang Shantz and Douglas Stebila
1685 (Sun Microsystems Laboratories)]
1687 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
1688 through methods ('mul', 'precompute_mult').
1690 The generic implementations (now internally called 'ec_wNAF_mul'
1691 and 'ec_wNAF_precomputed_mult') remain the default if these
1692 methods are undefined.
1694 [Sheueling Chang Shantz and Douglas Stebila
1695 (Sun Microsystems Laboratories)]
1697 *) New function EC_GROUP_get_degree, which is defined through
1698 EC_METHOD. For curves over prime fields, this returns the bit
1699 length of the modulus.
1701 [Sheueling Chang Shantz and Douglas Stebila
1702 (Sun Microsystems Laboratories)]
1704 *) New functions EC_GROUP_dup, EC_POINT_dup.
1705 (These simply call ..._new and ..._copy).
1707 [Sheueling Chang Shantz and Douglas Stebila
1708 (Sun Microsystems Laboratories)]
1710 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
1711 Polynomials are represented as BIGNUMs (where the sign bit is not
1712 used) in the following functions [macros]:
1715 BN_GF2m_sub [= BN_GF2m_add]
1716 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
1717 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
1718 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
1720 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
1721 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
1722 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
1723 BN_GF2m_cmp [= BN_ucmp]
1725 (Note that only the 'mod' functions are actually for fields GF(2^m).
1726 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
1728 For some functions, an the irreducible polynomial defining a
1729 field can be given as an 'unsigned int[]' with strictly
1730 decreasing elements giving the indices of those bits that are set;
1731 i.e., p[] represents the polynomial
1732 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
1734 p[0] > p[1] > ... > p[k] = 0.
1735 This applies to the following functions:
1740 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
1741 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
1743 BN_GF2m_mod_sqrt_arr
1744 BN_GF2m_mod_solve_quad_arr
1748 Conversion can be performed by the following functions:
1753 bntest.c has additional tests for binary polynomial arithmetic.
1755 Two implementations for BN_GF2m_mod_div() are available.
1756 The default algorithm simply uses BN_GF2m_mod_inv() and
1757 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
1758 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
1759 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
1761 [Sheueling Chang Shantz and Douglas Stebila
1762 (Sun Microsystems Laboratories)]
1764 *) Add new error code 'ERR_R_DISABLED' that can be used when some
1765 functionality is disabled at compile-time.
1766 [Douglas Stebila <douglas.stebila@sun.com>]
1768 *) Change default behaviour of 'openssl asn1parse' so that more
1769 information is visible when viewing, e.g., a certificate:
1771 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
1772 mode the content of non-printable OCTET STRINGs is output in a
1773 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
1774 avoid the appearance of a printable string.
1775 [Nils Larsch <nla@trustcenter.de>]
1777 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
1779 EC_GROUP_set_asn1_flag()
1780 EC_GROUP_get_asn1_flag()
1781 EC_GROUP_set_point_conversion_form()
1782 EC_GROUP_get_point_conversion_form()
1783 These control ASN1 encoding details:
1784 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
1785 has been set to OPENSSL_EC_NAMED_CURVE.
1786 - Points are encoded in uncompressed form by default; options for
1787 asn1_for are as for point2oct, namely
1788 POINT_CONVERSION_COMPRESSED
1789 POINT_CONVERSION_UNCOMPRESSED
1790 POINT_CONVERSION_HYBRID
1792 Also add 'seed' and 'seed_len' members to EC_GROUP with access
1795 EC_GROUP_get0_seed()
1796 EC_GROUP_get_seed_len()
1797 This is used only for ASN1 purposes (so far).
1798 [Nils Larsch <nla@trustcenter.de>]
1800 *) Add 'field_type' member to EC_METHOD, which holds the NID
1801 of the appropriate field type OID. The new function
1802 EC_METHOD_get_field_type() returns this value.
1803 [Nils Larsch <nla@trustcenter.de>]
1808 EC_POINT_point2hex()
1809 EC_POINT_hex2point()
1810 providing useful interfaces to EC_POINT_point2oct() and
1811 EC_POINT_oct2point().
1812 [Nils Larsch <nla@trustcenter.de>]
1814 *) Change internals of the EC library so that the functions
1815 EC_GROUP_set_generator()
1816 EC_GROUP_get_generator()
1817 EC_GROUP_get_order()
1818 EC_GROUP_get_cofactor()
1819 are implemented directly in crypto/ec/ec_lib.c and not dispatched
1820 to methods, which would lead to unnecessary code duplication when
1821 adding different types of curves.
1822 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
1824 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
1825 arithmetic, and such that modified wNAFs are generated
1826 (which avoid length expansion in many cases).
1829 *) Add a function EC_GROUP_check_discriminant() (defined via
1830 EC_METHOD) that verifies that the curve discriminant is non-zero.
1832 Add a function EC_GROUP_check() that makes some sanity tests
1833 on a EC_GROUP, its generator and order. This includes
1834 EC_GROUP_check_discriminant().
1835 [Nils Larsch <nla@trustcenter.de>]
1837 *) Add ECDSA in new directory crypto/ecdsa/.
1839 Add applications 'openssl ecparam' and 'openssl ecdsa'
1840 (these are based on 'openssl dsaparam' and 'openssl dsa').
1842 ECDSA support is also included in various other files across the
1843 library. Most notably,
1844 - 'openssl req' now has a '-newkey ecdsa:file' option;
1845 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
1846 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
1847 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
1848 them suitable for ECDSA where domain parameters must be
1849 extracted before the specific public key;
1850 - ECDSA engine support has been added.
1851 [Nils Larsch <nla@trustcenter.de>]
1853 *) Include some named elliptic curves, and add OIDs from X9.62,
1854 SECG, and WAP/WTLS. Each curve can be obtained from the new
1856 EC_GROUP_new_by_curve_name(),
1857 and the list of available named curves can be obtained with
1858 EC_get_builtin_curves().
1859 Also add a 'curve_name' member to EC_GROUP objects, which can be
1861 EC_GROUP_set_curve_name()
1862 EC_GROUP_get_curve_name()
1863 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
1865 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1866 was actually never needed) and in BN_mul(). The removal in BN_mul()
1867 required a small change in bn_mul_part_recursive() and the addition
1868 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1869 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1870 bn_sub_words() and bn_add_words() except they take arrays with
1874 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
1876 *) Cleanse PEM buffers before freeing them since they may contain
1878 [Benjamin Bennett <ben@psc.edu>]
1880 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1881 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1882 authentication-only ciphersuites.
1885 *) Since AES128 and AES256 share a single mask bit in the logic of
1886 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1887 kludge to work properly if AES128 is available and AES256 isn't.
1890 *) Expand security boundary to match 1.1.1 module.
1893 *) Remove redundant features: hash file source, editing of test vectors
1894 modify fipsld to use external fips_premain.c signature.
1897 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
1898 run algorithm test programs.
1901 *) Make algorithm test programs more tolerant of whitespace.
1904 *) Have SSL/TLS server implementation tolerate "mismatched" record
1905 protocol version while receiving ClientHello even if the
1906 ClientHello is fragmented. (The server can't insist on the
1907 particular protocol version it has chosen before the ServerHello
1908 message has informed the client about his choice.)
1911 *) Load error codes if they are not already present instead of using a
1912 static variable. This allows them to be cleanly unloaded and reloaded.
1915 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
1917 *) Introduce limits to prevent malicious keys being able to
1918 cause a denial of service. (CVE-2006-2940)
1919 [Steve Henson, Bodo Moeller]
1921 *) Fix ASN.1 parsing of certain invalid structures that can result
1922 in a denial of service. (CVE-2006-2937) [Steve Henson]
1924 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1925 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1927 *) Fix SSL client code which could crash if connecting to a
1928 malicious SSLv2 server. (CVE-2006-4343)
1929 [Tavis Ormandy and Will Drewry, Google Security Team]
1931 *) Change ciphersuite string processing so that an explicit
1932 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
1933 will no longer include "AES128-SHA"), and any other similar
1934 ciphersuite (same bitmap) from *other* protocol versions (so that
1935 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
1936 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
1937 changes from 0.9.8b and 0.9.8d.
1940 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
1942 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1943 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1945 *) Change the Unix randomness entropy gathering to use poll() when
1946 possible instead of select(), since the latter has some
1947 undesirable limitations.
1948 [Darryl Miles via Richard Levitte and Bodo Moeller]
1950 *) Disable rogue ciphersuites:
1952 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1953 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1954 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1956 The latter two were purportedly from
1957 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1960 Also deactive the remaining ciphersuites from
1961 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1962 unofficial, and the ID has long expired.
1965 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1966 dual-core machines) and other potential thread-safety issues.
1969 Changes between 0.9.7i and 0.9.7j [04 May 2006]
1971 *) Adapt fipsld and the build system to link against the validated FIPS
1972 module in FIPS mode.
1975 *) Fixes for VC++ 2005 build under Windows.
1978 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
1979 from a Windows bash shell such as MSYS. It is autodetected from the
1980 "config" script when run from a VC++ environment. Modify standard VC++
1981 build to use fipscanister.o from the GNU make build.
1984 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
1986 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
1987 The value now differs depending on if you build for FIPS or not.
1988 BEWARE! A program linked with a shared FIPSed libcrypto can't be
1989 safely run with a non-FIPSed libcrypto, as it may crash because of
1990 the difference induced by this change.
1993 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
1995 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1996 (part of SSL_OP_ALL). This option used to disable the
1997 countermeasure against man-in-the-middle protocol-version
1998 rollback in the SSL 2.0 server implementation, which is a bad
1999 idea. (CVE-2005-2969)
2001 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2002 for Information Security, National Institute of Advanced Industrial
2003 Science and Technology [AIST], Japan)]
2005 *) Minimal support for X9.31 signatures and PSS padding modes. This is
2006 mainly for FIPS compliance and not fully integrated at this stage.
2009 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
2010 the exponentiation using a fixed-length exponent. (Otherwise,
2011 the information leaked through timing could expose the secret key
2012 after many signatures; cf. Bleichenbacher's attack on DSA with
2016 *) Make a new fixed-window mod_exp implementation the default for
2017 RSA, DSA, and DH private-key operations so that the sequence of
2018 squares and multiplies and the memory access pattern are
2019 independent of the particular secret key. This will mitigate
2020 cache-timing and potential related attacks.
2022 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
2023 and this is automatically used by BN_mod_exp_mont() if the new flag
2024 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
2025 will use this BN flag for private exponents unless the flag
2026 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
2027 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
2029 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
2031 *) Change the client implementation for SSLv23_method() and
2032 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
2033 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
2034 (Previously, the SSL 2.0 backwards compatible Client Hello
2035 message format would be used even with SSL_OP_NO_SSLv2.)
2038 *) Add support for smime-type MIME parameter in S/MIME messages which some
2042 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
2043 a threadsafe manner. Modify rsa code to use new function and add calls
2044 to dsa and dh code (which had race conditions before).
2047 *) Include the fixed error library code in the C error file definitions
2048 instead of fixing them up at runtime. This keeps the error code
2049 structures constant.
2052 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
2054 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
2057 *) Fixes for newer kerberos headers. NB: the casts are needed because
2058 the 'length' field is signed on one version and unsigned on another
2059 with no (?) obvious way to tell the difference, without these VC++
2060 complains. Also the "definition" of FAR (blank) is no longer included
2061 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
2062 some needed definitions.
2065 *) Undo Cygwin change.
2068 *) Added support for proxy certificates according to RFC 3820.
2069 Because they may be a security thread to unaware applications,
2070 they must be explicitely allowed in run-time. See
2071 docs/HOWTO/proxy_certificates.txt for further information.
2074 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
2076 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
2077 server and client random values. Previously
2078 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
2079 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
2081 This change has negligible security impact because:
2083 1. Server and client random values still have 24 bytes of pseudo random
2086 2. Server and client random values are sent in the clear in the initial
2089 3. The master secret is derived using the premaster secret (48 bytes in
2090 size for static RSA ciphersuites) as well as client server and random
2093 The OpenSSL team would like to thank the UK NISCC for bringing this issue
2096 [Stephen Henson, reported by UK NISCC]
2098 *) Use Windows randomness collection on Cygwin.
2101 *) Fix hang in EGD/PRNGD query when communication socket is closed
2102 prematurely by EGD/PRNGD.
2103 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
2105 *) Prompt for pass phrases when appropriate for PKCS12 input format.
2108 *) Back-port of selected performance improvements from development
2109 branch, as well as improved support for PowerPC platforms.
2112 *) Add lots of checks for memory allocation failure, error codes to indicate
2113 failure and freeing up memory if a failure occurs.
2114 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
2116 *) Add new -passin argument to dgst.
2119 *) Perform some character comparisons of different types in X509_NAME_cmp:
2120 this is needed for some certificates that reencode DNs into UTF8Strings
2121 (in violation of RFC3280) and can't or wont issue name rollover
2125 *) Make an explicit check during certificate validation to see that
2126 the CA setting in each certificate on the chain is correct. As a
2127 side effect always do the following basic checks on extensions,
2128 not just when there's an associated purpose to the check:
2130 - if there is an unhandled critical extension (unless the user
2131 has chosen to ignore this fault)
2132 - if the path length has been exceeded (if one is set at all)
2133 - that certain extensions fit the associated purpose (if one has
2137 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
2139 *) Avoid a race condition when CRLs are checked in a multi threaded
2140 environment. This would happen due to the reordering of the revoked
2141 entries during signature checking and serial number lookup. Now the
2142 encoding is cached and the serial number sort performed under a lock.
2143 Add new STACK function sk_is_sorted().
2146 *) Add Delta CRL to the extension code.
2149 *) Various fixes to s3_pkt.c so alerts are sent properly.
2150 [David Holmes <d.holmes@f5.com>]
2152 *) Reduce the chances of duplicate issuer name and serial numbers (in
2153 violation of RFC3280) using the OpenSSL certificate creation utilities.
2154 This is done by creating a random 64 bit value for the initial serial
2155 number when a serial number file is created or when a self signed
2156 certificate is created using 'openssl req -x509'. The initial serial
2157 number file is created using 'openssl x509 -next_serial' in CA.pl
2158 rather than being initialized to 1.
2161 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
2163 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
2164 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
2165 [Joe Orton, Steve Henson]
2167 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
2169 [Joe Orton, Steve Henson]
2171 *) Make it possible to have multiple active certificates with the same
2172 subject in the CA index file. This is done only if the keyword
2173 'unique_subject' is set to 'no' in the main CA section (default
2174 if 'CA_default') of the configuration file. The value is saved
2175 with the database itself in a separate index attribute file,
2176 named like the index file with '.attr' appended to the name.
2179 *) X509 verify fixes. Disable broken certificate workarounds when
2180 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
2181 keyUsage extension present. Don't accept CRLs with unhandled critical
2182 extensions: since verify currently doesn't process CRL extensions this
2183 rejects a CRL with *any* critical extensions. Add new verify error codes
2187 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
2188 A clarification of RFC2560 will require the use of OCTET STRINGs and
2189 some implementations cannot handle the current raw format. Since OpenSSL
2190 copies and compares OCSP nonces as opaque blobs without any attempt at
2191 parsing them this should not create any compatibility issues.
2194 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
2195 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
2196 this HMAC (and other) operations are several times slower than OpenSSL
2200 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
2201 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2203 *) Use the correct content when signing type "other".
2206 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
2208 *) Fix various bugs revealed by running the NISCC test suite:
2210 Stop out of bounds reads in the ASN1 code when presented with
2211 invalid tags (CVE-2003-0543 and CVE-2003-0544).
2213 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
2215 If verify callback ignores invalid public key errors don't try to check
2216 certificate signature with the NULL public key.
2220 *) New -ignore_err option in ocsp application to stop the server
2221 exiting on the first error in a request.
2224 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
2225 if the server requested one: as stated in TLS 1.0 and SSL 3.0
2229 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
2230 extra data after the compression methods not only for TLS 1.0
2231 but also for SSL 3.0 (as required by the specification).
2232 [Bodo Moeller; problem pointed out by Matthias Loepfe]
2234 *) Change X509_certificate_type() to mark the key as exported/exportable
2235 when it's 512 *bits* long, not 512 bytes.
2238 *) Change AES_cbc_encrypt() so it outputs exact multiple of
2239 blocks during encryption.
2242 *) Various fixes to base64 BIO and non blocking I/O. On write
2243 flushes were not handled properly if the BIO retried. On read
2244 data was not being buffered properly and had various logic bugs.
2245 This also affects blocking I/O when the data being decoded is a
2249 *) Various S/MIME bugfixes and compatibility changes:
2250 output correct application/pkcs7 MIME type if
2251 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
2252 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
2253 of files as .eml work). Correctly handle very long lines in MIME
2257 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
2259 *) Countermeasure against the Klima-Pokorny-Rosa extension of
2260 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
2261 a protocol version number mismatch like a decryption error
2262 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
2265 *) Turn on RSA blinding by default in the default implementation
2266 to avoid a timing attack. Applications that don't want it can call
2267 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
2268 They would be ill-advised to do so in most cases.
2269 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
2271 *) Change RSA blinding code so that it works when the PRNG is not
2272 seeded (in this case, the secret RSA exponent is abused as
2273 an unpredictable seed -- if it is not unpredictable, there
2274 is no point in blinding anyway). Make RSA blinding thread-safe
2275 by remembering the creator's thread ID in rsa->blinding and
2276 having all other threads use local one-time blinding factors
2277 (this requires more computation than sharing rsa->blinding, but
2278 avoids excessive locking; and if an RSA object is not shared
2279 between threads, blinding will still be very fast).
2282 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
2283 ENGINE as defaults for all supported algorithms irrespective of
2284 the 'flags' parameter. 'flags' is now honoured, so applications
2285 should make sure they are passing it correctly.
2288 *) Target "mingw" now allows native Windows code to be generated in
2289 the Cygwin environment as well as with the MinGW compiler.
2292 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
2294 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
2295 via timing by performing a MAC computation even if incorrrect
2296 block cipher padding has been found. This is a countermeasure
2297 against active attacks where the attacker has to distinguish
2298 between bad padding and a MAC verification error. (CVE-2003-0078)
2300 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
2301 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
2302 Martin Vuagnoux (EPFL, Ilion)]
2304 *) Make the no-err option work as intended. The intention with no-err
2305 is not to have the whole error stack handling routines removed from
2306 libcrypto, it's only intended to remove all the function name and
2307 reason texts, thereby removing some of the footprint that may not
2308 be interesting if those errors aren't displayed anyway.
2310 NOTE: it's still possible for any application or module to have it's
2311 own set of error texts inserted. The routines are there, just not
2312 used by default when no-err is given.
2315 *) Add support for FreeBSD on IA64.
2316 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
2318 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
2319 Kerberos function mit_des_cbc_cksum(). Before this change,
2320 the value returned by DES_cbc_cksum() was like the one from
2321 mit_des_cbc_cksum(), except the bytes were swapped.
2322 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
2324 *) Allow an application to disable the automatic SSL chain building.
2325 Before this a rather primitive chain build was always performed in
2326 ssl3_output_cert_chain(): an application had no way to send the
2327 correct chain if the automatic operation produced an incorrect result.
2329 Now the chain builder is disabled if either:
2331 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
2333 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
2335 The reasoning behind this is that an application would not want the
2336 auto chain building to take place if extra chain certificates are
2337 present and it might also want a means of sending no additional
2338 certificates (for example the chain has two certificates and the
2342 *) Add the possibility to build without the ENGINE framework.
2343 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2345 *) Under Win32 gmtime() can return NULL: check return value in
2346 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
2349 *) DSA routines: under certain error conditions uninitialized BN objects
2350 could be freed. Solution: make sure initialization is performed early
2351 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
2352 Nils Larsch <nla@trustcenter.de> via PR#459)
2355 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
2356 checked on reconnect on the client side, therefore session resumption
2357 could still fail with a "ssl session id is different" error. This
2358 behaviour is masked when SSL_OP_ALL is used due to
2359 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
2360 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2361 followup to PR #377.
2364 *) IA-32 assembler support enhancements: unified ELF targets, support
2365 for SCO/Caldera platforms, fix for Cygwin shared build.
2368 *) Add support for FreeBSD on sparc64. As a consequence, support for
2369 FreeBSD on non-x86 processors is separate from x86 processors on
2370 the config script, much like the NetBSD support.
2371 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
2373 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
2375 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
2378 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
2379 code (06) was taken as the first octet of the session ID and the last
2380 octet was ignored consequently. As a result SSLv2 client side session
2381 caching could not have worked due to the session ID mismatch between
2383 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2387 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
2388 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
2392 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
2393 seems that in spite of existing for more than a year, many application
2394 author have done nothing to provide the necessary callbacks, which
2395 means that this particular engine will not work properly anywhere.
2396 This is a very unfortunate situation which forces us, in the name
2397 of usability, to give the hw_ncipher.c a static lock, which is part
2399 NOTE: This is for the 0.9.7 series ONLY. This hack will never
2400 appear in 0.9.8 or later. We EXPECT application authors to have
2401 dealt properly with this when 0.9.8 is released (unless we actually
2402 make such changes in the libcrypto locking code that changes will
2403 have to be made anyway).
2406 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
2407 octets have been read, EOF or an error occurs. Without this change
2408 some truncated ASN1 structures will not produce an error.
2411 *) Disable Heimdal support, since it hasn't been fully implemented.
2412 Still give the possibility to force the use of Heimdal, but with
2413 warnings and a request that patches get sent to openssl-dev.
2416 *) Add the VC-CE target, introduce the WINCE sysname, and add
2417 INSTALL.WCE and appropriate conditionals to make it build.
2418 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2420 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
2421 cygssl-x.y.z.dll, where x, y and z are the major, minor and
2422 edit numbers of the version.
2423 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2425 *) Introduce safe string copy and catenation functions
2426 (BUF_strlcpy() and BUF_strlcat()).
2427 [Ben Laurie (CHATS) and Richard Levitte]
2429 *) Avoid using fixed-size buffers for one-line DNs.
2430 [Ben Laurie (CHATS)]
2432 *) Add BUF_MEM_grow_clean() to avoid information leakage when
2433 resizing buffers containing secrets, and use where appropriate.
2434 [Ben Laurie (CHATS)]
2436 *) Avoid using fixed size buffers for configuration file location.
2437 [Ben Laurie (CHATS)]
2439 *) Avoid filename truncation for various CA files.
2440 [Ben Laurie (CHATS)]
2442 *) Use sizeof in preference to magic numbers.
2443 [Ben Laurie (CHATS)]
2445 *) Avoid filename truncation in cert requests.
2446 [Ben Laurie (CHATS)]
2448 *) Add assertions to check for (supposedly impossible) buffer
2450 [Ben Laurie (CHATS)]
2452 *) Don't cache truncated DNS entries in the local cache (this could
2453 potentially lead to a spoofing attack).
2454 [Ben Laurie (CHATS)]
2456 *) Fix various buffers to be large enough for hex/decimal
2457 representations in a platform independent manner.
2458 [Ben Laurie (CHATS)]
2460 *) Add CRYPTO_realloc_clean() to avoid information leakage when
2461 resizing buffers containing secrets, and use where appropriate.
2462 [Ben Laurie (CHATS)]
2464 *) Add BIO_indent() to avoid much slightly worrying code to do
2466 [Ben Laurie (CHATS)]
2468 *) Convert sprintf()/BIO_puts() to BIO_printf().
2469 [Ben Laurie (CHATS)]
2471 *) buffer_gets() could terminate with the buffer only half
2473 [Ben Laurie (CHATS)]
2475 *) Add assertions to prevent user-supplied crypto functions from
2476 overflowing internal buffers by having large block sizes, etc.
2477 [Ben Laurie (CHATS)]
2479 *) New OPENSSL_assert() macro (similar to assert(), but enabled
2481 [Ben Laurie (CHATS)]
2483 *) Eliminate unused copy of key in RC4.
2484 [Ben Laurie (CHATS)]
2486 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
2487 [Ben Laurie (CHATS)]
2489 *) Fix off-by-one error in EGD path.
2490 [Ben Laurie (CHATS)]
2492 *) If RANDFILE path is too long, ignore instead of truncating.
2493 [Ben Laurie (CHATS)]
2495 *) Eliminate unused and incorrectly sized X.509 structure
2497 [Ben Laurie (CHATS)]
2499 *) Eliminate unused and dangerous function knumber().
2500 [Ben Laurie (CHATS)]
2502 *) Eliminate unused and dangerous structure, KSSL_ERR.
2503 [Ben Laurie (CHATS)]
2505 *) Protect against overlong session ID context length in an encoded
2506 session object. Since these are local, this does not appear to be
2508 [Ben Laurie (CHATS)]
2510 *) Change from security patch (see 0.9.6e below) that did not affect
2511 the 0.9.6 release series:
2513 Remote buffer overflow in SSL3 protocol - an attacker could
2514 supply an oversized master key in Kerberos-enabled versions.
2516 [Ben Laurie (CHATS)]
2518 *) Change the SSL kerb5 codes to match RFC 2712.
2521 *) Make -nameopt work fully for req and add -reqopt switch.
2522 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
2524 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
2525 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
2527 *) Make sure tests can be performed even if the corresponding algorithms
2528 have been removed entirely. This was also the last step to make
2529 OpenSSL compilable with DJGPP under all reasonable conditions.
2530 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
2532 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
2533 to allow version independent disabling of normally unselected ciphers,
2534 which may be activated as a side-effect of selecting a single cipher.
2536 (E.g., cipher list string "RSA" enables ciphersuites that are left
2537 out of "ALL" because they do not provide symmetric encryption.
2538 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
2539 [Lutz Jaenicke, Bodo Moeller]
2541 *) Add appropriate support for separate platform-dependent build
2542 directories. The recommended way to make a platform-dependent
2543 build directory is the following (tested on Linux), maybe with
2546 # Place yourself outside of the OpenSSL source tree. In
2547 # this example, the environment variable OPENSSL_SOURCE
2548 # is assumed to contain the absolute OpenSSL source directory.
2549 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
2550 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
2551 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
2552 mkdir -p `dirname $F`
2553 ln -s $OPENSSL_SOURCE/$F $F
2556 To be absolutely sure not to disturb the source tree, a "make clean"
2557 is a good thing. If it isn't successfull, don't worry about it,
2558 it probably means the source directory is very clean.
2561 *) Make sure any ENGINE control commands make local copies of string
2562 pointers passed to them whenever necessary. Otherwise it is possible
2563 the caller may have overwritten (or deallocated) the original string
2564 data when a later ENGINE operation tries to use the stored values.
2565 [Götz Babin-Ebell <babinebell@trustcenter.de>]
2567 *) Improve diagnostics in file reading and command-line digests.
2568 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
2570 *) Add AES modes CFB and OFB to the object database. Correct an
2571 error in AES-CFB decryption.
2574 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
2575 allows existing EVP_CIPHER_CTX structures to be reused after
2576 calling EVP_*Final(). This behaviour is used by encryption
2577 BIOs and some applications. This has the side effect that
2578 applications must explicitly clean up cipher contexts with
2579 EVP_CIPHER_CTX_cleanup() or they will leak memory.
2582 *) Check the values of dna and dnb in bn_mul_recursive before calling
2583 bn_mul_comba (a non zero value means the a or b arrays do not contain
2584 n2 elements) and fallback to bn_mul_normal if either is not zero.
2587 *) Fix escaping of non-ASCII characters when using the -subj option
2588 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
2591 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
2592 form for "surname", serialNumber has no short form.
2593 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
2594 therefore remove "mail" short name for "internet 7".
2595 The OID for unique identifiers in X509 certificates is
2596 x500UniqueIdentifier, not uniqueIdentifier.
2597 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
2600 *) Add an "init" command to the ENGINE config module and auto initialize
2601 ENGINEs. Without any "init" command the ENGINE will be initialized
2602 after all ctrl commands have been executed on it. If init=1 the
2603 ENGINE is initailized at that point (ctrls before that point are run
2604 on the uninitialized ENGINE and after on the initialized one). If
2605 init=0 then the ENGINE will not be iniatialized at all.
2608 *) Fix the 'app_verify_callback' interface so that the user-defined
2609 argument is actually passed to the callback: In the
2610 SSL_CTX_set_cert_verify_callback() prototype, the callback
2611 declaration has been changed from
2614 int (*cb)(X509_STORE_CTX *,void *);
2615 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
2616 i=s->ctx->app_verify_callback(&ctx)
2617 has been changed into
2618 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
2620 To update applications using SSL_CTX_set_cert_verify_callback(),
2621 a dummy argument can be added to their callback functions.
2622 [D. K. Smetters <smetters@parc.xerox.com>]
2624 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
2625 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
2627 *) Add and OPENSSL_LOAD_CONF define which will cause
2628 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
2629 This allows older applications to transparently support certain
2630 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
2631 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
2632 load the config file and OPENSSL_add_all_algorithms_conf() which will
2633 always load it have also been added.
2636 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
2637 Adjust NIDs and EVP layer.
2638 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2640 *) Config modules support in openssl utility.
2642 Most commands now load modules from the config file,
2643 though in a few (such as version) this isn't done
2644 because it couldn't be used for anything.
2646 In the case of ca and req the config file used is
2647 the same as the utility itself: that is the -config
2648 command line option can be used to specify an
2652 *) Move default behaviour from OPENSSL_config(). If appname is NULL
2653 use "openssl_conf" if filename is NULL use default openssl config file.
2656 *) Add an argument to OPENSSL_config() to allow the use of an alternative
2657 config section name. Add a new flag to tolerate a missing config file
2658 and move code to CONF_modules_load_file().
2661 *) Support for crypto accelerator cards from Accelerated Encryption
2662 Processing, www.aep.ie. (Use engine 'aep')
2663 The support was copied from 0.9.6c [engine] and adapted/corrected
2664 to work with the new engine framework.
2665 [AEP Inc. and Richard Levitte]
2667 *) Support for SureWare crypto accelerator cards from Baltimore
2668 Technologies. (Use engine 'sureware')
2669 The support was copied from 0.9.6c [engine] and adapted
2670 to work with the new engine framework.
2673 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
2674 make the newer ENGINE framework commands for the CHIL engine work.
2675 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
2677 *) Make it possible to produce shared libraries on ReliantUNIX.
2678 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
2680 *) Add the configuration target debug-linux-ppro.
2681 Make 'openssl rsa' use the general key loading routines
2682 implemented in apps.c, and make those routines able to
2683 handle the key format FORMAT_NETSCAPE and the variant
2685 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2687 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2688 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2690 *) Add -keyform to rsautl, and document -engine.
2691 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
2693 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
2694 BIO_R_NO_SUCH_FILE error code rather than the generic
2695 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
2698 *) Add new functions
2700 ERR_peek_last_error_line
2701 ERR_peek_last_error_line_data.
2702 These are similar to
2705 ERR_peek_error_line_data,
2706 but report on the latest error recorded rather than the first one
2707 still in the error queue.
2708 [Ben Laurie, Bodo Moeller]
2710 *) default_algorithms option in ENGINE config module. This allows things
2712 default_algorithms = ALL
2713 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
2716 *) Prelminary ENGINE config module.
2719 *) New experimental application configuration code.
2722 *) Change the AES code to follow the same name structure as all other
2723 symmetric ciphers, and behave the same way. Move everything to
2724 the directory crypto/aes, thereby obsoleting crypto/rijndael.
2725 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2727 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
2728 [Ben Laurie and Theo de Raadt]
2730 *) Add option to output public keys in req command.
2731 [Massimiliano Pala madwolf@openca.org]
2733 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
2734 (up to about 10% better than before for P-192 and P-224).
2737 *) New functions/macros
2739 SSL_CTX_set_msg_callback(ctx, cb)
2740 SSL_CTX_set_msg_callback_arg(ctx, arg)
2741 SSL_set_msg_callback(ssl, cb)
2742 SSL_set_msg_callback_arg(ssl, arg)
2744 to request calling a callback function
2746 void cb(int write_p, int version, int content_type,
2747 const void *buf, size_t len, SSL *ssl, void *arg)
2749 whenever a protocol message has been completely received
2750 (write_p == 0) or sent (write_p == 1). Here 'version' is the
2751 protocol version according to which the SSL library interprets
2752 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
2753 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
2754 the content type as defined in the SSL 3.0/TLS 1.0 protocol
2755 specification (change_cipher_spec(20), alert(21), handshake(22)).
2756 'buf' and 'len' point to the actual message, 'ssl' to the
2757 SSL object, and 'arg' is the application-defined value set by
2758 SSL[_CTX]_set_msg_callback_arg().
2760 'openssl s_client' and 'openssl s_server' have new '-msg' options
2761 to enable a callback that displays all protocol messages.
2764 *) Change the shared library support so shared libraries are built as
2765 soon as the corresponding static library is finished, and thereby get
2766 openssl and the test programs linked against the shared library.
2767 This still only happens when the keyword "shard" has been given to
2768 the configuration scripts.
2770 NOTE: shared library support is still an experimental thing, and
2771 backward binary compatibility is still not guaranteed.
2772 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
2774 *) Add support for Subject Information Access extension.
2775 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2777 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
2778 additional bytes when new memory had to be allocated, not just
2779 when reusing an existing buffer.
2782 *) New command line and configuration option 'utf8' for the req command.
2783 This allows field values to be specified as UTF8 strings.
2786 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
2787 runs for the former and machine-readable output for the latter.
2790 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
2791 of the e-mail address in the DN (i.e., it will go into a certificate
2792 extension only). The new configuration file option 'email_in_dn = no'
2793 has the same effect.
2794 [Massimiliano Pala madwolf@openca.org]
2796 *) Change all functions with names starting with des_ to be starting
2797 with DES_ instead. Add wrappers that are compatible with libdes,
2798 but are named _ossl_old_des_*. Finally, add macros that map the
2799 des_* symbols to the corresponding _ossl_old_des_* if libdes
2800 compatibility is desired. If OpenSSL 0.9.6c compatibility is
2801 desired, the des_* symbols will be mapped to DES_*, with one
2804 Since we provide two compatibility mappings, the user needs to
2805 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
2806 compatibility is desired. The default (i.e., when that macro
2807 isn't defined) is OpenSSL 0.9.6c compatibility.
2809 There are also macros that enable and disable the support of old
2810 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
2811 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
2812 are defined, the default will apply: to support the old des routines.
2814 In either case, one must include openssl/des.h to get the correct
2815 definitions. Do not try to just include openssl/des_old.h, that
2818 NOTE: This is a major break of an old API into a new one. Software
2819 authors are encouraged to switch to the DES_ style functions. Some
2820 time in the future, des_old.h and the libdes compatibility functions
2821 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
2822 default), and then completely removed.
2825 *) Test for certificates which contain unsupported critical extensions.
2826 If such a certificate is found during a verify operation it is
2827 rejected by default: this behaviour can be overridden by either
2828 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
2829 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
2830 X509_supported_extension() has also been added which returns 1 if a
2831 particular extension is supported.
2834 *) Modify the behaviour of EVP cipher functions in similar way to digests
2835 to retain compatibility with existing code.
2838 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
2839 compatibility with existing code. In particular the 'ctx' parameter does
2840 not have to be to be initialized before the call to EVP_DigestInit() and
2841 it is tidied up after a call to EVP_DigestFinal(). New function
2842 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
2843 EVP_MD_CTX_copy() changed to not require the destination to be
2844 initialized valid and new function EVP_MD_CTX_copy_ex() added which
2845 requires the destination to be valid.
2847 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
2848 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
2851 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
2852 so that complete 'Handshake' protocol structures are kept in memory
2853 instead of overwriting 'msg_type' and 'length' with 'body' data.
2856 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
2857 [Massimo Santin via Richard Levitte]
2859 *) Major restructuring to the underlying ENGINE code. This includes
2860 reduction of linker bloat, separation of pure "ENGINE" manipulation
2861 (initialisation, etc) from functionality dealing with implementations
2862 of specific crypto iterfaces. This change also introduces integrated
2863 support for symmetric ciphers and digest implementations - so ENGINEs
2864 can now accelerate these by providing EVP_CIPHER and EVP_MD
2865 implementations of their own. This is detailed in crypto/engine/README
2866 as it couldn't be adequately described here. However, there are a few
2867 API changes worth noting - some RSA, DSA, DH, and RAND functions that
2868 were changed in the original introduction of ENGINE code have now
2869 reverted back - the hooking from this code to ENGINE is now a good
2870 deal more passive and at run-time, operations deal directly with
2871 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
2872 dereferencing through an ENGINE pointer any more. Also, the ENGINE
2873 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
2874 they were not being used by the framework as there is no concept of a
2875 BIGNUM_METHOD and they could not be generalised to the new
2876 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
2877 ENGINE_cpy() has been removed as it cannot be consistently defined in
2881 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
2884 *) Change mkdef.pl to sort symbols that get the same entry number,
2885 and make sure the automatically generated functions ERR_load_*
2886 become part of libeay.num as well.
2889 *) New function SSL_renegotiate_pending(). This returns true once
2890 renegotiation has been requested (either SSL_renegotiate() call
2891 or HelloRequest/ClientHello receveived from the peer) and becomes
2892 false once a handshake has been completed.
2893 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
2894 sends a HelloRequest, but does not ensure that a handshake takes
2895 place. SSL_renegotiate_pending() is useful for checking if the
2896 client has followed the request.)
2899 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
2900 By default, clients may request session resumption even during
2901 renegotiation (if session ID contexts permit); with this option,
2902 session resumption is possible only in the first handshake.
2904 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
2905 more bits available for options that should not be part of
2906 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
2909 *) Add some demos for certificate and certificate request creation.
2912 *) Make maximum certificate chain size accepted from the peer application
2913 settable (SSL*_get/set_max_cert_list()), as proposed by
2914 "Douglas E. Engert" <deengert@anl.gov>.
2917 *) Add support for shared libraries for Unixware-7
2918 (Boyd Lynn Gerber <gerberb@zenez.com>).
2921 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
2922 be done prior to destruction. Use this to unload error strings from
2923 ENGINEs that load their own error strings. NB: This adds two new API
2924 functions to "get" and "set" this destroy handler in an ENGINE.
2927 *) Alter all existing ENGINE implementations (except "openssl" and
2928 "openbsd") to dynamically instantiate their own error strings. This
2929 makes them more flexible to be built both as statically-linked ENGINEs
2930 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
2931 Also, add stub code to each that makes building them as self-contained
2932 shared-libraries easier (see README.ENGINE).
2935 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
2936 implementations into applications that are completely implemented in
2937 self-contained shared-libraries. The "dynamic" ENGINE exposes control
2938 commands that can be used to configure what shared-library to load and
2939 to control aspects of the way it is handled. Also, made an update to
2940 the README.ENGINE file that brings its information up-to-date and
2941 provides some information and instructions on the "dynamic" ENGINE
2942 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
2945 *) Make it possible to unload ranges of ERR strings with a new
2946 "ERR_unload_strings" function.
2949 *) Add a copy() function to EVP_MD.
2952 *) Make EVP_MD routines take a context pointer instead of just the
2953 md_data void pointer.
2956 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
2957 that the digest can only process a single chunk of data
2958 (typically because it is provided by a piece of
2959 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
2960 is only going to provide a single chunk of data, and hence the
2961 framework needn't accumulate the data for oneshot drivers.
2964 *) As with "ERR", make it possible to replace the underlying "ex_data"
2965 functions. This change also alters the storage and management of global
2966 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
2967 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
2968 index counters. The API functions that use this state have been changed
2969 to take a "class_index" rather than pointers to the class's local STACK
2970 and counter, and there is now an API function to dynamically create new
2971 classes. This centralisation allows us to (a) plug a lot of the
2972 thread-safety problems that existed, and (b) makes it possible to clean
2973 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
2974 such data would previously have always leaked in application code and
2975 workarounds were in place to make the memory debugging turn a blind eye
2976 to it. Application code that doesn't use this new function will still
2977 leak as before, but their memory debugging output will announce it now
2978 rather than letting it slide.
2980 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
2981 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
2982 has a return value to indicate success or failure.
2985 *) Make it possible to replace the underlying "ERR" functions such that the
2986 global state (2 LHASH tables and 2 locks) is only used by the "default"
2987 implementation. This change also adds two functions to "get" and "set"
2988 the implementation prior to it being automatically set the first time
2989 any other ERR function takes place. Ie. an application can call "get",
2990 pass the return value to a module it has just loaded, and that module
2991 can call its own "set" function using that value. This means the
2992 module's "ERR" operations will use (and modify) the error state in the
2993 application and not in its own statically linked copy of OpenSSL code.
2996 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
2997 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
2998 the operation, and provides a more encapsulated way for external code
2999 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
3000 to use these functions rather than manually incrementing the counts.
3002 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
3005 *) Add EVP test program.
3008 *) Add symmetric cipher support to ENGINE. Expect the API to change!
3011 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
3012 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
3013 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
3014 These allow a CRL to be built without having to access X509_CRL fields
3015 directly. Modify 'ca' application to use new functions.
3018 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
3019 bug workarounds. Rollback attack detection is a security feature.
3020 The problem will only arise on OpenSSL servers when TLSv1 is not
3021 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
3022 Software authors not wanting to support TLSv1 will have special reasons
3023 for their choice and can explicitly enable this option.
3024 [Bodo Moeller, Lutz Jaenicke]
3026 *) Rationalise EVP so it can be extended: don't include a union of
3027 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
3028 (similar to those existing for EVP_CIPHER_CTX).
3033 EVP_MD_CTX_init(&md); /* new function call */
3034 EVP_DigestInit(&md, EVP_sha1());
3035 EVP_DigestUpdate(&md, in, len);
3036 EVP_DigestFinal(&md, out, NULL);
3037 EVP_MD_CTX_cleanup(&md); /* new function call */
3041 *) Make DES key schedule conform to the usual scheme, as well as
3042 correcting its structure. This means that calls to DES functions
3043 now have to pass a pointer to a des_key_schedule instead of a
3044 plain des_key_schedule (which was actually always a pointer
3047 des_key_schedule ks;
3049 des_set_key_checked(..., &ks);
3050 des_ncbc_encrypt(..., &ks, ...);
3052 (Note that a later change renames 'des_...' into 'DES_...'.)
3055 *) Initial reduction of linker bloat: the use of some functions, such as
3056 PEM causes large amounts of unused functions to be linked in due to
3057 poor organisation. For example pem_all.c contains every PEM function
3058 which has a knock on effect of linking in large amounts of (unused)
3059 ASN1 code. Grouping together similar functions and splitting unrelated
3060 functions prevents this.
3063 *) Cleanup of EVP macros.
3066 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
3067 correct _ecb suffix.
3070 *) Add initial OCSP responder support to ocsp application. The
3071 revocation information is handled using the text based index
3072 use by the ca application. The responder can either handle
3073 requests generated internally, supplied in files (for example
3074 via a CGI script) or using an internal minimal server.
3077 *) Add configuration choices to get zlib compression for TLS.
3080 *) Changes to Kerberos SSL for RFC 2712 compliance:
3081 1. Implemented real KerberosWrapper, instead of just using
3082 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
3083 2. Implemented optional authenticator field of KerberosWrapper.
3085 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
3086 and authenticator structs; see crypto/krb5/.
3088 Generalized Kerberos calls to support multiple Kerberos libraries.
3089 [Vern Staats <staatsvr@asc.hpc.mil>,
3090 Jeffrey Altman <jaltman@columbia.edu>
3091 via Richard Levitte]
3093 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
3094 already does with RSA. testdsa.h now has 'priv_key/pub_key'
3095 values for each of the key sizes rather than having just
3096 parameters (and 'speed' generating keys each time).
3099 *) Speed up EVP routines.
3102 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
3103 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
3104 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
3105 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
3107 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
3108 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
3109 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
3112 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
3114 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
3117 *) Added the OS2-EMX target.
3118 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
3120 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
3121 to support NCONF routines in extension code. New function CONF_set_nconf()
3122 to allow functions which take an NCONF to also handle the old LHASH
3123 structure: this means that the old CONF compatible routines can be
3124 retained (in particular wrt extensions) without having to duplicate the
3125 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
3128 *) Enhance the general user interface with mechanisms for inner control
3129 and with possibilities to have yes/no kind of prompts.
3132 *) Change all calls to low level digest routines in the library and
3133 applications to use EVP. Add missing calls to HMAC_cleanup() and
3134 don't assume HMAC_CTX can be copied using memcpy().
3135 [Verdon Walker <VWalker@novell.com>, Steve Henson]
3137 *) Add the possibility to control engines through control names but with
3138 arbitrary arguments instead of just a string.
3139 Change the key loaders to take a UI_METHOD instead of a callback
3140 function pointer. NOTE: this breaks binary compatibility with earlier
3141 versions of OpenSSL [engine].
3142 Adapt the nCipher code for these new conditions and add a card insertion
3146 *) Enhance the general user interface with mechanisms to better support
3147 dialog box interfaces, application-defined prompts, the possibility
3148 to use defaults (for example default passwords from somewhere else)
3149 and interrupts/cancellations.
3152 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
3153 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
3156 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
3157 tidy up some unnecessarily weird code in 'sk_new()').
3158 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
3160 *) Change the key loading routines for ENGINEs to use the same kind
3161 callback (pem_password_cb) as all other routines that need this
3165 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
3166 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
3167 than this minimum value is recommended.
3170 *) New random seeder for OpenVMS, using the system process statistics
3171 that are easily reachable.
3174 *) Windows apparently can't transparently handle global
3175 variables defined in DLLs. Initialisations such as:
3177 const ASN1_ITEM *it = &ASN1_INTEGER_it;
3179 wont compile. This is used by the any applications that need to
3180 declare their own ASN1 modules. This was fixed by adding the option
3181 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
3182 needed for static libraries under Win32.
3185 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
3186 setting of purpose and trust fields. New X509_STORE trust and
3187 purpose functions and tidy up setting in other SSL functions.
3190 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
3191 structure. These are inherited by X509_STORE_CTX when it is
3192 initialised. This allows various defaults to be set in the
3193 X509_STORE structure (such as flags for CRL checking and custom
3194 purpose or trust settings) for functions which only use X509_STORE_CTX
3195 internally such as S/MIME.
3197 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
3198 trust settings if they are not set in X509_STORE. This allows X509_STORE
3199 purposes and trust (in S/MIME for example) to override any set by default.
3201 Add command line options for CRL checking to smime, s_client and s_server
3205 *) Initial CRL based revocation checking. If the CRL checking flag(s)
3206 are set then the CRL is looked up in the X509_STORE structure and
3207 its validity and signature checked, then if the certificate is found
3208 in the CRL the verify fails with a revoked error.
3210 Various new CRL related callbacks added to X509_STORE_CTX structure.
3212 Command line options added to 'verify' application to support this.
3214 This needs some additional work, such as being able to handle multiple
3215 CRLs with different times, extension based lookup (rather than just
3216 by subject name) and ultimately more complete V2 CRL extension
3220 *) Add a general user interface API (crypto/ui/). This is designed
3221 to replace things like des_read_password and friends (backward
3222 compatibility functions using this new API are provided).
3223 The purpose is to remove prompting functions from the DES code
3224 section as well as provide for prompting through dialog boxes in
3225 a window system and the like.
3228 *) Add "ex_data" support to ENGINE so implementations can add state at a
3229 per-structure level rather than having to store it globally.
3232 *) Make it possible for ENGINE structures to be copied when retrieved by
3233 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
3234 This causes the "original" ENGINE structure to act like a template,
3235 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
3236 operational state can be localised to each ENGINE structure, despite the
3237 fact they all share the same "methods". New ENGINE structures returned in
3238 this case have no functional references and the return value is the single
3239 structural reference. This matches the single structural reference returned
3240 by ENGINE_by_id() normally, when it is incremented on the pre-existing
3244 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
3245 needs to match any other type at all we need to manually clear the
3249 *) Changes to the "openssl engine" utility to include;
3250 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
3251 about an ENGINE's available control commands.
3252 - executing control commands from command line arguments using the
3253 '-pre' and '-post' switches. '-post' is only used if '-t' is
3254 specified and the ENGINE is successfully initialised. The syntax for
3255 the individual commands are colon-separated, for example;
3256 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
3259 *) New dynamic control command support for ENGINEs. ENGINEs can now
3260 declare their own commands (numbers), names (strings), descriptions,
3261 and input types for run-time discovery by calling applications. A
3262 subset of these commands are implicitly classed as "executable"
3263 depending on their input type, and only these can be invoked through
3264 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
3265 can be based on user input, config files, etc). The distinction is
3266 that "executable" commands cannot return anything other than a boolean
3267 result and can only support numeric or string input, whereas some
3268 discoverable commands may only be for direct use through
3269 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
3270 pointers, or other custom uses. The "executable" commands are to
3271 support parameterisations of ENGINE behaviour that can be
3272 unambiguously defined by ENGINEs and used consistently across any
3273 OpenSSL-based application. Commands have been added to all the
3274 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
3275 control over shared-library paths without source code alterations.
3278 *) Changed all ENGINE implementations to dynamically allocate their
3279 ENGINEs rather than declaring them statically. Apart from this being
3280 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
3281 this also allows the implementations to compile without using the
3282 internal engine_int.h header.
3285 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
3286 'const' value. Any code that should be able to modify a RAND_METHOD
3287 should already have non-const pointers to it (ie. they should only
3288 modify their own ones).
3291 *) Made a variety of little tweaks to the ENGINE code.
3292 - "atalla" and "ubsec" string definitions were moved from header files
3293 to C code. "nuron" string definitions were placed in variables
3294 rather than hard-coded - allowing parameterisation of these values
3295 later on via ctrl() commands.
3296 - Removed unused "#if 0"'d code.
3297 - Fixed engine list iteration code so it uses ENGINE_free() to release
3298 structural references.
3299 - Constified the RAND_METHOD element of ENGINE structures.
3300 - Constified various get/set functions as appropriate and added
3301 missing functions (including a catch-all ENGINE_cpy that duplicates
3302 all ENGINE values onto a new ENGINE except reference counts/state).
3303 - Removed NULL parameter checks in get/set functions. Setting a method
3304 or function to NULL is a way of cancelling out a previously set
3305 value. Passing a NULL ENGINE parameter is just plain stupid anyway
3306 and doesn't justify the extra error symbols and code.
3307 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
3308 flags from engine_int.h to engine.h.
3309 - Changed prototypes for ENGINE handler functions (init(), finish(),
3310 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
3313 *) Implement binary inversion algorithm for BN_mod_inverse in addition
3314 to the algorithm using long division. The binary algorithm can be
3315 used only if the modulus is odd. On 32-bit systems, it is faster
3316 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
3317 roughly 5-15% for 256-bit moduli), so we use it only for moduli
3318 up to 450 bits. In 64-bit environments, the binary algorithm
3319 appears to be advantageous for much longer moduli; here we use it
3320 for moduli up to 2048 bits.
3323 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
3324 could not support the combine flag in choice fields.
3327 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
3328 extensions from a certificate request to the certificate.
3331 *) Allow multiple 'certopt' and 'nameopt' options to be separated
3332 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
3333 file: this allows the display of the certificate about to be
3334 signed to be customised, to allow certain fields to be included
3335 or excluded and extension details. The old system didn't display
3336 multicharacter strings properly, omitted fields not in the policy
3337 and couldn't display additional details such as extensions.
3340 *) Function EC_POINTs_mul for multiple scalar multiplication
3341 of an arbitrary number of elliptic curve points
3342 \sum scalars[i]*points[i],
3343 optionally including the generator defined for the EC_GROUP:
3344 scalar*generator + \sum scalars[i]*points[i].
3346 EC_POINT_mul is a simple wrapper function for the typical case
3347 that the point list has just one item (besides the optional
3351 *) First EC_METHODs for curves over GF(p):
3353 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
3354 operations and provides various method functions that can also
3355 operate with faster implementations of modular arithmetic.
3357 EC_GFp_mont_method() reuses most functions that are part of
3358 EC_GFp_simple_method, but uses Montgomery arithmetic.
3360 [Bodo Moeller; point addition and point doubling
3361 implementation directly derived from source code provided by
3362 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
3364 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
3365 crypto/ec/ec_lib.c):
3367 Curves are EC_GROUP objects (with an optional group generator)
3368 based on EC_METHODs that are built into the library.
3370 Points are EC_POINT objects based on EC_GROUP objects.
3372 Most of the framework would be able to handle curves over arbitrary
3373 finite fields, but as there are no obvious types for fields other
3374 than GF(p), some functions are limited to that for now.
3377 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
3378 that the file contains a complete HTTP response.
3381 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
3382 change the def and num file printf format specifier from "%-40sXXX"
3383 to "%-39s XXX". The latter will always guarantee a space after the
3384 field while the former will cause them to run together if the field
3385 is 40 of more characters long.
3388 *) Constify the cipher and digest 'method' functions and structures
3389 and modify related functions to take constant EVP_MD and EVP_CIPHER
3393 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
3394 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
3397 *) Modify EVP_Digest*() routines so they now return values. Although the
3398 internal software routines can never fail additional hardware versions
3402 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
3404 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
3405 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
3408 ERR_R_NESTED_ASN1_ERROR
3410 ERR_R_MISSING_ASN1_EOS
3411 were 4 .. 9, conflicting with
3412 ERR_LIB_RSA (= ERR_R_RSA_LIB)
3414 ERR_LIB_PEM (= ERR_R_PEM_LIB).
3415 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
3417 Add new error code 'ERR_R_INTERNAL_ERROR'.
3420 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
3424 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
3425 sets the subject name for a new request or supersedes the
3426 subject name in a given request. Formats that can be parsed are
3427 'CN=Some Name, OU=myOU, C=IT'
3429 'CN=Some Name/OU=myOU/C=IT'.
3431 Add options '-batch' and '-verbose' to 'openssl req'.
3432 [Massimiliano Pala <madwolf@hackmasters.net>]
3434 *) Introduce the possibility to access global variables through
3435 functions on platform were that's the best way to handle exporting
3436 global variables in shared libraries. To enable this functionality,
3437 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
3438 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
3439 is normally done by Configure or something similar).
3441 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
3442 in the source file (foo.c) like this:
3444 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
3445 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
3447 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
3448 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
3450 OPENSSL_DECLARE_GLOBAL(int,foo);
3451 #define foo OPENSSL_GLOBAL_REF(foo)
3452 OPENSSL_DECLARE_GLOBAL(double,bar);
3453 #define bar OPENSSL_GLOBAL_REF(bar)
3455 The #defines are very important, and therefore so is including the
3456 header file everywhere where the defined globals are used.
3458 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
3459 of ASN.1 items, but that structure is a bit different.
3461 The largest change is in util/mkdef.pl which has been enhanced with
3462 better and easier to understand logic to choose which symbols should
3463 go into the Windows .def files as well as a number of fixes and code
3464 cleanup (among others, algorithm keywords are now sorted
3465 lexicographically to avoid constant rewrites).
3468 *) In BN_div() keep a copy of the sign of 'num' before writing the
3469 result to 'rm' because if rm==num the value will be overwritten
3470 and produce the wrong result if 'num' is negative: this caused
3471 problems with BN_mod() and BN_nnmod().
3474 *) Function OCSP_request_verify(). This checks the signature on an
3475 OCSP request and verifies the signer certificate. The signer
3476 certificate is just checked for a generic purpose and OCSP request
3480 *) Add OCSP_check_validity() function to check the validity of OCSP
3481 responses. OCSP responses are prepared in real time and may only
3482 be a few seconds old. Simply checking that the current time lies
3483 between thisUpdate and nextUpdate max reject otherwise valid responses
3484 caused by either OCSP responder or client clock inaccuracy. Instead
3485 we allow thisUpdate and nextUpdate to fall within a certain period of
3486 the current time. The age of the response can also optionally be
3487 checked. Two new options -validity_period and -status_age added to
3491 *) If signature or public key algorithm is unrecognized print out its
3492 OID rather that just UNKNOWN.
3495 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
3496 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
3497 ID to be generated from the issuer certificate alone which can then be
3498 passed to OCSP_id_issuer_cmp().
3501 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
3502 ASN1 modules to export functions returning ASN1_ITEM pointers
3503 instead of the ASN1_ITEM structures themselves. This adds several
3504 new macros which allow the underlying ASN1 function/structure to
3505 be accessed transparently. As a result code should not use ASN1_ITEM
3506 references directly (such as &X509_it) but instead use the relevant
3507 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
3508 use of the new ASN1 code on platforms where exporting structures
3509 is problematical (for example in shared libraries) but exporting
3510 functions returning pointers to structures is not.
3513 *) Add support for overriding the generation of SSL/TLS session IDs.
3514 These callbacks can be registered either in an SSL_CTX or per SSL.
3515 The purpose of this is to allow applications to control, if they wish,
3516 the arbitrary values chosen for use as session IDs, particularly as it
3517 can be useful for session caching in multiple-server environments. A
3518 command-line switch for testing this (and any client code that wishes
3519 to use such a feature) has been added to "s_server".
3520 [Geoff Thorpe, Lutz Jaenicke]
3522 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
3523 of the form '#if defined(...) || defined(...) || ...' and
3524 '#if !defined(...) && !defined(...) && ...'. This also avoids
3525 the growing number of special cases it was previously handling.
3528 *) Make all configuration macros available for application by making
3529 sure they are available in opensslconf.h, by giving them names starting
3530 with "OPENSSL_" to avoid conflicts with other packages and by making
3531 sure e_os2.h will cover all platform-specific cases together with
3533 Additionally, it is now possible to define configuration/platform-
3534 specific names (called "system identities"). In the C code, these
3535 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
3536 macro with the name beginning with "OPENSSL_SYS_", which is determined
3537 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
3541 *) New option -set_serial to 'req' and 'x509' this allows the serial
3542 number to use to be specified on the command line. Previously self
3543 signed certificates were hard coded with serial number 0 and the
3544 CA options of 'x509' had to use a serial number in a file which was
3548 *) New options to 'ca' utility to support V2 CRL entry extensions.
3549 Currently CRL reason, invalidity date and hold instruction are
3550 supported. Add new CRL extensions to V3 code and some new objects.
3553 *) New function EVP_CIPHER_CTX_set_padding() this is used to
3554 disable standard block padding (aka PKCS#5 padding) in the EVP
3555 API, which was previously mandatory. This means that the data is
3556 not padded in any way and so the total length much be a multiple
3557 of the block size, otherwise an error occurs.
3560 *) Initial (incomplete) OCSP SSL support.
3563 *) New function OCSP_parse_url(). This splits up a URL into its host,
3564 port and path components: primarily to parse OCSP URLs. New -url
3565 option to ocsp utility.
3568 *) New nonce behavior. The return value of OCSP_check_nonce() now
3569 reflects the various checks performed. Applications can decide
3570 whether to tolerate certain situations such as an absent nonce
3571 in a response when one was present in a request: the ocsp application
3572 just prints out a warning. New function OCSP_add1_basic_nonce()
3573 this is to allow responders to include a nonce in a response even if
3574 the request is nonce-less.
3577 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
3578 skipped when using openssl x509 multiple times on a single input file,
3579 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
3582 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
3583 set string type: to handle setting ASN1_TIME structures. Fix ca
3584 utility to correctly initialize revocation date of CRLs.
3587 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
3588 the clients preferred ciphersuites and rather use its own preferences.
3589 Should help to work around M$ SGC (Server Gated Cryptography) bug in
3590 Internet Explorer by ensuring unchanged hash method during stepup.
3591 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
3594 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
3595 to aes and add a new 'exist' option to print out symbols that don't
3599 *) Additional options to ocsp utility to allow flags to be set and
3600 additional certificates supplied.
3603 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
3604 OCSP client a number of certificate to only verify the response
3608 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
3609 handle the new API. Currently only ECB, CBC modes supported. Add new
3612 Add TLS AES ciphersuites as described in RFC3268, "Advanced
3613 Encryption Standard (AES) Ciphersuites for Transport Layer
3614 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
3615 not enabled by default and were not part of the "ALL" ciphersuite
3616 alias because they were not yet official; they could be
3617 explicitly requested by specifying the "AESdraft" ciphersuite
3618 group alias. In the final release of OpenSSL 0.9.7, the group
3619 alias is called "AES" and is part of "ALL".)
3620 [Ben Laurie, Steve Henson, Bodo Moeller]
3622 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
3623 request to response.
3626 *) Functions for OCSP responders. OCSP_request_onereq_count(),
3627 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
3628 extract information from a certificate request. OCSP_response_create()
3629 creates a response and optionally adds a basic response structure.
3630 OCSP_basic_add1_status() adds a complete single response to a basic
3631 response and returns the OCSP_SINGLERESP structure just added (to allow
3632 extensions to be included for example). OCSP_basic_add1_cert() adds a
3633 certificate to a basic response and OCSP_basic_sign() signs a basic
3634 response with various flags. New helper functions ASN1_TIME_check()
3635 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
3636 (converts ASN1_TIME to GeneralizedTime).
3639 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
3640 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
3641 structure from a certificate. X509_pubkey_digest() digests the public_key
3642 contents: this is used in various key identifiers.
3645 *) Make sk_sort() tolerate a NULL argument.
3646 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
3648 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
3649 passed by the function are trusted implicitly. If any of them signed the
3650 response then it is assumed to be valid and is not verified.
3653 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
3654 to data. This was previously part of the PKCS7 ASN1 code. This
3655 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
3656 [Steve Henson, reported by Kenneth R. Robinette
3657 <support@securenetterm.com>]
3659 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
3660 routines: without these tracing memory leaks is very painful.
3661 Fix leaks in PKCS12 and PKCS7 routines.
3664 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
3665 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
3666 effectively meant GeneralizedTime would never be used. Now it
3667 is initialised to -1 but X509_time_adj() now has to check the value
3668 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
3669 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
3670 [Steve Henson, reported by Kenneth R. Robinette
3671 <support@securenetterm.com>]
3673 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
3674 result in a zero length in the ASN1_INTEGER structure which was
3675 not consistent with the structure when d2i_ASN1_INTEGER() was used
3676 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
3677 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
3678 where it did not print out a minus for negative ASN1_INTEGER.
3681 *) Add summary printout to ocsp utility. The various functions which
3682 convert status values to strings have been renamed to:
3683 OCSP_response_status_str(), OCSP_cert_status_str() and
3684 OCSP_crl_reason_str() and are no longer static. New options
3685 to verify nonce values and to disable verification. OCSP response
3686 printout format cleaned up.
3689 *) Add additional OCSP certificate checks. These are those specified
3690 in RFC2560. This consists of two separate checks: the CA of the
3691 certificate being checked must either be the OCSP signer certificate
3692 or the issuer of the OCSP signer certificate. In the latter case the
3693 OCSP signer certificate must contain the OCSP signing extended key
3694 usage. This check is performed by attempting to match the OCSP
3695 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
3696 in the OCSP_CERTID structures of the response.
3699 *) Initial OCSP certificate verification added to OCSP_basic_verify()
3700 and related routines. This uses the standard OpenSSL certificate
3701 verify routines to perform initial checks (just CA validity) and
3702 to obtain the certificate chain. Then additional checks will be
3703 performed on the chain. Currently the root CA is checked to see
3704 if it is explicitly trusted for OCSP signing. This is used to set
3705 a root CA as a global signing root: that is any certificate that
3706 chains to that CA is an acceptable OCSP signing certificate.
3709 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
3710 extensions from a separate configuration file.
3711 As when reading extensions from the main configuration file,
3712 the '-extensions ...' option may be used for specifying the
3714 [Massimiliano Pala <madwolf@comune.modena.it>]
3716 *) New OCSP utility. Allows OCSP requests to be generated or
3717 read. The request can be sent to a responder and the output
3718 parsed, outputed or printed in text form. Not complete yet:
3719 still needs to check the OCSP response validity.
3722 *) New subcommands for 'openssl ca':
3723 'openssl ca -status <serial>' prints the status of the cert with
3724 the given serial number (according to the index file).
3725 'openssl ca -updatedb' updates the expiry status of certificates
3727 [Massimiliano Pala <madwolf@comune.modena.it>]
3729 *) New '-newreq-nodes' command option to CA.pl. This is like
3730 '-newreq', but calls 'openssl req' with the '-nodes' option
3731 so that the resulting key is not encrypted.
3732 [Damien Miller <djm@mindrot.org>]
3734 *) New configuration for the GNU Hurd.
3735 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
3737 *) Initial code to implement OCSP basic response verify. This
3738 is currently incomplete. Currently just finds the signer's
3739 certificate and verifies the signature on the response.
3742 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
3743 value of OPENSSLDIR. This is available via the new '-d' option
3744 to 'openssl version', and is also included in 'openssl version -a'.
3747 *) Allowing defining memory allocation callbacks that will be given
3748 file name and line number information in additional arguments
3749 (a const char* and an int). The basic functionality remains, as
3750 well as the original possibility to just replace malloc(),
3751 realloc() and free() by functions that do not know about these
3752 additional arguments. To register and find out the current
3753 settings for extended allocation functions, the following
3754 functions are provided:
3756 CRYPTO_set_mem_ex_functions
3757 CRYPTO_set_locked_mem_ex_functions
3758 CRYPTO_get_mem_ex_functions
3759 CRYPTO_get_locked_mem_ex_functions
3761 These work the same way as CRYPTO_set_mem_functions and friends.
3762 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
3763 extended allocation function is enabled.
3764 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
3765 a conventional allocation function is enabled.
3766 [Richard Levitte, Bodo Moeller]
3768 *) Finish off removing the remaining LHASH function pointer casts.
3769 There should no longer be any prototype-casting required when using
3770 the LHASH abstraction, and any casts that remain are "bugs". See
3771 the callback types and macros at the head of lhash.h for details
3772 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
3775 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
3776 If /dev/[u]random devices are not available or do not return enough
3777 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
3779 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
3780 /etc/entropy will be queried once each in this sequence, quering stops
3781 when enough entropy was collected without querying more sockets.
3784 *) Change the Unix RAND_poll() variant to be able to poll several
3785 random devices, as specified by DEVRANDOM, until a sufficient amount
3786 of data has been collected. We spend at most 10 ms on each file
3787 (select timeout) and read in non-blocking mode. DEVRANDOM now
3788 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
3789 (previously it was just the string "/dev/urandom"), so on typical
3790 platforms the 10 ms delay will never occur.
3791 Also separate out the Unix variant to its own file, rand_unix.c.
3792 For VMS, there's a currently-empty rand_vms.c.
3795 *) Move OCSP client related routines to ocsp_cl.c. These
3796 provide utility functions which an application needing
3797 to issue a request to an OCSP responder and analyse the
3798 response will typically need: as opposed to those which an
3799 OCSP responder itself would need which will be added later.
3801 OCSP_request_sign() signs an OCSP request with an API similar
3802 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
3803 response. OCSP_response_get1_basic() extracts basic response
3804 from response. OCSP_resp_find_status(): finds and extracts status
3805 information from an OCSP_CERTID structure (which will be created
3806 when the request structure is built). These are built from lower
3807 level functions which work on OCSP_SINGLERESP structures but
3808 wont normally be used unless the application wishes to examine
3809 extensions in the OCSP response for example.
3811 Replace nonce routines with a pair of functions.
3812 OCSP_request_add1_nonce() adds a nonce value and optionally
3813 generates a random value. OCSP_check_nonce() checks the
3814 validity of the nonce in an OCSP response.
3817 *) Change function OCSP_request_add() to OCSP_request_add0_id().
3818 This doesn't copy the supplied OCSP_CERTID and avoids the
3819 need to free up the newly created id. Change return type
3820 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
3821 This can then be used to add extensions to the request.
3822 Deleted OCSP_request_new(), since most of its functionality
3823 is now in OCSP_REQUEST_new() (and the case insensitive name
3824 clash) apart from the ability to set the request name which
3825 will be added elsewhere.
3828 *) Update OCSP API. Remove obsolete extensions argument from
3829 various functions. Extensions are now handled using the new
3830 OCSP extension code. New simple OCSP HTTP function which
3831 can be used to send requests and parse the response.
3834 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
3835 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
3836 uses the special reorder version of SET OF to sort the attributes
3837 and reorder them to match the encoded order. This resolves a long
3838 standing problem: a verify on a PKCS7 structure just after signing
3839 it used to fail because the attribute order did not match the
3840 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
3841 it uses the received order. This is necessary to tolerate some broken
3842 software that does not order SET OF. This is handled by encoding
3843 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
3844 to produce the required SET OF.
3847 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
3848 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
3849 files to get correct declarations of the ASN.1 item variables.
3852 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
3853 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
3854 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
3855 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
3856 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
3857 ASN1_ITEM and no wrapper functions.
3860 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
3861 replace the old function pointer based I/O routines. Change most of
3862 the *_d2i_bio() and *_d2i_fp() functions to use these.
3865 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
3866 lines, recognice more "algorithms" that can be deselected, and make
3867 it complain about algorithm deselection that isn't recognised.
3870 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
3871 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
3872 to use new functions. Add NO_ASN1_OLD which can be set to remove
3873 some old style ASN1 functions: this can be used to determine if old
3874 code will still work when these eventually go away.
3877 *) New extension functions for OCSP structures, these follow the
3878 same conventions as certificates and CRLs.
3881 *) New function X509V3_add1_i2d(). This automatically encodes and
3882 adds an extension. Its behaviour can be customised with various
3883 flags to append, replace or delete. Various wrappers added for
3884 certifcates and CRLs.
3887 *) Fix to avoid calling the underlying ASN1 print routine when
3888 an extension cannot be parsed. Correct a typo in the
3889 OCSP_SERVICELOC extension. Tidy up print OCSP format.
3892 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
3893 entries for variables.
3896 *) Add functionality to apps/openssl.c for detecting locking
3897 problems: As the program is single-threaded, all we have
3898 to do is register a locking callback using an array for
3899 storing which locks are currently held by the program.
3902 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
3903 SSL_get_ex_data_X509_STORE_idx(), which is used in
3904 ssl_verify_cert_chain() and thus can be called at any time
3905 during TLS/SSL handshakes so that thread-safety is essential.
3906 Unfortunately, the ex_data design is not at all suited
3907 for multi-threaded use, so it probably should be abolished.
3910 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
3911 [Broadcom, tweaked and integrated by Geoff Thorpe]
3913 *) Move common extension printing code to new function
3914 X509V3_print_extensions(). Reorganise OCSP print routines and
3915 implement some needed OCSP ASN1 functions. Add OCSP extensions.
3918 *) New function X509_signature_print() to remove duplication in some
3922 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
3923 set (this was treated exactly the same as SET OF previously). This
3924 is used to reorder the STACK representing the structure to match the
3925 encoding. This will be used to get round a problem where a PKCS7
3926 structure which was signed could not be verified because the STACK
3927 order did not reflect the encoded order.
3930 *) Reimplement the OCSP ASN1 module using the new code.
3933 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
3934 for its ASN1 operations. The old style function pointers still exist
3935 for now but they will eventually go away.
3938 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
3939 completely replaces the old ASN1 functionality with a table driven
3940 encoder and decoder which interprets an ASN1_ITEM structure describing
3941 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
3942 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
3943 has also been converted to the new form.
3946 *) Change BN_mod_exp_recp so that negative moduli are tolerated
3947 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
3948 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
3949 for negative moduli.
3952 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
3953 of not touching the result's sign bit.
3956 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
3960 *) Changed the LHASH code to use prototypes for callbacks, and created
3961 macros to declare and implement thin (optionally static) functions
3962 that provide type-safety and avoid function pointer casting for the
3963 type-specific callbacks.
3966 *) Added Kerberos Cipher Suites to be used with TLS, as written in
3968 [Veers Staats <staatsvr@asc.hpc.mil>,
3969 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
3971 *) Reformat the FAQ so the different questions and answers can be divided
3972 in sections depending on the subject.
3975 *) Have the zlib compression code load ZLIB.DLL dynamically under
3979 *) New function BN_mod_sqrt for computing square roots modulo a prime
3980 (using the probabilistic Tonelli-Shanks algorithm unless
3981 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
3982 be handled deterministically).
3983 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3985 *) Make BN_mod_inverse faster by explicitly handling small quotients
3986 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
3987 512 bits], about 30% for larger ones [1024 or 2048 bits].)
3990 *) New function BN_kronecker.
3993 *) Fix BN_gcd so that it works on negative inputs; the result is
3994 positive unless both parameters are zero.
3995 Previously something reasonably close to an infinite loop was
3996 possible because numbers could be growing instead of shrinking
3997 in the implementation of Euclid's algorithm.
4000 *) Fix BN_is_word() and BN_is_one() macros to take into account the
4001 sign of the number in question.
4003 Fix BN_is_word(a,w) to work correctly for w == 0.
4005 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
4006 because its test if the absolute value of 'a' equals 'w'.
4007 Note that BN_abs_is_word does *not* handle w == 0 reliably;
4008 it exists mostly for use in the implementations of BN_is_zero(),
4009 BN_is_one(), and BN_is_word().
4012 *) New function BN_swap.
4015 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
4016 the exponentiation functions are more likely to produce reasonable
4017 results on negative inputs.
4020 *) Change BN_mod_mul so that the result is always non-negative.
4021 Previously, it could be negative if one of the factors was negative;
4022 I don't think anyone really wanted that behaviour.
4025 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
4026 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
4027 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
4028 and add new functions:
4037 BN_mod_lshift1_quick
4041 These functions always generate non-negative results.
4043 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
4044 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
4046 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
4047 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
4048 be reduced modulo m.
4049 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
4052 The following entry accidentily appeared in the CHANGES file
4053 distributed with OpenSSL 0.9.7. The modifications described in
4054 it do *not* apply to OpenSSL 0.9.7.
4056 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4057 was actually never needed) and in BN_mul(). The removal in BN_mul()
4058 required a small change in bn_mul_part_recursive() and the addition
4059 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4060 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4061 bn_sub_words() and bn_add_words() except they take arrays with
4066 *) In 'openssl passwd', verify passwords read from the terminal
4067 unless the '-salt' option is used (which usually means that
4068 verification would just waste user's time since the resulting
4069 hash is going to be compared with some given password hash)
4070 or the new '-noverify' option is used.
4072 This is an incompatible change, but it does not affect
4073 non-interactive use of 'openssl passwd' (passwords on the command
4074 line, '-stdin' option, '-in ...' option) and thus should not
4078 *) Remove all references to RSAref, since there's no more need for it.
4081 *) Make DSO load along a path given through an environment variable
4082 (SHLIB_PATH) with shl_load().
4085 *) Constify the ENGINE code as a result of BIGNUM constification.
4086 Also constify the RSA code and most things related to it. In a
4087 few places, most notable in the depth of the ASN.1 code, ugly
4088 casts back to non-const were required (to be solved at a later
4092 *) Make it so the openssl application has all engines loaded by default.
4095 *) Constify the BIGNUM routines a little more.
4098 *) Add the following functions:
4100 ENGINE_load_cswift()
4102 ENGINE_load_atalla()
4104 ENGINE_load_builtin_engines()
4106 That way, an application can itself choose if external engines that
4107 are built-in in OpenSSL shall ever be used or not. The benefit is
4108 that applications won't have to be linked with libdl or other dso
4109 libraries unless it's really needed.
4111 Changed 'openssl engine' to load all engines on demand.
4112 Changed the engine header files to avoid the duplication of some
4113 declarations (they differed!).
4116 *) 'openssl engine' can now list capabilities.
4119 *) Better error reporting in 'openssl engine'.
4122 *) Never call load_dh_param(NULL) in s_server.
4125 *) Add engine application. It can currently list engines by name and
4126 identity, and test if they are actually available.
4129 *) Improve RPM specification file by forcing symbolic linking and making
4130 sure the installed documentation is also owned by root.root.
4131 [Damien Miller <djm@mindrot.org>]
4133 *) Give the OpenSSL applications more possibilities to make use of
4134 keys (public as well as private) handled by engines.
4137 *) Add OCSP code that comes from CertCo.
4140 *) Add VMS support for the Rijndael code.
4143 *) Added untested support for Nuron crypto accelerator.
4146 *) Add support for external cryptographic devices. This code was
4147 previously distributed separately as the "engine" branch.
4148 [Geoff Thorpe, Richard Levitte]
4150 *) Rework the filename-translation in the DSO code. It is now possible to
4151 have far greater control over how a "name" is turned into a filename
4152 depending on the operating environment and any oddities about the
4153 different shared library filenames on each system.
4156 *) Support threads on FreeBSD-elf in Configure.
4159 *) Fix for SHA1 assembly problem with MASM: it produces
4160 warnings about corrupt line number information when assembling
4161 with debugging information. This is caused by the overlapping
4163 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
4166 NCONF_get_number() has no error checking at all. As a replacement,
4167 NCONF_get_number_e() is defined (_e for "error checking") and is
4168 promoted strongly. The old NCONF_get_number is kept around for
4169 binary backward compatibility.
4170 Make it possible for methods to load from something other than a BIO,
4171 by providing a function pointer that is given a name instead of a BIO.
4172 For example, this could be used to load configuration data from an
4176 *) Fix for non blocking accept BIOs. Added new I/O special reason
4177 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
4178 with non blocking I/O was not possible because no retry code was
4179 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
4183 *) Added the beginnings of Rijndael support.
4186 *) Fix for bug in DirectoryString mask setting. Add support for
4187 X509_NAME_print_ex() in 'req' and X509_print_ex() function
4188 to allow certificate printing to more controllable, additional
4189 'certopt' option to 'x509' to allow new printing options to be
4193 *) Clean old EAY MD5 hack from e_os.h.
4196 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
4198 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4199 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4200 [Joe Orton, Steve Henson]
4202 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
4204 *) Fix additional bug revealed by the NISCC test suite:
4206 Stop bug triggering large recursion when presented with
4207 certain ASN.1 tags (CVE-2003-0851)
4210 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
4212 *) Fix various bugs revealed by running the NISCC test suite:
4214 Stop out of bounds reads in the ASN1 code when presented with
4215 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4217 If verify callback ignores invalid public key errors don't try to check
4218 certificate signature with the NULL public key.
4222 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4223 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4227 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4228 extra data after the compression methods not only for TLS 1.0
4229 but also for SSL 3.0 (as required by the specification).
4230 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4232 *) Change X509_certificate_type() to mark the key as exported/exportable
4233 when it's 512 *bits* long, not 512 bytes.
4236 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
4238 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4239 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4240 a protocol version number mismatch like a decryption error
4241 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4244 *) Turn on RSA blinding by default in the default implementation
4245 to avoid a timing attack. Applications that don't want it can call
4246 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4247 They would be ill-advised to do so in most cases.
4248 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4250 *) Change RSA blinding code so that it works when the PRNG is not
4251 seeded (in this case, the secret RSA exponent is abused as
4252 an unpredictable seed -- if it is not unpredictable, there
4253 is no point in blinding anyway). Make RSA blinding thread-safe
4254 by remembering the creator's thread ID in rsa->blinding and
4255 having all other threads use local one-time blinding factors
4256 (this requires more computation than sharing rsa->blinding, but
4257 avoids excessive locking; and if an RSA object is not shared
4258 between threads, blinding will still be very fast).
4261 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
4263 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4264 via timing by performing a MAC computation even if incorrrect
4265 block cipher padding has been found. This is a countermeasure
4266 against active attacks where the attacker has to distinguish
4267 between bad padding and a MAC verification error. (CVE-2003-0078)
4269 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4270 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4271 Martin Vuagnoux (EPFL, Ilion)]
4273 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
4275 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
4276 memory from it's contents. This is done with a counter that will
4277 place alternating values in each byte. This can be used to solve
4278 two issues: 1) the removal of calls to memset() by highly optimizing
4279 compilers, and 2) cleansing with other values than 0, since those can
4280 be read through on certain media, for example a swap space on disk.
4283 *) Bugfix: client side session caching did not work with external caching,
4284 because the session->cipher setting was not restored when reloading
4285 from the external cache. This problem was masked, when
4286 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
4287 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
4290 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
4291 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
4292 [Zeev Lieber <zeev-l@yahoo.com>]
4294 *) Undo an undocumented change introduced in 0.9.6e which caused
4295 repeated calls to OpenSSL_add_all_ciphers() and
4296 OpenSSL_add_all_digests() to be ignored, even after calling
4300 *) Change the default configuration reader to deal with last line not
4301 being properly terminated.
4304 *) Change X509_NAME_cmp() so it applies the special rules on handling
4305 DN values that are of type PrintableString, as well as RDNs of type
4306 emailAddress where the value has the type ia5String.
4307 [stefank@valicert.com via Richard Levitte]
4309 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
4310 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
4311 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
4312 the bitwise-OR of the two for use by the majority of applications
4313 wanting this behaviour, and update the docs. The documented
4314 behaviour and actual behaviour were inconsistent and had been
4315 changing anyway, so this is more a bug-fix than a behavioural
4317 [Geoff Thorpe, diagnosed by Nadav Har'El]
4319 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
4320 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
4323 *) Fix initialization code race conditions in
4324 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
4325 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
4326 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
4327 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
4328 ssl2_get_cipher_by_char(),
4329 ssl3_get_cipher_by_char().
4330 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
4332 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
4333 the cached sessions are flushed, as the remove_cb() might use ex_data
4334 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
4335 (see [openssl.org #212]).
4336 [Geoff Thorpe, Lutz Jaenicke]
4338 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
4339 length, instead of the encoding length to d2i_ASN1_OBJECT.
4342 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
4344 *) [In 0.9.6g-engine release:]
4345 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
4346 [Lynn Gazis <lgazis@rainbow.com>]
4348 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
4350 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
4351 and get fix the header length calculation.
4352 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
4353 Alon Kantor <alonk@checkpoint.com> (and others),
4356 *) Use proper error handling instead of 'assertions' in buffer
4357 overflow checks added in 0.9.6e. This prevents DoS (the
4358 assertions could call abort()).
4359 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
4361 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
4363 *) Add various sanity checks to asn1_get_length() to reject
4364 the ASN1 length bytes if they exceed sizeof(long), will appear
4365 negative or the content length exceeds the length of the
4367 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4369 *) Fix cipher selection routines: ciphers without encryption had no flags
4370 for the cipher strength set and where therefore not handled correctly
4371 by the selection routines (PR #130).
4374 *) Fix EVP_dsa_sha macro.
4378 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
4379 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
4380 that was added in OpenSSL 0.9.6d.
4382 As the countermeasure turned out to be incompatible with some
4383 broken SSL implementations, the new option is part of SSL_OP_ALL.
4384 SSL_OP_ALL is usually employed when compatibility with weird SSL
4385 implementations is desired (e.g. '-bugs' option to 's_client' and
4386 's_server'), so the new option is automatically set in many
4390 *) Changes in security patch:
4392 Changes marked "(CHATS)" were sponsored by the Defense Advanced
4393 Research Projects Agency (DARPA) and Air Force Research Laboratory,
4394 Air Force Materiel Command, USAF, under agreement number
4397 *) Add various sanity checks to asn1_get_length() to reject
4398 the ASN1 length bytes if they exceed sizeof(long), will appear
4399 negative or the content length exceeds the length of the
4400 supplied buffer. (CVE-2002-0659)
4401 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4403 *) Assertions for various potential buffer overflows, not known to
4405 [Ben Laurie (CHATS)]
4407 *) Various temporary buffers to hold ASCII versions of integers were
4408 too small for 64 bit platforms. (CVE-2002-0655)
4409 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
4411 *) Remote buffer overflow in SSL3 protocol - an attacker could
4412 supply an oversized session ID to a client. (CVE-2002-0656)
4413 [Ben Laurie (CHATS)]
4415 *) Remote buffer overflow in SSL2 protocol - an attacker could
4416 supply an oversized client master key. (CVE-2002-0656)
4417 [Ben Laurie (CHATS)]
4419 Changes between 0.9.6c and 0.9.6d [9 May 2002]
4421 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
4422 encoded as NULL) with id-dsa-with-sha1.
4423 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
4425 *) Check various X509_...() return values in apps/req.c.
4426 [Nils Larsch <nla@trustcenter.de>]
4428 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
4429 an end-of-file condition would erronously be flagged, when the CRLF
4430 was just at the end of a processed block. The bug was discovered when
4431 processing data through a buffering memory BIO handing the data to a
4432 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
4433 <ptsekov@syntrex.com> and Nedelcho Stanev.
4436 *) Implement a countermeasure against a vulnerability recently found
4437 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
4438 before application data chunks to avoid the use of known IVs
4439 with data potentially chosen by the attacker.
4442 *) Fix length checks in ssl3_get_client_hello().
4445 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
4446 to prevent ssl3_read_internal() from incorrectly assuming that
4447 ssl3_read_bytes() found application data while handshake
4448 processing was enabled when in fact s->s3->in_read_app_data was
4449 merely automatically cleared during the initial handshake.
4450 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
4452 *) Fix object definitions for Private and Enterprise: they were not
4453 recognized in their shortname (=lowercase) representation. Extend
4454 obj_dat.pl to issue an error when using undefined keywords instead
4455 of silently ignoring the problem (Svenning Sorensen
4456 <sss@sss.dnsalias.net>).
4459 *) Fix DH_generate_parameters() so that it works for 'non-standard'
4460 generators, i.e. generators other than 2 and 5. (Previously, the
4461 code did not properly initialise the 'add' and 'rem' values to
4462 BN_generate_prime().)
4464 In the new general case, we do not insist that 'generator' is
4465 actually a primitive root: This requirement is rather pointless;
4466 a generator of the order-q subgroup is just as good, if not
4470 *) Map new X509 verification errors to alerts. Discovered and submitted by
4471 Tom Wu <tom@arcot.com>.
4474 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
4475 returning non-zero before the data has been completely received
4476 when using non-blocking I/O.
4477 [Bodo Moeller; problem pointed out by John Hughes]
4479 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
4480 [Ben Laurie, Lutz Jaenicke]
4482 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
4483 Yoram Zahavi <YoramZ@gilian.com>).
4486 *) Add information about CygWin 1.3 and on, and preserve proper
4487 configuration for the versions before that.
4488 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4490 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
4491 check whether we deal with a copy of a session and do not delete from
4492 the cache in this case. Problem reported by "Izhar Shoshani Levi"
4493 <izhar@checkpoint.com>.
4496 *) Do not store session data into the internal session cache, if it
4497 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
4498 flag is set). Proposed by Aslam <aslam@funk.com>.
4501 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
4505 *) [In 0.9.6d-engine release:]
4506 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4507 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4509 *) Add the configuration target linux-s390x.
4510 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
4512 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
4513 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
4514 variable as an indication that a ClientHello message has been
4515 received. As the flag value will be lost between multiple
4516 invocations of ssl3_accept when using non-blocking I/O, the
4517 function may not be aware that a handshake has actually taken
4518 place, thus preventing a new session from being added to the
4521 To avoid this problem, we now set s->new_session to 2 instead of
4522 using a local variable.
4523 [Lutz Jaenicke, Bodo Moeller]
4525 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
4526 if the SSL_R_LENGTH_MISMATCH error is detected.
4527 [Geoff Thorpe, Bodo Moeller]
4529 *) New 'shared_ldflag' column in Configure platform table.
4532 *) Fix EVP_CIPHER_mode macro.
4533 ["Dan S. Camper" <dan@bti.net>]
4535 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
4536 type, we must throw them away by setting rr->length to 0.
4537 [D P Chang <dpc@qualys.com>]
4539 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
4541 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
4542 <Dominikus.Scherkl@biodata.com>. (The previous implementation
4543 worked incorrectly for those cases where range = 10..._2 and
4544 3*range is two bits longer than range.)
4547 *) Only add signing time to PKCS7 structures if it is not already
4551 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
4552 OBJ_ld_ce should be OBJ_id_ce.
4553 Also some ip-pda OIDs in crypto/objects/objects.txt were
4554 incorrect (cf. RFC 3039).
4555 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
4557 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
4558 returns early because it has nothing to do.
4559 [Andy Schneider <andy.schneider@bjss.co.uk>]
4561 *) [In 0.9.6c-engine release:]
4562 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
4563 [Andy Schneider <andy.schneider@bjss.co.uk>]
4565 *) [In 0.9.6c-engine release:]
4566 Add support for Cryptographic Appliance's keyserver technology.
4567 (Use engine 'keyclient')
4568 [Cryptographic Appliances and Geoff Thorpe]
4570 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
4571 is called via tools/c89.sh because arguments have to be
4572 rearranged (all '-L' options must appear before the first object
4574 [Richard Shapiro <rshapiro@abinitio.com>]
4576 *) [In 0.9.6c-engine release:]
4577 Add support for Broadcom crypto accelerator cards, backported
4579 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
4581 *) [In 0.9.6c-engine release:]
4582 Add support for SureWare crypto accelerator cards from
4583 Baltimore Technologies. (Use engine 'sureware')
4584 [Baltimore Technologies and Mark Cox]
4586 *) [In 0.9.6c-engine release:]
4587 Add support for crypto accelerator cards from Accelerated
4588 Encryption Processing, www.aep.ie. (Use engine 'aep')
4589 [AEP Inc. and Mark Cox]
4591 *) Add a configuration entry for gcc on UnixWare.
4592 [Gary Benson <gbenson@redhat.com>]
4594 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
4595 messages are stored in a single piece (fixed-length part and
4596 variable-length part combined) and fix various bugs found on the way.
4599 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
4600 instead. BIO_gethostbyname() does not know what timeouts are
4601 appropriate, so entries would stay in cache even when they have
4603 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
4605 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
4606 faced with a pathologically small ClientHello fragment that does
4607 not contain client_version: Instead of aborting with an error,
4608 simply choose the highest available protocol version (i.e.,
4609 TLS 1.0 unless it is disabled). In practice, ClientHello
4610 messages are never sent like this, but this change gives us
4611 strictly correct behaviour at least for TLS.
4614 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
4615 never resets s->method to s->ctx->method when called from within
4616 one of the SSL handshake functions.
4617 [Bodo Moeller; problem pointed out by Niko Baric]
4619 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
4620 (sent using the client's version number) if client_version is
4621 smaller than the protocol version in use. Also change
4622 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
4623 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
4624 the client will at least see that alert.
4627 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
4631 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
4632 client receives HelloRequest while in a handshake.
4633 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
4635 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
4636 should end in 'break', not 'goto end' which circuments various
4637 cleanups done in state SSL_ST_OK. But session related stuff
4638 must be disabled for SSL_ST_OK in the case that we just sent a
4641 Also avoid some overhead by not calling ssl_init_wbio_buffer()
4642 before just sending a HelloRequest.
4643 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
4645 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
4646 reveal whether illegal block cipher padding was found or a MAC
4647 verification error occured. (Neither SSLerr() codes nor alerts
4648 are directly visible to potential attackers, but the information
4649 may leak via logfiles.)
4651 Similar changes are not required for the SSL 2.0 implementation
4652 because the number of padding bytes is sent in clear for SSL 2.0,
4653 and the extra bytes are just ignored. However ssl/s2_pkt.c
4654 failed to verify that the purported number of padding bytes is in
4658 *) Add OpenUNIX-8 support including shared libraries
4659 (Boyd Lynn Gerber <gerberb@zenez.com>).
4662 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
4663 'wristwatch attack' using huge encoding parameters (cf.
4664 James H. Manger's CRYPTO 2001 paper). Note that the
4665 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
4666 encoding parameters and hence was not vulnerable.
4669 *) BN_sqr() bug fix.
4670 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
4672 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
4673 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
4674 followed by modular reduction.
4675 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
4677 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
4678 equivalent based on BN_pseudo_rand() instead of BN_rand().
4681 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
4682 This function was broken, as the check for a new client hello message
4683 to handle SGC did not allow these large messages.
4684 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
4687 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
4690 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
4691 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
4694 *) Rework the configuration and shared library support for Tru64 Unix.
4695 The configuration part makes use of modern compiler features and
4696 still retains old compiler behavior for those that run older versions
4697 of the OS. The shared library support part includes a variant that
4698 uses the RPATH feature, and is available through the special
4699 configuration target "alpha-cc-rpath", which will never be selected
4701 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
4703 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
4704 with the same message size as in ssl3_get_certificate_request().
4705 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
4706 messages might inadvertently be reject as too long.
4707 [Petr Lampa <lampa@fee.vutbr.cz>]
4709 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
4712 *) Modified SSL library such that the verify_callback that has been set
4713 specificly for an SSL object with SSL_set_verify() is actually being
4714 used. Before the change, a verify_callback set with this function was
4715 ignored and the verify_callback() set in the SSL_CTX at the time of
4716 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
4717 to allow the necessary settings.
4720 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
4721 explicitly to NULL, as at least on Solaris 8 this seems not always to be
4722 done automatically (in contradiction to the requirements of the C
4723 standard). This made problems when used from OpenSSH.
4726 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
4727 dh->length and always used
4729 BN_rand_range(priv_key, dh->p).
4731 BN_rand_range() is not necessary for Diffie-Hellman, and this
4732 specific range makes Diffie-Hellman unnecessarily inefficient if
4733 dh->length (recommended exponent length) is much smaller than the
4734 length of dh->p. We could use BN_rand_range() if the order of
4735 the subgroup was stored in the DH structure, but we only have
4740 BN_rand(priv_key, l, ...)
4742 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
4748 RSA_eay_public_encrypt
4749 RSA_eay_private_decrypt
4750 RSA_eay_private_encrypt (signing)
4751 RSA_eay_public_decrypt (signature verification)
4753 (default implementations for RSA_public_encrypt,
4754 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
4755 always reject numbers >= n.
4758 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
4759 to synchronize access to 'locking_thread'. This is necessary on
4760 systems where access to 'locking_thread' (an 'unsigned long'
4761 variable) is not atomic.
4764 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
4765 *before* setting the 'crypto_lock_rand' flag. The previous code had
4766 a race condition if 0 is a valid thread ID.
4767 [Travis Vitek <vitek@roguewave.com>]
4769 *) Add support for shared libraries under Irix.
4770 [Albert Chin-A-Young <china@thewrittenword.com>]
4772 *) Add configuration option to build on Linux on both big-endian and
4774 [Ralf Baechle <ralf@uni-koblenz.de>]
4776 *) Add the possibility to create shared libraries on HP-UX.
4779 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
4781 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
4782 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
4783 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
4784 PRNG state recovery was possible based on the output of
4785 one PRNG request appropriately sized to gain knowledge on
4786 'md' followed by enough consecutive 1-byte PRNG requests
4787 to traverse all of 'state'.
4789 1. When updating 'md_local' (the current thread's copy of 'md')
4790 during PRNG output generation, hash all of the previous
4791 'md_local' value, not just the half used for PRNG output.
4793 2. Make the number of bytes from 'state' included into the hash
4794 independent from the number of PRNG bytes requested.
4796 The first measure alone would be sufficient to avoid
4797 Markku-Juhani's attack. (Actually it had never occurred
4798 to me that the half of 'md_local' used for chaining was the
4799 half from which PRNG output bytes were taken -- I had always
4800 assumed that the secret half would be used.) The second
4801 measure makes sure that additional data from 'state' is never
4802 mixed into 'md_local' in small portions; this heuristically
4803 further strengthens the PRNG.
4806 *) Fix crypto/bn/asm/mips3.s.
4809 *) When only the key is given to "enc", the IV is undefined. Print out
4810 an error message in this case.
4813 *) Handle special case when X509_NAME is empty in X509 printing routines.
4816 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
4817 positive and less than q.
4820 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
4821 used: it isn't thread safe and the add_lock_callback should handle
4823 [Paul Rose <Paul.Rose@bridge.com>]
4825 *) Verify that incoming data obeys the block size in
4826 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
4830 [Ulf Möller, Bodo Möller]
4832 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
4833 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
4834 when fixing the server behaviour for backwards-compatible 'client
4835 hello' messages. (Note that the attack is impractical against
4836 SSL 3.0 and TLS 1.0 anyway because length and version checking
4837 means that the probability of guessing a valid ciphertext is
4838 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
4841 Before 0.9.5, the countermeasure (hide the error by generating a
4842 random 'decryption result') did not work properly because
4843 ERR_clear_error() was missing, meaning that SSL_get_error() would
4844 detect the supposedly ignored error.
4846 Both problems are now fixed.
4849 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
4850 (previously it was 1024).
4853 *) Fix for compatibility mode trust settings: ignore trust settings
4854 unless some valid trust or reject settings are present.
4857 *) Fix for blowfish EVP: its a variable length cipher.
4860 *) Fix various bugs related to DSA S/MIME verification. Handle missing
4861 parameters in DSA public key structures and return an error in the
4862 DSA routines if parameters are absent.
4865 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
4866 in the current directory if neither $RANDFILE nor $HOME was set.
4867 RAND_file_name() in 0.9.6a returned NULL in this case. This has
4868 caused some confusion to Windows users who haven't defined $HOME.
4869 Thus RAND_file_name() is changed again: e_os.h can define a
4870 DEFAULT_HOME, which will be used if $HOME is not set.
4871 For Windows, we use "C:"; on other platforms, we still require
4872 environment variables.
4874 *) Move 'if (!initialized) RAND_poll()' into regions protected by
4875 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
4876 having multiple threads call RAND_poll() concurrently.
4879 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
4880 combination of a flag and a thread ID variable.
4881 Otherwise while one thread is in ssleay_rand_bytes (which sets the
4882 flag), *other* threads can enter ssleay_add_bytes without obeying
4883 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
4884 that they do not hold after the first thread unsets add_do_not_lock).
4887 *) Change bctest again: '-x' expressions are not available in all
4891 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
4893 *) Fix a couple of memory leaks in PKCS7_dataDecode()
4894 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
4896 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
4897 the default extension for executables, if any. Also, make the perl
4898 scripts that use symlink() to test if it really exists and use "cp"
4899 if it doesn't. All this made OpenSSL compilable and installable in
4903 *) Fix for asn1_GetSequence() for indefinite length constructed data.
4904 If SEQUENCE is length is indefinite just set c->slen to the total
4905 amount of data available.
4906 [Steve Henson, reported by shige@FreeBSD.org]
4907 [This change does not apply to 0.9.7.]
4909 *) Change bctest to avoid here-documents inside command substitution
4910 (workaround for FreeBSD /bin/sh bug).
4911 For compatibility with Ultrix, avoid shell functions (introduced
4912 in the bctest version that searches along $PATH).
4915 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
4916 with des_encrypt() defined on some operating systems, like Solaris
4920 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
4921 On the Importance of Eliminating Errors in Cryptographic
4922 Computations, J. Cryptology 14 (2001) 2, 101-119,
4923 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
4926 *) MIPS assembler BIGNUM division bug fix.
4929 *) Disabled incorrect Alpha assembler code.
4932 *) Fix PKCS#7 decode routines so they correctly update the length
4933 after reading an EOC for the EXPLICIT tag.
4935 [This change does not apply to 0.9.7.]
4937 *) Fix bug in PKCS#12 key generation routines. This was triggered
4938 if a 3DES key was generated with a 0 initial byte. Include
4939 PKCS12_BROKEN_KEYGEN compilation option to retain the old
4940 (but broken) behaviour.
4943 *) Enhance bctest to search for a working bc along $PATH and print
4945 [Tim Rice <tim@multitalents.net> via Richard Levitte]
4947 *) Fix memory leaks in err.c: free err_data string if necessary;
4948 don't write to the wrong index in ERR_set_error_data.
4951 *) Implement ssl23_peek (analogous to ssl23_read), which previously
4955 *) Replace rdtsc with _emit statements for VC++ version 5.
4956 [Jeremy Cooper <jeremy@baymoo.org>]
4958 *) Make it possible to reuse SSLv2 sessions.
4961 *) In copy_email() check for >= 0 as a return value for
4962 X509_NAME_get_index_by_NID() since 0 is a valid index.
4963 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
4965 *) Avoid coredump with unsupported or invalid public keys by checking if
4966 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
4967 PKCS7_verify() fails with non detached data.
4970 *) Don't use getenv in library functions when run as setuid/setgid.
4971 New function OPENSSL_issetugid().
4974 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
4975 due to incorrect handling of multi-threading:
4977 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
4979 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
4981 3. Count how many times MemCheck_off() has been called so that
4982 nested use can be treated correctly. This also avoids
4983 inband-signalling in the previous code (which relied on the
4984 assumption that thread ID 0 is impossible).
4987 *) Add "-rand" option also to s_client and s_server.
4990 *) Fix CPU detection on Irix 6.x.
4991 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
4992 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4994 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
4997 [This change does not apply to 0.9.7.]
4999 *) Use the cached encoding of an X509_NAME structure rather than
5000 copying it. This is apparently the reason for the libsafe "errors"
5001 but the code is actually correct.
5004 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
5005 Bleichenbacher's DSA attack.
5006 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
5007 to be set and top=0 forces the highest bit to be set; top=-1 is new
5008 and leaves the highest bit random.
5009 [Ulf Moeller, Bodo Moeller]
5011 *) In the NCONF_...-based implementations for CONF_... queries
5012 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
5013 a temporary CONF structure with the data component set to NULL
5014 (which gives segmentation faults in lh_retrieve).
5015 Instead, use NULL for the CONF pointer in CONF_get_string and
5016 CONF_get_number (which may use environment variables) and directly
5017 return NULL from CONF_get_section.
5020 *) Fix potential buffer overrun for EBCDIC.
5023 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
5024 keyUsage if basicConstraints absent for a CA.
5027 *) Make SMIME_write_PKCS7() write mail header values with a format that
5028 is more generally accepted (no spaces before the semicolon), since
5029 some programs can't parse those values properly otherwise. Also make
5030 sure BIO's that break lines after each write do not create invalid
5034 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
5035 macros previously used would not encode an empty SEQUENCE OF
5036 and break the signature.
5038 [This change does not apply to 0.9.7.]
5040 *) Zero the premaster secret after deriving the master secret in
5044 *) Add some EVP_add_digest_alias registrations (as found in
5045 OpenSSL_add_all_digests()) to SSL_library_init()
5046 aka OpenSSL_add_ssl_algorithms(). This provides improved
5047 compatibility with peers using X.509 certificates
5048 with unconventional AlgorithmIdentifier OIDs.
5051 *) Fix for Irix with NO_ASM.
5052 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
5054 *) ./config script fixes.
5055 [Ulf Moeller, Richard Levitte]
5057 *) Fix 'openssl passwd -1'.
5060 *) Change PKCS12_key_gen_asc() so it can cope with non null
5061 terminated strings whose length is passed in the passlen
5062 parameter, for example from PEM callbacks. This was done
5063 by adding an extra length parameter to asc2uni().
5064 [Steve Henson, reported by <oddissey@samsung.co.kr>]
5066 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
5067 call failed, free the DSA structure.
5070 *) Fix to uni2asc() to cope with zero length Unicode strings.
5071 These are present in some PKCS#12 files.
5074 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
5075 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
5076 when writing a 32767 byte record.
5077 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
5079 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
5080 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
5082 (RSA objects have a reference count access to which is protected
5083 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
5084 so they are meant to be shared between threads.)
5085 [Bodo Moeller, Geoff Thorpe; original patch submitted by
5086 "Reddie, Steven" <Steven.Reddie@ca.com>]
5088 *) Fix a deadlock in CRYPTO_mem_leaks().
5091 *) Use better test patterns in bntest.
5094 *) rand_win.c fix for Borland C.
5097 *) BN_rshift bugfix for n == 0.
5100 *) Add a 'bctest' script that checks for some known 'bc' bugs
5101 so that 'make test' does not abort just because 'bc' is broken.
5104 *) Store verify_result within SSL_SESSION also for client side to
5105 avoid potential security hole. (Re-used sessions on the client side
5106 always resulted in verify_result==X509_V_OK, not using the original
5107 result of the server certificate verification.)
5110 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
5111 SSL3_RT_APPLICATION_DATA, return 0.
5112 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
5116 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
5117 releases, have been re-implemented by renaming the previous
5118 implementations of ssl2_read and ssl3_read to ssl2_read_internal
5119 and ssl3_read_internal, respectively, and adding 'peek' parameters
5120 to them. The new ssl[23]_{read,peek} functions are calls to
5121 ssl[23]_read_internal with the 'peek' flag set appropriately.
5122 A 'peek' parameter has also been added to ssl3_read_bytes, which
5123 does the actual work for ssl3_read_internal.
5126 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
5127 the method-specific "init()" handler. Also clean up ex_data after
5128 calling the method-specific "finish()" handler. Previously, this was
5129 happening the other way round.
5132 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
5133 The previous value, 12, was not always sufficient for BN_mod_exp().
5136 *) Make sure that shared libraries get the internal name engine with
5137 the full version number and not just 0. This should mark the
5138 shared libraries as not backward compatible. Of course, this should
5139 be changed again when we can guarantee backward binary compatibility.
5142 *) Fix typo in get_cert_by_subject() in by_dir.c
5143 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
5145 *) Rework the system to generate shared libraries:
5147 - Make note of the expected extension for the shared libraries and
5148 if there is a need for symbolic links from for example libcrypto.so.0
5149 to libcrypto.so.0.9.7. There is extended info in Configure for
5152 - Make as few rebuilds of the shared libraries as possible.
5154 - Still avoid linking the OpenSSL programs with the shared libraries.
5156 - When installing, install the shared libraries separately from the
5160 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
5162 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
5163 and not in SSL_clear because the latter is also used by the
5164 accept/connect functions; previously, the settings made by
5165 SSL_set_read_ahead would be lost during the handshake.
5166 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
5168 *) Correct util/mkdef.pl to be selective about disabled algorithms.
5169 Previously, it would create entries for disableed algorithms no
5173 *) Added several new manual pages for SSL_* function.
5176 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
5178 *) In ssl23_get_client_hello, generate an error message when faced
5179 with an initial SSL 3.0/TLS record that is too small to contain the
5180 first two bytes of the ClientHello message, i.e. client_version.
5181 (Note that this is a pathologic case that probably has never happened
5182 in real life.) The previous approach was to use the version number
5183 from the record header as a substitute; but our protocol choice
5184 should not depend on that one because it is not authenticated
5185 by the Finished messages.
5188 *) More robust randomness gathering functions for Windows.
5189 [Jeffrey Altman <jaltman@columbia.edu>]
5191 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
5192 not set then we don't setup the error code for issuer check errors
5193 to avoid possibly overwriting other errors which the callback does
5194 handle. If an application does set the flag then we assume it knows
5195 what it is doing and can handle the new informational codes
5199 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
5200 a general "ANY" type, as such it should be able to decode anything
5201 including tagged types. However it didn't check the class so it would
5202 wrongly interpret tagged types in the same way as their universal
5203 counterpart and unknown types were just rejected. Changed so that the
5204 tagged and unknown types are handled in the same way as a SEQUENCE:
5205 that is the encoding is stored intact. There is also a new type
5206 "V_ASN1_OTHER" which is used when the class is not universal, in this
5207 case we have no idea what the actual type is so we just lump them all
5211 *) On VMS, stdout may very well lead to a file that is written to
5212 in a record-oriented fashion. That means that every write() will
5213 write a separate record, which will be read separately by the
5214 programs trying to read from it. This can be very confusing.
5216 The solution is to put a BIO filter in the way that will buffer
5217 text until a linefeed is reached, and then write everything a
5218 line at a time, so every record written will be an actual line,
5219 not chunks of lines and not (usually doesn't happen, but I've
5220 seen it once) several lines in one record. BIO_f_linebuffer() is
5223 Currently, it's a VMS-only method, because that's where it has
5224 been tested well enough.
5227 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
5228 it can return incorrect results.
5229 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
5230 but it was in 0.9.6-beta[12].)
5233 *) Disable the check for content being present when verifying detached
5234 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
5235 include zero length content when signing messages.
5238 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
5239 BIO_ctrl (for BIO pairs).
5242 *) Add DSO method for VMS.
5245 *) Bug fix: Montgomery multiplication could produce results with the
5249 *) Add RPM specification openssl.spec and modify it to build three
5250 packages. The default package contains applications, application
5251 documentation and run-time libraries. The devel package contains
5252 include files, static libraries and function documentation. The
5253 doc package contains the contents of the doc directory. The original
5254 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
5257 *) Add a large number of documentation files for many SSL routines.
5258 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5260 *) Add a configuration entry for Sony News 4.
5261 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
5263 *) Don't set the two most significant bits to one when generating a
5264 random number < q in the DSA library.
5267 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
5268 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
5269 the underlying transport is blocking) if a handshake took place.
5270 (The default behaviour is needed by applications such as s_client
5271 and s_server that use select() to determine when to use SSL_read;
5272 but for applications that know in advance when to expect data, it
5273 just makes things more complicated.)
5276 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
5280 *) Add a few more EBCDIC conditionals that make `req' and `x509'
5281 work better on such systems.
5282 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5284 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
5285 Update PKCS12_parse() so it copies the friendlyName and the
5286 keyid to the certificates aux info.
5289 *) Fix bug in PKCS7_verify() which caused an infinite loop
5290 if there was more than one signature.
5291 [Sven Uszpelkat <su@celocom.de>]
5293 *) Major change in util/mkdef.pl to include extra information
5294 about each symbol, as well as presentig variables as well
5295 as functions. This change means that there's n more need
5296 to rebuild the .num files when some algorithms are excluded.
5299 *) Allow the verify time to be set by an application,
5300 rather than always using the current time.
5303 *) Phase 2 verify code reorganisation. The certificate
5304 verify code now looks up an issuer certificate by a
5305 number of criteria: subject name, authority key id
5306 and key usage. It also verifies self signed certificates
5307 by the same criteria. The main comparison function is
5308 X509_check_issued() which performs these checks.
5310 Lot of changes were necessary in order to support this
5311 without completely rewriting the lookup code.
5313 Authority and subject key identifier are now cached.
5315 The LHASH 'certs' is X509_STORE has now been replaced
5316 by a STACK_OF(X509_OBJECT). This is mainly because an
5317 LHASH can't store or retrieve multiple objects with
5318 the same hash value.
5320 As a result various functions (which were all internal
5321 use only) have changed to handle the new X509_STORE
5322 structure. This will break anything that messed round
5323 with X509_STORE internally.
5325 The functions X509_STORE_add_cert() now checks for an
5326 exact match, rather than just subject name.
5328 The X509_STORE API doesn't directly support the retrieval
5329 of multiple certificates matching a given criteria, however
5330 this can be worked round by performing a lookup first
5331 (which will fill the cache with candidate certificates)
5332 and then examining the cache for matches. This is probably
5333 the best we can do without throwing out X509_LOOKUP
5334 entirely (maybe later...).
5336 The X509_VERIFY_CTX structure has been enhanced considerably.
5338 All certificate lookup operations now go via a get_issuer()
5339 callback. Although this currently uses an X509_STORE it
5340 can be replaced by custom lookups. This is a simple way
5341 to bypass the X509_STORE hackery necessary to make this
5342 work and makes it possible to use more efficient techniques
5343 in future. A very simple version which uses a simple
5344 STACK for its trusted certificate store is also provided
5345 using X509_STORE_CTX_trusted_stack().
5347 The verify_cb() and verify() callbacks now have equivalents
5348 in the X509_STORE_CTX structure.
5350 X509_STORE_CTX also has a 'flags' field which can be used
5351 to customise the verify behaviour.
5354 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
5355 excludes S/MIME capabilities.
5358 *) When a certificate request is read in keep a copy of the
5359 original encoding of the signed data and use it when outputing
5360 again. Signatures then use the original encoding rather than
5361 a decoded, encoded version which may cause problems if the
5362 request is improperly encoded.
5365 *) For consistency with other BIO_puts implementations, call
5366 buffer_write(b, ...) directly in buffer_puts instead of calling
5369 In BIO_puts, increment b->num_write as in BIO_write.
5370 [Peter.Sylvester@EdelWeb.fr]
5372 *) Fix BN_mul_word for the case where the word is 0. (We have to use
5373 BN_zero, we may not return a BIGNUM with an array consisting of
5377 *) Avoid calling abort() from within the library when problems are
5378 detected, except if preprocessor symbols have been defined
5379 (such as REF_CHECK, BN_DEBUG etc.).
5382 *) New openssl application 'rsautl'. This utility can be
5383 used for low level RSA operations. DER public key
5384 BIO/fp routines also added.
5387 *) New Configure entry and patches for compiling on QNX 4.
5388 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
5390 *) A demo state-machine implementation was sponsored by
5391 Nuron (http://www.nuron.com/) and is now available in
5392 demos/state_machine.
5395 *) New options added to the 'dgst' utility for signature
5396 generation and verification.
5399 *) Unrecognized PKCS#7 content types are now handled via a
5400 catch all ASN1_TYPE structure. This allows unsupported
5401 types to be stored as a "blob" and an application can
5402 encode and decode it manually.
5405 *) Fix various signed/unsigned issues to make a_strex.c
5407 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
5409 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
5410 length if passed a buffer. ASN1_INTEGER_to_BN failed
5411 if passed a NULL BN and its argument was negative.
5412 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
5414 *) Modification to PKCS#7 encoding routines to output definite
5415 length encoding. Since currently the whole structures are in
5416 memory there's not real point in using indefinite length
5417 constructed encoding. However if OpenSSL is compiled with
5418 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
5421 *) Added BIO_vprintf() and BIO_vsnprintf().
5424 *) Added more prefixes to parse for in the the strings written
5425 through a logging bio, to cover all the levels that are available
5426 through syslog. The prefixes are now:
5428 PANIC, EMERG, EMR => LOG_EMERG
5429 ALERT, ALR => LOG_ALERT
5430 CRIT, CRI => LOG_CRIT
5431 ERROR, ERR => LOG_ERR
5432 WARNING, WARN, WAR => LOG_WARNING
5433 NOTICE, NOTE, NOT => LOG_NOTICE
5434 INFO, INF => LOG_INFO
5435 DEBUG, DBG => LOG_DEBUG
5437 and as before, if none of those prefixes are present at the
5438 beginning of the string, LOG_ERR is chosen.
5440 On Win32, the LOG_* levels are mapped according to this:
5442 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
5443 LOG_WARNING => EVENTLOG_WARNING_TYPE
5444 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
5448 *) Made it possible to reconfigure with just the configuration
5449 argument "reconf" or "reconfigure". The command line arguments
5450 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
5451 and are retrieved from there when reconfiguring.
5455 [Assar Westerlund <assar@sics.se>, Richard Levitte]
5457 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
5460 *) The obj_dat.pl script was messing up the sorting of object
5461 names. The reason was that it compared the quoted version
5462 of strings as a result "OCSP" > "OCSP Signing" because
5463 " > SPACE. Changed script to store unquoted versions of
5464 names and add quotes on output. It was also omitting some
5465 names from the lookup table if they were given a default
5466 value (that is if SN is missing it is given the same
5467 value as LN and vice versa), these are now added on the
5468 grounds that if an object has a name we should be able to
5469 look it up. Finally added warning output when duplicate
5470 short or long names are found.
5473 *) Changes needed for Tandem NSK.
5474 [Scott Uroff <scott@xypro.com>]
5476 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
5477 RSA_padding_check_SSLv23(), special padding was never detected
5478 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
5479 version rollback attacks was not effective.
5481 In s23_clnt.c, don't use special rollback-attack detection padding
5482 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
5483 client; similarly, in s23_srvr.c, don't do the rollback check if
5484 SSL 2.0 is the only protocol enabled in the server.
5487 *) Make it possible to get hexdumps of unprintable data with 'openssl
5488 asn1parse'. By implication, the functions ASN1_parse_dump() and
5489 BIO_dump_indent() are added.
5492 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
5493 these print out strings and name structures based on various
5494 flags including RFC2253 support and proper handling of
5495 multibyte characters. Added options to the 'x509' utility
5496 to allow the various flags to be set.
5499 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
5500 Also change the functions X509_cmp_current_time() and
5501 X509_gmtime_adj() work with an ASN1_TIME structure,
5502 this will enable certificates using GeneralizedTime in validity
5503 dates to be checked.
5506 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
5507 negative public key encodings) on by default,
5508 NO_NEG_PUBKEY_BUG can be set to disable it.
5511 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
5512 content octets. An i2c_ASN1_OBJECT is unnecessary because
5513 the encoding can be trivially obtained from the structure.
5516 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
5517 not read locks (CRYPTO_r_[un]lock).
5520 *) A first attempt at creating official support for shared
5521 libraries through configuration. I've kept it so the
5522 default is static libraries only, and the OpenSSL programs
5523 are always statically linked for now, but there are
5524 preparations for dynamic linking in place.
5525 This has been tested on Linux and Tru64.
5528 *) Randomness polling function for Win9x, as described in:
5529 Peter Gutmann, Software Generation of Practically Strong
5533 *) Fix so PRNG is seeded in req if using an already existing
5537 *) New options to smime application. -inform and -outform
5538 allow alternative formats for the S/MIME message including
5539 PEM and DER. The -content option allows the content to be
5540 specified separately. This should allow things like Netscape
5541 form signing output easier to verify.
5544 *) Fix the ASN1 encoding of tags using the 'long form'.
5547 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
5548 STRING types. These convert content octets to and from the
5549 underlying type. The actual tag and length octets are
5550 already assumed to have been read in and checked. These
5551 are needed because all other string types have virtually
5552 identical handling apart from the tag. By having versions
5553 of the ASN1 functions that just operate on content octets
5554 IMPLICIT tagging can be handled properly. It also allows
5555 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
5556 and ASN1_INTEGER are identical apart from the tag.
5559 *) Change the handling of OID objects as follows:
5561 - New object identifiers are inserted in objects.txt, following
5562 the syntax given in objects.README.
5563 - objects.pl is used to process obj_mac.num and create a new
5565 - obj_dat.pl is used to create a new obj_dat.h, using the data in
5568 This is currently kind of a hack, and the perl code in objects.pl
5569 isn't very elegant, but it works as I intended. The simplest way
5570 to check that it worked correctly is to look in obj_dat.h and
5571 check the array nid_objs and make sure the objects haven't moved
5572 around (this is important!). Additions are OK, as well as
5573 consistent name changes.
5576 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
5579 *) Addition of the command line parameter '-rand file' to 'openssl req'.
5580 The given file adds to whatever has already been seeded into the
5581 random pool through the RANDFILE configuration file option or
5582 environment variable, or the default random state file.
5585 *) mkstack.pl now sorts each macro group into lexical order.
5586 Previously the output order depended on the order the files
5587 appeared in the directory, resulting in needless rewriting
5591 *) Patches to make OpenSSL compile under Win32 again. Mostly
5592 work arounds for the VC++ problem that it treats func() as
5593 func(void). Also stripped out the parts of mkdef.pl that
5594 added extra typesafe functions: these no longer exist.
5597 *) Reorganisation of the stack code. The macros are now all
5598 collected in safestack.h . Each macro is defined in terms of
5599 a "stack macro" of the form SKM_<name>(type, a, b). The
5600 DEBUG_SAFESTACK is now handled in terms of function casts,
5601 this has the advantage of retaining type safety without the
5602 use of additional functions. If DEBUG_SAFESTACK is not defined
5603 then the non typesafe macros are used instead. Also modified the
5604 mkstack.pl script to handle the new form. Needs testing to see
5605 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
5606 the default if no major problems. Similar behaviour for ASN1_SET_OF
5607 and PKCS12_STACK_OF.
5610 *) When some versions of IIS use the 'NET' form of private key the
5611 key derivation algorithm is different. Normally MD5(password) is
5612 used as a 128 bit RC4 key. In the modified case
5613 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
5614 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
5615 as the old Netscape_RSA functions except they have an additional
5616 'sgckey' parameter which uses the modified algorithm. Also added
5617 an -sgckey command line option to the rsa utility. Thanks to
5618 Adrian Peck <bertie@ncipher.com> for posting details of the modified
5619 algorithm to openssl-dev.
5622 *) The evp_local.h macros were using 'c.##kname' which resulted in
5623 invalid expansion on some systems (SCO 5.0.5 for example).
5624 Corrected to 'c.kname'.
5625 [Phillip Porch <root@theporch.com>]
5627 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
5628 a STACK of email addresses from a certificate or request, these look
5629 in the subject name and the subject alternative name extensions and
5630 omit any duplicate addresses.
5633 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
5634 This makes DSA verification about 2 % faster.
5637 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
5638 (meaning that now 2^5 values will be precomputed, which is only 4 KB
5639 plus overhead for 1024 bit moduli).
5640 This makes exponentiations about 0.5 % faster for 1024 bit
5641 exponents (as measured by "openssl speed rsa2048").
5644 *) Rename memory handling macros to avoid conflicts with other
5646 Malloc => OPENSSL_malloc
5647 Malloc_locked => OPENSSL_malloc_locked
5648 Realloc => OPENSSL_realloc
5649 Free => OPENSSL_free
5652 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
5653 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
5656 *) CygWin32 support.
5657 [John Jarvie <jjarvie@newsguy.com>]
5659 *) The type-safe stack code has been rejigged. It is now only compiled
5660 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
5661 by default all type-specific stack functions are "#define"d back to
5662 standard stack functions. This results in more streamlined output
5663 but retains the type-safety checking possibilities of the original
5667 *) The STACK code has been cleaned up, and certain type declarations
5668 that didn't make a lot of sense have been brought in line. This has
5669 also involved a cleanup of sorts in safestack.h to more correctly
5670 map type-safe stack functions onto their plain stack counterparts.
5671 This work has also resulted in a variety of "const"ifications of
5672 lots of the code, especially "_cmp" operations which should normally
5673 be prototyped with "const" parameters anyway.
5676 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
5677 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
5678 (The PRNG state consists of two parts, the large pool 'state' and 'md',
5679 where all of 'md' is used each time the PRNG is used, but 'state'
5680 is used only indexed by a cyclic counter. As entropy may not be
5681 well distributed from the beginning, 'md' is important as a
5682 chaining variable. However, the output function chains only half
5683 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
5684 all of 'md', and seeding with STATE_SIZE dummy bytes will result
5685 in all of 'state' being rewritten, with the new values depending
5686 on virtually all of 'md'. This overcomes the 80 bit limitation.)
5689 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
5690 the handshake is continued after ssl_verify_cert_chain();
5691 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
5692 can lead to 'unexplainable' connection aborts later.
5693 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
5695 *) Major EVP API cipher revision.
5696 Add hooks for extra EVP features. This allows various cipher
5697 parameters to be set in the EVP interface. Support added for variable
5698 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
5699 setting of RC2 and RC5 parameters.
5701 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
5704 Remove lots of duplicated code from the EVP library. For example *every*
5705 cipher init() function handles the 'iv' in the same way according to the
5706 cipher mode. They also all do nothing if the 'key' parameter is NULL and
5707 for CFB and OFB modes they zero ctx->num.
5709 New functionality allows removal of S/MIME code RC2 hack.
5711 Most of the routines have the same form and so can be declared in terms
5714 By shifting this to the top level EVP_CipherInit() it can be removed from
5715 all individual ciphers. If the cipher wants to handle IVs or keys
5716 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
5719 Change lots of functions like EVP_EncryptUpdate() to now return a
5720 value: although software versions of the algorithms cannot fail
5721 any installed hardware versions can.
5724 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
5725 this option is set, tolerate broken clients that send the negotiated
5726 protocol version number instead of the requested protocol version
5730 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
5731 i.e. non-zero for export ciphersuites, zero otherwise.
5732 Previous versions had this flag inverted, inconsistent with
5733 rsa_tmp_cb (..._TMP_RSA_CB).
5734 [Bodo Moeller; problem reported by Amit Chopra]
5736 *) Add missing DSA library text string. Work around for some IIS
5737 key files with invalid SEQUENCE encoding.
5740 *) Add a document (doc/standards.txt) that list all kinds of standards
5741 and so on that are implemented in OpenSSL.
5744 *) Enhance c_rehash script. Old version would mishandle certificates
5745 with the same subject name hash and wouldn't handle CRLs at all.
5746 Added -fingerprint option to crl utility, to support new c_rehash
5750 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
5753 *) Fix for SSL server purpose checking. Server checking was
5754 rejecting certificates which had extended key usage present
5755 but no ssl client purpose.
5756 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
5758 *) Make PKCS#12 code work with no password. The PKCS#12 spec
5759 is a little unclear about how a blank password is handled.
5760 Since the password in encoded as a BMPString with terminating
5761 double NULL a zero length password would end up as just the
5762 double NULL. However no password at all is different and is
5763 handled differently in the PKCS#12 key generation code. NS
5764 treats a blank password as zero length. MSIE treats it as no
5765 password on export: but it will try both on import. We now do
5766 the same: PKCS12_parse() tries zero length and no password if
5767 the password is set to "" or NULL (NULL is now a valid password:
5768 it wasn't before) as does the pkcs12 application.
5771 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
5772 perror when PEM_read_bio_X509_REQ fails, the error message must
5773 be obtained from the error queue.
5776 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
5777 it in ERR_remove_state if appropriate, and change ERR_get_state
5778 accordingly to avoid race conditions (this is necessary because
5779 thread_hash is no longer constant once set).
5782 *) Bugfix for linux-elf makefile.one.
5785 *) RSA_get_default_method() will now cause a default
5786 RSA_METHOD to be chosen if one doesn't exist already.
5787 Previously this was only set during a call to RSA_new()
5788 or RSA_new_method(NULL) meaning it was possible for
5789 RSA_get_default_method() to return NULL.
5792 *) Added native name translation to the existing DSO code
5793 that will convert (if the flag to do so is set) filenames
5794 that are sufficiently small and have no path information
5795 into a canonical native form. Eg. "blah" converted to
5796 "libblah.so" or "blah.dll" etc.
5799 *) New function ERR_error_string_n(e, buf, len) which is like
5800 ERR_error_string(e, buf), but writes at most 'len' bytes
5801 including the 0 terminator. For ERR_error_string_n, 'buf'
5803 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
5805 *) CONF library reworked to become more general. A new CONF
5806 configuration file reader "class" is implemented as well as a
5807 new functions (NCONF_*, for "New CONF") to handle it. The now
5808 old CONF_* functions are still there, but are reimplemented to
5809 work in terms of the new functions. Also, a set of functions
5810 to handle the internal storage of the configuration data is
5811 provided to make it easier to write new configuration file
5812 reader "classes" (I can definitely see something reading a
5813 configuration file in XML format, for example), called _CONF_*,
5814 or "the configuration storage API"...
5816 The new configuration file reading functions are:
5818 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
5819 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
5821 NCONF_default, NCONF_WIN32
5823 NCONF_dump_fp, NCONF_dump_bio
5825 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
5826 NCONF_new creates a new CONF object. This works in the same way
5827 as other interfaces in OpenSSL, like the BIO interface.
5828 NCONF_dump_* dump the internal storage of the configuration file,
5829 which is useful for debugging. All other functions take the same
5830 arguments as the old CONF_* functions wth the exception of the
5831 first that must be a `CONF *' instead of a `LHASH *'.
5833 To make it easer to use the new classes with the old CONF_* functions,
5834 the function CONF_set_default_method is provided.
5837 *) Add '-tls1' option to 'openssl ciphers', which was already
5838 mentioned in the documentation but had not been implemented.
5839 (This option is not yet really useful because even the additional
5840 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
5843 *) Initial DSO code added into libcrypto for letting OpenSSL (and
5844 OpenSSL-based applications) load shared libraries and bind to
5845 them in a portable way.
5846 [Geoff Thorpe, with contributions from Richard Levitte]
5848 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
5850 *) Make sure _lrotl and _lrotr are only used with MSVC.
5852 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
5853 (the default implementation of RAND_status).
5855 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
5856 to '-clrext' (= clear extensions), as intended and documented.
5857 [Bodo Moeller; inconsistency pointed out by Michael Attili
5860 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
5861 was larger than the MD block size.
5862 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
5864 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
5865 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
5866 using the passed key: if the passed key was a private key the result
5867 of X509_print(), for example, would be to print out all the private key
5871 *) des_quad_cksum() byte order bug fix.
5872 [Ulf Möller, using the problem description in krb4-0.9.7, where
5873 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
5875 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
5877 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
5879 *) For easily testing in shell scripts whether some command
5880 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
5881 returns with exit code 0 iff no command of the given name is available.
5882 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
5883 the output goes to stdout and nothing is printed to stderr.
5884 Additional arguments are always ignored.
5886 Since for each cipher there is a command of the same name,
5887 the 'no-cipher' compilation switches can be tested this way.
5889 ('openssl no-XXX' is not able to detect pseudo-commands such
5890 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
5893 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
5896 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
5897 is set; it will be thrown away anyway because each handshake creates
5899 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
5900 to parameters -- in previous versions (since OpenSSL 0.9.3) the
5901 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
5902 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
5905 *) New s_client option -ign_eof: EOF at stdin is ignored, and
5906 'Q' and 'R' lose their special meanings (quit/renegotiate).
5907 This is part of what -quiet does; unlike -quiet, -ign_eof
5908 does not suppress any output.
5911 *) Add compatibility options to the purpose and trust code. The
5912 purpose X509_PURPOSE_ANY is "any purpose" which automatically
5913 accepts a certificate or CA, this was the previous behaviour,
5914 with all the associated security issues.
5916 X509_TRUST_COMPAT is the old trust behaviour: only and
5917 automatically trust self signed roots in certificate store. A
5918 new trust setting X509_TRUST_DEFAULT is used to specify that
5919 a purpose has no associated trust setting and it should instead
5920 use the value in the default purpose.
5923 *) Fix the PKCS#8 DSA private key code so it decodes keys again
5924 and fix a memory leak.
5927 *) In util/mkerr.pl (which implements 'make errors'), preserve
5928 reason strings from the previous version of the .c file, as
5929 the default to have only downcase letters (and digits) in
5930 automatically generated reasons codes is not always appropriate.
5933 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
5934 using strerror. Previously, ERR_reason_error_string() returned
5935 library names as reason strings for SYSerr; but SYSerr is a special
5936 case where small numbers are errno values, not library numbers.
5939 *) Add '-dsaparam' option to 'openssl dhparam' application. This
5940 converts DSA parameters into DH parameters. (When creating parameters,
5941 DSA_generate_parameters is used.)
5944 *) Include 'length' (recommended exponent length) in C code generated
5945 by 'openssl dhparam -C'.
5948 *) The second argument to set_label in perlasm was already being used
5949 so couldn't be used as a "file scope" flag. Moved to third argument
5953 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
5954 instead of RAND_bytes for encryption IVs and salts.
5957 *) Include RAND_status() into RAND_METHOD instead of implementing
5958 it only for md_rand.c Otherwise replacing the PRNG by calling
5959 RAND_set_rand_method would be impossible.
5962 *) Don't let DSA_generate_key() enter an infinite loop if the random
5963 number generation fails.
5966 *) New 'rand' application for creating pseudo-random output.
5969 *) Added configuration support for Linux/IA64
5970 [Rolf Haberrecker <rolf@suse.de>]
5972 *) Assembler module support for Mingw32.
5975 *) Shared library support for HPUX (in shlib/).
5976 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
5978 *) Shared library support for Solaris gcc.
5979 [Lutz Behnke <behnke@trustcenter.de>]
5981 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
5983 *) PKCS7_encrypt() was adding text MIME headers twice because they
5984 were added manually and by SMIME_crlf_copy().
5987 *) In bntest.c don't call BN_rand with zero bits argument.
5988 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
5990 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
5991 case was implemented. This caused BN_div_recp() to fail occasionally.
5994 *) Add an optional second argument to the set_label() in the perl
5995 assembly language builder. If this argument exists and is set
5996 to 1 it signals that the assembler should use a symbol whose
5997 scope is the entire file, not just the current function. This
5998 is needed with MASM which uses the format label:: for this scope.
5999 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
6001 *) Change the ASN1 types so they are typedefs by default. Before
6002 almost all types were #define'd to ASN1_STRING which was causing
6003 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
6007 *) Change names of new functions to the new get1/get0 naming
6008 convention: After 'get1', the caller owns a reference count
6009 and has to call ..._free; 'get0' returns a pointer to some
6010 data structure without incrementing reference counters.
6011 (Some of the existing 'get' functions increment a reference
6012 counter, some don't.)
6013 Similarly, 'set1' and 'add1' functions increase reference
6014 counters or duplicate objects.
6017 *) Allow for the possibility of temp RSA key generation failure:
6018 the code used to assume it always worked and crashed on failure.
6021 *) Fix potential buffer overrun problem in BIO_printf().
6022 [Ulf Möller, using public domain code by Patrick Powell; problem
6023 pointed out by David Sacerdote <das33@cornell.edu>]
6025 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
6026 RAND_egd() and RAND_status(). In the command line application,
6027 the EGD socket can be specified like a seed file using RANDFILE
6031 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
6032 Some CAs (e.g. Verisign) distribute certificates in this form.
6035 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
6036 list to exclude them. This means that no special compilation option
6037 is needed to use anonymous DH: it just needs to be included in the
6041 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
6042 EVP_MD_type. The old functionality is available in a new macro called
6043 EVP_MD_md(). Change code that uses it and update docs.
6046 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
6047 where the 'void *' argument is replaced by a function pointer argument.
6048 Previously 'void *' was abused to point to functions, which works on
6049 many platforms, but is not correct. As these functions are usually
6050 called by macros defined in OpenSSL header files, most source code
6051 should work without changes.
6054 *) <openssl/opensslconf.h> (which is created by Configure) now contains
6055 sections with information on -D... compiler switches used for
6056 compiling the library so that applications can see them. To enable
6057 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
6058 must be defined. E.g.,
6059 #define OPENSSL_ALGORITHM_DEFINES
6060 #include <openssl/opensslconf.h>
6061 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
6062 [Richard Levitte, Ulf and Bodo Möller]
6064 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
6068 *) Change the 'other' type in certificate aux info to a STACK_OF
6069 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
6070 the required ASN1 format: arbitrary types determined by an OID.
6073 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
6074 argument to 'req'. This is not because the function is newer or
6075 better than others it just uses the work 'NEW' in the certificate
6076 request header lines. Some software needs this.
6079 *) Reorganise password command line arguments: now passwords can be
6080 obtained from various sources. Delete the PEM_cb function and make
6081 it the default behaviour: i.e. if the callback is NULL and the
6082 usrdata argument is not NULL interpret it as a null terminated pass
6083 phrase. If usrdata and the callback are NULL then the pass phrase
6084 is prompted for as usual.
6087 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
6088 the support is automatically enabled. The resulting binaries will
6089 autodetect the card and use it if present.
6090 [Ben Laurie and Compaq Inc.]
6092 *) Work around for Netscape hang bug. This sends certificate request
6093 and server done in one record. Since this is perfectly legal in the
6094 SSL/TLS protocol it isn't a "bug" option and is on by default. See
6095 the bugs/SSLv3 entry for more info.
6098 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
6101 *) Add -rand argument to smime and pkcs12 applications and read/write
6105 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
6108 *) Add command line password options to the remaining applications.
6111 *) Bug fix for BN_div_recp() for numerators with an even number of
6115 *) More tests in bntest.c, and changed test_bn output.
6118 *) ./config recognizes MacOS X now.
6121 *) Bug fix for BN_div() when the first words of num and divsor are
6122 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
6125 *) Add support for various broken PKCS#8 formats, and command line
6126 options to produce them.
6129 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
6130 get temporary BIGNUMs from a BN_CTX.
6133 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
6137 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
6138 include a #define from the old name to the new. The original intent
6139 was that statically linked binaries could for example just call
6140 SSLeay_add_all_ciphers() to just add ciphers to the table and not
6141 link with digests. This never worked becayse SSLeay_add_all_digests()
6142 and SSLeay_add_all_ciphers() were in the same source file so calling
6143 one would link with the other. They are now in separate source files.
6146 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
6149 *) Use a less unusual form of the Miller-Rabin primality test (it used
6150 a binary algorithm for exponentiation integrated into the Miller-Rabin
6151 loop, our standard modexp algorithms are faster).
6154 *) Support for the EBCDIC character set completed.
6155 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
6157 *) Source code cleanups: use const where appropriate, eliminate casts,
6158 use void * instead of char * in lhash.
6161 *) Bugfix: ssl3_send_server_key_exchange was not restartable
6162 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
6163 this the server could overwrite ephemeral keys that the client
6167 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
6168 using 50 iterations of the Rabin-Miller test.
6170 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
6171 iterations of the Rabin-Miller test as required by the appendix
6172 to FIPS PUB 186[-1]) instead of DSA_is_prime.
6173 As BN_is_prime_fasttest includes trial division, DSA parameter
6174 generation becomes much faster.
6176 This implies a change for the callback functions in DSA_is_prime
6177 and DSA_generate_parameters: The callback function is called once
6178 for each positive witness in the Rabin-Miller test, not just
6179 occasionally in the inner loop; and the parameters to the
6180 callback function now provide an iteration count for the outer
6181 loop rather than for the current invocation of the inner loop.
6182 DSA_generate_parameters additionally can call the callback
6183 function with an 'iteration count' of -1, meaning that a
6184 candidate has passed the trial division test (when q is generated
6185 from an application-provided seed, trial division is skipped).
6188 *) New function BN_is_prime_fasttest that optionally does trial
6189 division before starting the Rabin-Miller test and has
6190 an additional BN_CTX * argument (whereas BN_is_prime always
6191 has to allocate at least one BN_CTX).
6192 'callback(1, -1, cb_arg)' is called when a number has passed the
6193 trial division stage.
6196 *) Fix for bug in CRL encoding. The validity dates weren't being handled
6200 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
6203 *) New function BN_pseudo_rand().
6206 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
6207 bignum version of BN_from_montgomery() with the working code from
6208 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
6212 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
6213 made it impossible to use the same SSL_SESSION data structure in
6214 SSL2 clients in multiple threads.
6217 *) The return value of RAND_load_file() no longer counts bytes obtained
6218 by stat(). RAND_load_file(..., -1) is new and uses the complete file
6219 to seed the PRNG (previously an explicit byte count was required).
6220 [Ulf Möller, Bodo Möller]
6222 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
6223 used (char *) instead of (void *) and had casts all over the place.
6226 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
6229 *) Retain source code compatibility for BN_prime_checks macro:
6230 BN_is_prime(..., BN_prime_checks, ...) now uses
6231 BN_prime_checks_for_size to determine the appropriate number of
6232 Rabin-Miller iterations.
6235 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
6236 DH_CHECK_P_NOT_SAFE_PRIME.
6237 (Check if this is true? OpenPGP calls them "strong".)
6240 *) Merge the functionality of "dh" and "gendh" programs into a new program
6241 "dhparam". The old programs are retained for now but will handle DH keys
6242 (instead of parameters) in future.
6245 *) Make the ciphers, s_server and s_client programs check the return values
6246 when a new cipher list is set.
6249 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
6250 ciphers. Before when the 56bit ciphers were enabled the sorting was
6253 The syntax for the cipher sorting has been extended to support sorting by
6254 cipher-strength (using the strength_bits hard coded in the tables).
6255 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
6257 Fix a bug in the cipher-command parser: when supplying a cipher command
6258 string with an "undefined" symbol (neither command nor alphanumeric
6259 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
6260 an error is flagged.
6262 Due to the strength-sorting extension, the code of the
6263 ssl_create_cipher_list() function was completely rearranged. I hope that
6264 the readability was also increased :-)
6265 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6267 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
6268 for the first serial number and places 2 in the serial number file. This
6269 avoids problems when the root CA is created with serial number zero and
6270 the first user certificate has the same issuer name and serial number
6274 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
6275 the new code. Add documentation for this stuff.
6278 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
6279 X509_*() to X509at_*() on the grounds that they don't handle X509
6280 structures and behave in an analagous way to the X509v3 functions:
6281 they shouldn't be called directly but wrapper functions should be used
6284 So we also now have some wrapper functions that call the X509at functions
6285 when passed certificate requests. (TO DO: similar things can be done with
6286 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
6287 things. Some of these need some d2i or i2d and print functionality
6288 because they handle more complex structures.)
6291 *) Add missing #ifndefs that caused missing symbols when building libssl
6292 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
6293 NO_RSA in ssl/s2*.c.
6294 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
6296 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
6297 has a return value which indicates the quality of the random data
6298 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
6299 error queue. New function RAND_pseudo_bytes() generates output that is
6300 guaranteed to be unique but not unpredictable. RAND_add is like
6301 RAND_seed, but takes an extra argument for an entropy estimate
6302 (RAND_seed always assumes full entropy).
6305 *) Do more iterations of Rabin-Miller probable prime test (specifically,
6306 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
6307 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
6308 in crypto/bn/bn_prime.c for the complete table). This guarantees a
6309 false-positive rate of at most 2^-80 for random input.
6312 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
6315 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
6316 in the 0.9.5 release), this returns the chain
6317 from an X509_CTX structure with a dup of the stack and all
6318 the X509 reference counts upped: so the stack will exist
6319 after X509_CTX_cleanup() has been called. Modify pkcs12.c
6322 Also make SSL_SESSION_print() print out the verify return
6326 *) Add manpage for the pkcs12 command. Also change the default
6327 behaviour so MAC iteration counts are used unless the new
6328 -nomaciter option is used. This improves file security and
6329 only older versions of MSIE (4.0 for example) need it.
6332 *) Honor the no-xxx Configure options when creating .DEF files.
6335 *) Add PKCS#10 attributes to field table: challengePassword,
6336 unstructuredName and unstructuredAddress. These are taken from
6337 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
6338 international characters are used.
6340 More changes to X509_ATTRIBUTE code: allow the setting of types
6341 based on strings. Remove the 'loc' parameter when adding
6342 attributes because these will be a SET OF encoding which is sorted
6346 *) Initial changes to the 'req' utility to allow request generation
6347 automation. This will allow an application to just generate a template
6348 file containing all the field values and have req construct the
6351 Initial support for X509_ATTRIBUTE handling. Stacks of these are
6352 used all over the place including certificate requests and PKCS#7
6353 structures. They are currently handled manually where necessary with
6354 some primitive wrappers for PKCS#7. The new functions behave in a
6355 manner analogous to the X509 extension functions: they allow
6356 attributes to be looked up by NID and added.
6358 Later something similar to the X509V3 code would be desirable to
6359 automatically handle the encoding, decoding and printing of the
6360 more complex types. The string types like challengePassword can
6361 be handled by the string table functions.
6363 Also modified the multi byte string table handling. Now there is
6364 a 'global mask' which masks out certain types. The table itself
6365 can use the flag STABLE_NO_MASK to ignore the mask setting: this
6366 is useful when for example there is only one permissible type
6367 (as in countryName) and using the mask might result in no valid
6371 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
6372 SSL_get_peer_finished to allow applications to obtain the latest
6373 Finished messages sent to the peer or expected from the peer,
6374 respectively. (SSL_get_peer_finished is usually the Finished message
6375 actually received from the peer, otherwise the protocol will be aborted.)
6377 As the Finished message are message digests of the complete handshake
6378 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
6379 be used for external authentication procedures when the authentication
6380 provided by SSL/TLS is not desired or is not enough.
6383 *) Enhanced support for Alpha Linux is added. Now ./config checks if
6384 the host supports BWX extension and if Compaq C is present on the
6385 $PATH. Just exploiting of the BWX extension results in 20-30%
6386 performance kick for some algorithms, e.g. DES and RC4 to mention
6387 a couple. Compaq C in turn generates ~20% faster code for MD5 and
6391 *) Add support for MS "fast SGC". This is arguably a violation of the
6392 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
6393 weak crypto and after checking the certificate is SGC a second one
6394 with strong crypto. MS SGC stops the first handshake after receiving
6395 the server certificate message and sends a second client hello. Since
6396 a server will typically do all the time consuming operations before
6397 expecting any further messages from the client (server key exchange
6398 is the most expensive) there is little difference between the two.
6400 To get OpenSSL to support MS SGC we have to permit a second client
6401 hello message after we have sent server done. In addition we have to
6402 reset the MAC if we do get this second client hello.
6405 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
6406 if a DER encoded private key is RSA or DSA traditional format. Changed
6407 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
6408 format DER encoded private key. Newer code should use PKCS#8 format which
6409 has the key type encoded in the ASN1 structure. Added DER private key
6410 support to pkcs8 application.
6413 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
6414 ciphersuites has been selected (as required by the SSL 3/TLS 1
6415 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
6416 is set, we interpret this as a request to violate the specification
6417 (the worst that can happen is a handshake failure, and 'correct'
6418 behaviour would result in a handshake failure anyway).
6421 *) In SSL_CTX_add_session, take into account that there might be multiple
6422 SSL_SESSION structures with the same session ID (e.g. when two threads
6423 concurrently obtain them from an external cache).
6424 The internal cache can handle only one SSL_SESSION with a given ID,
6425 so if there's a conflict, we now throw out the old one to achieve
6429 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
6430 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
6431 some routines that use cipher OIDs: some ciphers do not have OIDs
6432 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
6436 *) Simplify the trust setting structure and code. Now we just have
6437 two sequences of OIDs for trusted and rejected settings. These will
6438 typically have values the same as the extended key usage extension
6439 and any application specific purposes.
6441 The trust checking code now has a default behaviour: it will just
6442 check for an object with the same NID as the passed id. Functions can
6443 be provided to override either the default behaviour or the behaviour
6444 for a given id. SSL client, server and email already have functions
6445 in place for compatibility: they check the NID and also return "trusted"
6446 if the certificate is self signed.
6449 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
6450 traditional format into an EVP_PKEY structure.
6453 *) Add a password callback function PEM_cb() which either prompts for
6454 a password if usr_data is NULL or otherwise assumes it is a null
6455 terminated password. Allow passwords to be passed on command line
6456 environment or config files in a few more utilities.
6459 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
6460 keys. Add some short names for PKCS#8 PBE algorithms and allow them
6461 to be specified on the command line for the pkcs8 and pkcs12 utilities.
6462 Update documentation.
6465 *) Support for ASN1 "NULL" type. This could be handled before by using
6466 ASN1_TYPE but there wasn't any function that would try to read a NULL
6467 and produce an error if it couldn't. For compatibility we also have
6468 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
6469 don't allocate anything because they don't need to.
6472 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
6474 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
6476 *) Rebuild of the memory allocation routines used by OpenSSL code and
6477 possibly others as well. The purpose is to make an interface that
6478 provide hooks so anyone can build a separate set of allocation and
6479 deallocation routines to be used by OpenSSL, for example memory
6480 pool implementations, or something else, which was previously hard
6481 since Malloc(), Realloc() and Free() were defined as macros having
6482 the values malloc, realloc and free, respectively (except for Win32
6483 compilations). The same is provided for memory debugging code.
6484 OpenSSL already comes with functionality to find memory leaks, but
6485 this gives people a chance to debug other memory problems.
6487 With these changes, a new set of functions and macros have appeared:
6489 CRYPTO_set_mem_debug_functions() [F]
6490 CRYPTO_get_mem_debug_functions() [F]
6491 CRYPTO_dbg_set_options() [F]
6492 CRYPTO_dbg_get_options() [F]
6493 CRYPTO_malloc_debug_init() [M]
6495 The memory debug functions are NULL by default, unless the library
6496 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
6497 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
6498 gives the standard debugging functions that come with OpenSSL) or
6499 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
6500 provided by the library user) must be used. When the standard
6501 debugging functions are used, CRYPTO_dbg_set_options can be used to
6502 request additional information:
6503 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
6504 the CRYPTO_MDEBUG_xxx macro when compiling the library.
6506 Also, things like CRYPTO_set_mem_functions will always give the
6507 expected result (the new set of functions is used for allocation
6508 and deallocation) at all times, regardless of platform and compiler
6511 To finish it up, some functions that were never use in any other
6512 way than through macros have a new API and new semantic:
6515 CRYPTO_dbg_realloc()
6518 All macros of value have retained their old syntax.
6519 [Richard Levitte and Bodo Moeller]
6521 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
6522 ordering of SMIMECapabilities wasn't in "strength order" and there
6523 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
6527 *) Some ASN1 types with illegal zero length encoding (INTEGER,
6528 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
6529 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
6531 *) Merge in my S/MIME library for OpenSSL. This provides a simple
6532 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
6533 functionality to handle multipart/signed properly) and a utility
6534 called 'smime' to call all this stuff. This is based on code I
6535 originally wrote for Celo who have kindly allowed it to be
6536 included in OpenSSL.
6539 *) Add variants des_set_key_checked and des_set_key_unchecked of
6540 des_set_key (aka des_key_sched). Global variable des_check_key
6541 decides which of these is called by des_set_key; this way
6542 des_check_key behaves as it always did, but applications and
6543 the library itself, which was buggy for des_check_key == 1,
6544 have a cleaner way to pick the version they need.
6547 *) New function PKCS12_newpass() which changes the password of a
6551 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
6552 dynamic mix. In both cases the ids can be used as an index into the
6553 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
6554 functions so they accept a list of the field values and the
6555 application doesn't need to directly manipulate the X509_TRUST
6559 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
6563 *) Modify the way the V3 extension code looks up extensions. This now
6564 works in a similar way to the object code: we have some "standard"
6565 extensions in a static table which is searched with OBJ_bsearch()
6566 and the application can add dynamic ones if needed. The file
6567 crypto/x509v3/ext_dat.h now has the info: this file needs to be
6568 updated whenever a new extension is added to the core code and kept
6569 in ext_nid order. There is a simple program 'tabtest.c' which checks
6570 this. New extensions are not added too often so this file can readily
6571 be maintained manually.
6573 There are two big advantages in doing things this way. The extensions
6574 can be looked up immediately and no longer need to be "added" using
6575 X509V3_add_standard_extensions(): this function now does nothing.
6576 [Side note: I get *lots* of email saying the extension code doesn't
6577 work because people forget to call this function]
6578 Also no dynamic allocation is done unless new extensions are added:
6579 so if we don't add custom extensions there is no need to call
6580 X509V3_EXT_cleanup().
6583 *) Modify enc utility's salting as follows: make salting the default. Add a
6584 magic header, so unsalted files fail gracefully instead of just decrypting
6585 to garbage. This is because not salting is a big security hole, so people
6586 should be discouraged from doing it.
6589 *) Fixes and enhancements to the 'x509' utility. It allowed a message
6590 digest to be passed on the command line but it only used this
6591 parameter when signing a certificate. Modified so all relevant
6592 operations are affected by the digest parameter including the
6593 -fingerprint and -x509toreq options. Also -x509toreq choked if a
6594 DSA key was used because it didn't fix the digest.
6597 *) Initial certificate chain verify code. Currently tests the untrusted
6598 certificates for consistency with the verify purpose (which is set
6599 when the X509_STORE_CTX structure is set up) and checks the pathlength.
6601 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
6602 this is because it will reject chains with invalid extensions whereas
6603 every previous version of OpenSSL and SSLeay made no checks at all.
6605 Trust code: checks the root CA for the relevant trust settings. Trust
6606 settings have an initial value consistent with the verify purpose: e.g.
6607 if the verify purpose is for SSL client use it expects the CA to be
6608 trusted for SSL client use. However the default value can be changed to
6609 permit custom trust settings: one example of this would be to only trust
6610 certificates from a specific "secure" set of CAs.
6612 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
6613 which should be used for version portability: especially since the
6614 verify structure is likely to change more often now.
6616 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
6617 to set them. If not set then assume SSL clients will verify SSL servers
6620 Two new options to the verify program: -untrusted allows a set of
6621 untrusted certificates to be passed in and -purpose which sets the
6622 intended purpose of the certificate. If a purpose is set then the
6623 new chain verify code is used to check extension consistency.
6626 *) Support for the authority information access extension.
6629 *) Modify RSA and DSA PEM read routines to transparently handle
6630 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
6631 public keys in a format compatible with certificate
6632 SubjectPublicKeyInfo structures. Unfortunately there were already
6633 functions called *_PublicKey_* which used various odd formats so
6634 these are retained for compatibility: however the DSA variants were
6635 never in a public release so they have been deleted. Changed dsa/rsa
6636 utilities to handle the new format: note no releases ever handled public
6637 keys so we should be OK.
6639 The primary motivation for this change is to avoid the same fiasco
6640 that dogs private keys: there are several incompatible private key
6641 formats some of which are standard and some OpenSSL specific and
6642 require various evil hacks to allow partial transparent handling and
6643 even then it doesn't work with DER formats. Given the option anything
6644 other than PKCS#8 should be dumped: but the other formats have to
6645 stay in the name of compatibility.
6647 With public keys and the benefit of hindsight one standard format
6648 is used which works with EVP_PKEY, RSA or DSA structures: though
6649 it clearly returns an error if you try to read the wrong kind of key.
6651 Added a -pubkey option to the 'x509' utility to output the public key.
6652 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
6653 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
6654 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
6655 that do the same as the EVP_PKEY_assign_*() except they up the
6656 reference count of the added key (they don't "swallow" the
6660 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
6661 CRLs would fail if the file contained no certificates or no CRLs:
6662 added a new function to read in both types and return the number
6663 read: this means that if none are read it will be an error. The
6664 DER versions of the certificate and CRL reader would always fail
6665 because it isn't possible to mix certificates and CRLs in DER format
6666 without choking one or the other routine. Changed this to just read
6667 a certificate: this is the best we can do. Also modified the code
6668 in apps/verify.c to take notice of return codes: it was previously
6669 attempting to read in certificates from NULL pointers and ignoring
6670 any errors: this is one reason why the cert and CRL reader seemed
6671 to work. It doesn't check return codes from the default certificate
6672 routines: these may well fail if the certificates aren't installed.
6675 *) Code to support otherName option in GeneralName.
6678 *) First update to verify code. Change the verify utility
6679 so it warns if it is passed a self signed certificate:
6680 for consistency with the normal behaviour. X509_verify
6681 has been modified to it will now verify a self signed
6682 certificate if *exactly* the same certificate appears
6683 in the store: it was previously impossible to trust a
6684 single self signed certificate. This means that:
6685 openssl verify ss.pem
6686 now gives a warning about a self signed certificate but
6687 openssl verify -CAfile ss.pem ss.pem
6691 *) For servers, store verify_result in SSL_SESSION data structure
6692 (and add it to external session representation).
6693 This is needed when client certificate verifications fails,
6694 but an application-provided verification callback (set by
6695 SSL_CTX_set_cert_verify_callback) allows accepting the session
6696 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
6697 but returns 1): When the session is reused, we have to set
6698 ssl->verify_result to the appropriate error code to avoid
6700 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
6702 *) Fix a bug in the new PKCS#7 code: it didn't consider the
6703 case in PKCS7_dataInit() where the signed PKCS7 structure
6704 didn't contain any existing data because it was being created.
6705 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
6707 *) Add a salt to the key derivation routines in enc.c. This
6708 forms the first 8 bytes of the encrypted file. Also add a
6709 -S option to allow a salt to be input on the command line.
6712 *) New function X509_cmp(). Oddly enough there wasn't a function
6713 to compare two certificates. We do this by working out the SHA1
6714 hash and comparing that. X509_cmp() will be needed by the trust
6718 *) SSL_get1_session() is like SSL_get_session(), but increments
6719 the reference count in the SSL_SESSION returned.
6720 [Geoff Thorpe <geoff@eu.c2.net>]
6722 *) Fix for 'req': it was adding a null to request attributes.
6723 Also change the X509_LOOKUP and X509_INFO code to handle
6724 certificate auxiliary information.
6727 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
6731 *) Add the possibility to add extra information to the memory leak
6732 detecting output, to form tracebacks, showing from where each
6733 allocation was originated: CRYPTO_push_info("constant string") adds
6734 the string plus current file name and line number to a per-thread
6735 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
6736 is like calling CYRPTO_pop_info() until the stack is empty.
6737 Also updated memory leak detection code to be multi-thread-safe.
6740 *) Add options -text and -noout to pkcs7 utility and delete the
6741 encryption options which never did anything. Update docs.
6744 *) Add options to some of the utilities to allow the pass phrase
6745 to be included on either the command line (not recommended on
6746 OSes like Unix) or read from the environment. Update the
6747 manpages and fix a few bugs.
6750 *) Add a few manpages for some of the openssl commands.
6753 *) Fix the -revoke option in ca. It was freeing up memory twice,
6754 leaking and not finding already revoked certificates.
6757 *) Extensive changes to support certificate auxiliary information.
6758 This involves the use of X509_CERT_AUX structure and X509_AUX
6759 functions. An X509_AUX function such as PEM_read_X509_AUX()
6760 can still read in a certificate file in the usual way but it
6761 will also read in any additional "auxiliary information". By
6762 doing things this way a fair degree of compatibility can be
6763 retained: existing certificates can have this information added
6764 using the new 'x509' options.
6766 Current auxiliary information includes an "alias" and some trust
6767 settings. The trust settings will ultimately be used in enhanced
6768 certificate chain verification routines: currently a certificate
6769 can only be trusted if it is self signed and then it is trusted
6773 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
6774 The problem was that one of the replacement routines had not been working
6775 since SSLeay releases. For now the offending routine has been replaced
6776 with non-optimised assembler. Even so, this now gives around 95%
6777 performance improvement for 1024 bit RSA signs.
6780 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
6781 handling. Most clients have the effective key size in bits equal to
6782 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
6783 A few however don't do this and instead use the size of the decrypted key
6784 to determine the RC2 key length and the AlgorithmIdentifier to determine
6785 the effective key length. In this case the effective key length can still
6786 be 40 bits but the key length can be 168 bits for example. This is fixed
6787 by manually forcing an RC2 key into the EVP_PKEY structure because the
6788 EVP code can't currently handle unusual RC2 key sizes: it always assumes
6789 the key length and effective key length are equal.
6792 *) Add a bunch of functions that should simplify the creation of
6793 X509_NAME structures. Now you should be able to do:
6794 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
6795 and have it automatically work out the correct field type and fill in
6796 the structures. The more adventurous can try:
6797 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
6798 and it will (hopefully) work out the correct multibyte encoding.
6801 *) Change the 'req' utility to use the new field handling and multibyte
6802 copy routines. Before the DN field creation was handled in an ad hoc
6803 way in req, ca, and x509 which was rather broken and didn't support
6804 BMPStrings or UTF8Strings. Since some software doesn't implement
6805 BMPStrings or UTF8Strings yet, they can be enabled using the config file
6806 using the dirstring_type option. See the new comment in the default
6807 openssl.cnf for more info.
6810 *) Make crypto/rand/md_rand.c more robust:
6811 - Assure unique random numbers after fork().
6812 - Make sure that concurrent threads access the global counter and
6813 md serializably so that we never lose entropy in them
6814 or use exactly the same state in multiple threads.
6815 Access to the large state is not always serializable because
6816 the additional locking could be a performance killer, and
6817 md should be large enough anyway.
6820 *) New file apps/app_rand.c with commonly needed functionality
6821 for handling the random seed file.
6823 Use the random seed file in some applications that previously did not:
6825 dsaparam -genkey (which also ignored its '-rand' option),
6828 x509 (when signing).
6829 Except on systems with /dev/urandom, it is crucial to have a random
6830 seed file at least for key creation, DSA signing, and for DH exchanges;
6831 for RSA signatures we could do without one.
6833 gendh and gendsa (unlike genrsa) used to read only the first byte
6834 of each file listed in the '-rand' option. The function as previously
6835 found in genrsa is now in app_rand.c and is used by all programs
6836 that support '-rand'.
6839 *) In RAND_write_file, use mode 0600 for creating files;
6840 don't just chmod when it may be too late.
6843 *) Report an error from X509_STORE_load_locations
6844 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
6847 *) New function ASN1_mbstring_copy() this copies a string in either
6848 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
6849 into an ASN1_STRING type. A mask of permissible types is passed
6850 and it chooses the "minimal" type to use or an error if not type
6854 *) Add function equivalents to the various macros in asn1.h. The old
6855 macros are retained with an M_ prefix. Code inside the library can
6856 use the M_ macros. External code (including the openssl utility)
6857 should *NOT* in order to be "shared library friendly".
6860 *) Add various functions that can check a certificate's extensions
6861 to see if it usable for various purposes such as SSL client,
6862 server or S/MIME and CAs of these types. This is currently
6863 VERY EXPERIMENTAL but will ultimately be used for certificate chain
6864 verification. Also added a -purpose flag to x509 utility to
6865 print out all the purposes.
6868 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
6872 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
6873 for, obtain and decode and extension and obtain its critical flag.
6874 This allows all the necessary extension code to be handled in a
6875 single function call.
6878 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
6879 platforms. See crypto/rc4/rc4_enc.c for further details.
6882 *) New -noout option to asn1parse. This causes no output to be produced
6883 its main use is when combined with -strparse and -out to extract data
6884 from a file (which may not be in ASN.1 format).
6887 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
6888 when producing the local key id.
6889 [Richard Levitte <levitte@stacken.kth.se>]
6891 *) New option -dhparam in s_server. This allows a DH parameter file to be
6892 stated explicitly. If it is not stated then it tries the first server
6893 certificate file. The previous behaviour hard coded the filename
6897 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
6898 a public key to be input or output. For example:
6899 openssl rsa -in key.pem -pubout -out pubkey.pem
6900 Also added necessary DSA public key functions to handle this.
6903 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
6904 in the message. This was handled by allowing
6905 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
6906 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
6908 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
6909 to the end of the strings whereas this didn't. This would cause problems
6910 if strings read with d2i_ASN1_bytes() were later modified.
6911 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
6913 *) Fix for base64 decode bug. When a base64 bio reads only one line of
6914 data and it contains EOF it will end up returning an error. This is
6915 caused by input 46 bytes long. The cause is due to the way base64
6916 BIOs find the start of base64 encoded data. They do this by trying a
6917 trial decode on each line until they find one that works. When they
6918 do a flag is set and it starts again knowing it can pass all the
6919 data directly through the decoder. Unfortunately it doesn't reset
6920 the context it uses. This means that if EOF is reached an attempt
6921 is made to pass two EOFs through the context and this causes the
6922 resulting error. This can also cause other problems as well. As is
6923 usual with these problems it takes *ages* to find and the fix is
6924 trivial: move one line.
6925 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
6927 *) Ugly workaround to get s_client and s_server working under Windows. The
6928 old code wouldn't work because it needed to select() on sockets and the
6929 tty (for keypresses and to see if data could be written). Win32 only
6930 supports select() on sockets so we select() with a 1s timeout on the
6931 sockets and then see if any characters are waiting to be read, if none
6932 are present then we retry, we also assume we can always write data to
6933 the tty. This isn't nice because the code then blocks until we've
6934 received a complete line of data and it is effectively polling the
6935 keyboard at 1s intervals: however it's quite a bit better than not
6936 working at all :-) A dedicated Windows application might handle this
6937 with an event loop for example.
6940 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
6941 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
6942 will be called when RSA_sign() and RSA_verify() are used. This is useful
6943 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
6944 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
6945 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
6946 This necessitated the support of an extra signature type NID_md5_sha1
6947 for SSL signatures and modifications to the SSL library to use it instead
6948 of calling RSA_public_decrypt() and RSA_private_encrypt().
6951 *) Add new -verify -CAfile and -CApath options to the crl program, these
6952 will lookup a CRL issuers certificate and verify the signature in a
6953 similar way to the verify program. Tidy up the crl program so it
6954 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
6955 less strict. It will now permit CRL extensions even if it is not
6956 a V2 CRL: this will allow it to tolerate some broken CRLs.
6959 *) Initialize all non-automatic variables each time one of the openssl
6960 sub-programs is started (this is necessary as they may be started
6961 multiple times from the "OpenSSL>" prompt).
6962 [Lennart Bang, Bodo Moeller]
6964 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
6965 removing all other RSA functionality (this is what NO_RSA does). This
6966 is so (for example) those in the US can disable those operations covered
6967 by the RSA patent while allowing storage and parsing of RSA keys and RSA
6971 *) Non-copying interface to BIO pairs.
6972 (still largely untested)
6975 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
6976 ASCII string. This was handled independently in various places before.
6979 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
6980 UTF8 strings a character at a time.
6983 *) Use client_version from client hello to select the protocol
6984 (s23_srvr.c) and for RSA client key exchange verification
6985 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
6988 *) Add various utility functions to handle SPKACs, these were previously
6989 handled by poking round in the structure internals. Added new function
6990 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
6991 print, verify and generate SPKACs. Based on an original idea from
6992 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
6995 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
6998 *) Allow the config file extension section to be overwritten on the
6999 command line. Based on an original idea from Massimiliano Pala
7000 <madwolf@comune.modena.it>. The new option is called -extensions
7001 and can be applied to ca, req and x509. Also -reqexts to override
7002 the request extensions in req and -crlexts to override the crl extensions
7006 *) Add new feature to the SPKAC handling in ca. Now you can include
7007 the same field multiple times by preceding it by "XXXX." for example:
7010 this is the same syntax as used in the req config file.
7013 *) Allow certificate extensions to be added to certificate requests. These
7014 are specified in a 'req_extensions' option of the req section of the
7015 config file. They can be printed out with the -text option to req but
7016 are otherwise ignored at present.
7019 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
7020 data read consists of only the final block it would not decrypted because
7021 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
7022 A misplaced 'break' also meant the decrypted final block might not be
7023 copied until the next read.
7026 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
7027 a few extra parameters to the DH structure: these will be useful if
7028 for example we want the value of 'q' or implement X9.42 DH.
7031 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
7032 provides hooks that allow the default DSA functions or functions on a
7033 "per key" basis to be replaced. This allows hardware acceleration and
7034 hardware key storage to be handled without major modification to the
7035 library. Also added low level modexp hooks and CRYPTO_EX structure and
7036 associated functions.
7039 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
7040 as "read only": it can't be written to and the buffer it points to will
7041 not be freed. Reading from a read only BIO is much more efficient than
7042 a normal memory BIO. This was added because there are several times when
7043 an area of memory needs to be read from a BIO. The previous method was
7044 to create a memory BIO and write the data to it, this results in two
7045 copies of the data and an O(n^2) reading algorithm. There is a new
7046 function BIO_new_mem_buf() which creates a read only memory BIO from
7047 an area of memory. Also modified the PKCS#7 routines to use read only
7051 *) Bugfix: ssl23_get_client_hello did not work properly when called in
7052 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
7053 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
7054 but a retry condition occured while trying to read the rest.
7057 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
7058 NID_pkcs7_encrypted by default: this was wrong since this should almost
7059 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
7060 the encrypted data type: this is a more sensible place to put it and it
7061 allows the PKCS#12 code to be tidied up that duplicated this
7065 *) Changed obj_dat.pl script so it takes its input and output files on
7066 the command line. This should avoid shell escape redirection problems
7070 *) Initial support for certificate extension requests, these are included
7071 in things like Xenroll certificate requests. Included functions to allow
7072 extensions to be obtained and added.
7075 *) -crlf option to s_client and s_server for sending newlines as
7076 CRLF (as required by many protocols).
7079 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
7081 *) Install libRSAglue.a when OpenSSL is built with RSAref.
7082 [Ralf S. Engelschall]
7084 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
7085 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
7087 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
7091 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
7092 DH parameters/keys (q is lost during that conversion, but the resulting
7093 DH parameters contain its length).
7095 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
7096 much faster than DH_generate_parameters (which creates parameters
7097 where p = 2*q + 1), and also the smaller q makes DH computations
7098 much more efficient (160-bit exponentiation instead of 1024-bit
7099 exponentiation); so this provides a convenient way to support DHE
7100 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
7101 utter importance to use
7102 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7104 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7105 when such DH parameters are used, because otherwise small subgroup
7106 attacks may become possible!
7109 *) Avoid memory leak in i2d_DHparams.
7112 *) Allow the -k option to be used more than once in the enc program:
7113 this allows the same encrypted message to be read by multiple recipients.
7116 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
7117 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
7118 it will always use the numerical form of the OID, even if it has a short
7122 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
7123 method only got called if p,q,dmp1,dmq1,iqmp components were present,
7124 otherwise bn_mod_exp was called. In the case of hardware keys for example
7125 no private key components need be present and it might store extra data
7126 in the RSA structure, which cannot be accessed from bn_mod_exp.
7127 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
7128 private key operations.
7131 *) Added support for SPARC Linux.
7134 *) pem_password_cb function type incompatibly changed from
7135 typedef int pem_password_cb(char *buf, int size, int rwflag);
7137 ....(char *buf, int size, int rwflag, void *userdata);
7138 so that applications can pass data to their callbacks:
7139 The PEM[_ASN1]_{read,write}... functions and macros now take an
7140 additional void * argument, which is just handed through whenever
7141 the password callback is called.
7142 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
7144 New function SSL_CTX_set_default_passwd_cb_userdata.
7146 Compatibility note: As many C implementations push function arguments
7147 onto the stack in reverse order, the new library version is likely to
7148 interoperate with programs that have been compiled with the old
7149 pem_password_cb definition (PEM_whatever takes some data that
7150 happens to be on the stack as its last argument, and the callback
7151 just ignores this garbage); but there is no guarantee whatsoever that
7154 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
7155 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
7156 problems not only on Windows, but also on some Unix platforms.
7157 To avoid problematic command lines, these definitions are now in an
7158 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
7159 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
7162 *) MIPS III/IV assembler module is reimplemented.
7165 *) More DES library cleanups: remove references to srand/rand and
7166 delete an unused file.
7169 *) Add support for the the free Netwide assembler (NASM) under Win32,
7170 since not many people have MASM (ml) and it can be hard to obtain.
7171 This is currently experimental but it seems to work OK and pass all
7172 the tests. Check out INSTALL.W32 for info.
7175 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
7176 without temporary keys kept an extra copy of the server key,
7177 and connections with temporary keys did not free everything in case
7181 *) New function RSA_check_key and new openssl rsa option -check
7182 for verifying the consistency of RSA keys.
7183 [Ulf Moeller, Bodo Moeller]
7185 *) Various changes to make Win32 compile work:
7186 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
7187 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
7188 comparison" warnings.
7189 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
7192 *) Add a debugging option to PKCS#5 v2 key generation function: when
7193 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
7194 derived keys are printed to stderr.
7197 *) Copy the flags in ASN1_STRING_dup().
7198 [Roman E. Pavlov <pre@mo.msk.ru>]
7200 *) The x509 application mishandled signing requests containing DSA
7201 keys when the signing key was also DSA and the parameters didn't match.
7203 It was supposed to omit the parameters when they matched the signing key:
7204 the verifying software was then supposed to automatically use the CA's
7205 parameters if they were absent from the end user certificate.
7207 Omitting parameters is no longer recommended. The test was also
7208 the wrong way round! This was probably due to unusual behaviour in
7209 EVP_cmp_parameters() which returns 1 if the parameters match.
7210 This meant that parameters were omitted when they *didn't* match and
7211 the certificate was useless. Certificates signed with 'ca' didn't have
7213 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
7215 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
7216 The interface is as follows:
7217 Applications can use
7218 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
7219 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
7220 "off" is now the default.
7221 The library internally uses
7222 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
7223 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
7224 to disable memory-checking temporarily.
7226 Some inconsistent states that previously were possible (and were
7227 even the default) are now avoided.
7229 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
7230 with each memory chunk allocated; this is occasionally more helpful
7231 than just having a counter.
7233 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
7235 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
7239 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
7240 which largely parallels "options", but is for changing API behaviour,
7241 whereas "options" are about protocol behaviour.
7242 Initial "mode" flags are:
7244 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
7245 a single record has been written.
7246 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
7247 retries use the same buffer location.
7248 (But all of the contents must be
7252 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
7255 *) Fix problems with no-hmac etc.
7256 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
7258 *) New functions RSA_get_default_method(), RSA_set_method() and
7259 RSA_get_method(). These allows replacement of RSA_METHODs without having
7260 to mess around with the internals of an RSA structure.
7263 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
7264 Also really enable memory leak checks in openssl.c and in some
7266 [Chad C. Mulligan, Bodo Moeller]
7268 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
7269 up the length of negative integers. This has now been simplified to just
7270 store the length when it is first determined and use it later, rather
7271 than trying to keep track of where data is copied and updating it to
7273 [Steve Henson, reported by Brien Wheeler
7274 <bwheeler@authentica-security.com>]
7276 *) Add a new function PKCS7_signatureVerify. This allows the verification
7277 of a PKCS#7 signature but with the signing certificate passed to the
7278 function itself. This contrasts with PKCS7_dataVerify which assumes the
7279 certificate is present in the PKCS#7 structure. This isn't always the
7280 case: certificates can be omitted from a PKCS#7 structure and be
7281 distributed by "out of band" means (such as a certificate database).
7284 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
7285 function prototypes in pem.h, also change util/mkdef.pl to add the
7286 necessary function names.
7289 *) mk1mf.pl (used by Windows builds) did not properly read the
7290 options set by Configure in the top level Makefile, and Configure
7291 was not even able to write more than one option correctly.
7292 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
7295 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
7296 file to be loaded from a BIO or FILE pointer. The BIO version will
7297 for example allow memory BIOs to contain config info.
7300 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
7301 Whoever hopes to achieve shared-library compatibility across versions
7302 must use this, not the compile-time macro.
7303 (Exercise 0.9.4: Which is the minimum library version required by
7305 Note: All this applies only to multi-threaded programs, others don't
7309 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
7310 through a BIO pair triggered the default case, i.e.
7311 SSLerr(...,SSL_R_UNKNOWN_STATE).
7314 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
7315 can use the SSL library even if none of the specific BIOs is
7319 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
7320 for the encoded length.
7321 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
7323 *) Add initial documentation of the X509V3 functions.
7326 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
7327 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
7328 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
7329 secure PKCS#8 private key format with a high iteration count.
7332 *) Fix determination of Perl interpreter: A perl or perl5
7333 _directory_ in $PATH was also accepted as the interpreter.
7334 [Ralf S. Engelschall]
7336 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
7337 wrong with it but it was very old and did things like calling
7338 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
7342 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
7343 to use the new extension code.
7346 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
7347 with macros. This should make it easier to change their form, add extra
7348 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
7352 *) Add to configuration table a new entry that can specify an alternative
7353 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
7354 according to Mark Crispin <MRC@Panda.COM>.
7358 *) DES CBC did not update the IV. Weird.
7361 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
7362 Changing the behaviour of the former might break existing programs --
7363 where IV updating is needed, des_ncbc_encrypt can be used.
7366 *) When bntest is run from "make test" it drives bc to check its
7367 calculations, as well as internally checking them. If an internal check
7368 fails, it needs to cause bc to give a non-zero result or make test carries
7369 on without noticing the failure. Fixed.
7372 *) DES library cleanups.
7375 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
7376 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
7377 ciphers. NOTE: although the key derivation function has been verified
7378 against some published test vectors it has not been extensively tested
7379 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
7383 *) Instead of "mkdir -p", which is not fully portable, use new
7384 Perl script "util/mkdir-p.pl".
7387 *) Rewrite the way password based encryption (PBE) is handled. It used to
7388 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
7389 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
7390 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
7391 the 'parameter' field of the AlgorithmIdentifier is passed to the
7392 underlying key generation function so it must do its own ASN1 parsing.
7393 This has also changed the EVP_PBE_CipherInit() function which now has a
7394 'parameter' argument instead of literal salt and iteration count values
7395 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
7398 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
7399 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
7400 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
7401 KEY" because this clashed with PKCS#8 unencrypted string. Since this
7402 value was just used as a "magic string" and not used directly its
7403 value doesn't matter.
7406 *) Introduce some semblance of const correctness to BN. Shame C doesn't
7410 *) "linux-sparc64" configuration (ultrapenguin).
7411 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
7412 "linux-sparc" configuration.
7413 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
7415 *) config now generates no-xxx options for missing ciphers.
7418 *) Support the EBCDIC character set (work in progress).
7419 File ebcdic.c not yet included because it has a different license.
7420 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7422 *) Support BS2000/OSD-POSIX.
7423 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7425 *) Make callbacks for key generation use void * instead of char *.
7428 *) Make S/MIME samples compile (not yet tested).
7431 *) Additional typesafe stacks.
7434 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
7438 Changes between 0.9.3 and 0.9.3a [29 May 1999]
7440 *) New configuration variant "sco5-gcc".
7442 *) Updated some demos.
7443 [Sean O Riordain, Wade Scholine]
7445 *) Add missing BIO_free at exit of pkcs12 application.
7448 *) Fix memory leak in conf.c.
7451 *) Updates for Win32 to assembler version of MD5.
7454 *) Set #! path to perl in apps/der_chop to where we found it
7455 instead of using a fixed path.
7458 *) SHA library changes for irix64-mips4-cc.
7461 *) Improvements for VMS support.
7465 Changes between 0.9.2b and 0.9.3 [24 May 1999]
7467 *) Bignum library bug fix. IRIX 6 passes "make test" now!
7468 This also avoids the problems with SC4.2 and unpatched SC5.
7469 [Andy Polyakov <appro@fy.chalmers.se>]
7471 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
7472 These are required because of the typesafe stack would otherwise break
7473 existing code. If old code used a structure member which used to be STACK
7474 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
7475 sk_num or sk_value it would produce an error because the num, data members
7476 are not present in STACK_OF. Now it just produces a warning. sk_set
7477 replaces the old method of assigning a value to sk_value
7478 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
7479 that does this will no longer work (and should use sk_set instead) but
7480 this could be regarded as a "questionable" behaviour anyway.
7483 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
7484 correctly handle encrypted S/MIME data.
7487 *) Change type of various DES function arguments from des_cblock
7488 (which means, in function argument declarations, pointer to char)
7489 to des_cblock * (meaning pointer to array with 8 char elements),
7490 which allows the compiler to do more typechecking; it was like
7491 that back in SSLeay, but with lots of ugly casts.
7493 Introduce new type const_des_cblock.
7496 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
7497 problems: find RecipientInfo structure that matches recipient certificate
7498 and initialise the ASN1 structures properly based on passed cipher.
7501 *) Belatedly make the BN tests actually check the results.
7504 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
7505 to and from BNs: it was completely broken. New compilation option
7506 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
7507 key elements as negative integers.
7510 *) Reorganize and speed up MD5.
7511 [Andy Polyakov <appro@fy.chalmers.se>]
7514 [Richard Levitte <richard@levitte.org>]
7516 *) New option -out to asn1parse to allow the parsed structure to be
7517 output to a file. This is most useful when combined with the -strparse
7518 option to examine the output of things like OCTET STRINGS.
7521 *) Make SSL library a little more fool-proof by not requiring any longer
7522 that SSL_set_{accept,connect}_state be called before
7523 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
7524 in many applications because usually everything *appeared* to work as
7525 intended anyway -- now it really works as intended).
7528 *) Move openssl.cnf out of lib/.
7531 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
7532 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
7533 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
7534 [Ralf S. Engelschall]
7536 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
7537 handle PKCS#7 enveloped data properly.
7538 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
7540 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
7541 copying pointers. The cert_st handling is changed by this in
7542 various ways (and thus what used to be known as ctx->default_cert
7543 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
7544 any longer when s->cert does not give us what we need).
7545 ssl_cert_instantiate becomes obsolete by this change.
7546 As soon as we've got the new code right (possibly it already is?),
7547 we have solved a couple of bugs of the earlier code where s->cert
7548 was used as if it could not have been shared with other SSL structures.
7550 Note that using the SSL API in certain dirty ways now will result
7551 in different behaviour than observed with earlier library versions:
7552 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
7553 does not influence s as it used to.
7555 In order to clean up things more thoroughly, inside SSL_SESSION
7556 we don't use CERT any longer, but a new structure SESS_CERT
7557 that holds per-session data (if available); currently, this is
7558 the peer's certificate chain and, for clients, the server's certificate
7559 and temporary key. CERT holds only those values that can have
7560 meaningful defaults in an SSL_CTX.
7563 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
7564 from the internal representation. Various PKCS#7 fixes: remove some
7565 evil casts and set the enc_dig_alg field properly based on the signing
7569 *) Allow PKCS#12 password to be set from the command line or the
7570 environment. Let 'ca' get its config file name from the environment
7571 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
7575 *) Allow certificate policies extension to use an IA5STRING for the
7576 organization field. This is contrary to the PKIX definition but
7577 VeriSign uses it and IE5 only recognises this form. Document 'x509'
7581 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
7582 without disallowing inline assembler and the like for non-pedantic builds.
7585 *) Support Borland C++ builder.
7586 [Janez Jere <jj@void.si>, modified by Ulf Möller]
7591 *) SHA-1 cleanups and performance enhancements.
7592 [Andy Polyakov <appro@fy.chalmers.se>]
7594 *) Sparc v8plus assembler for the bignum library.
7595 [Andy Polyakov <appro@fy.chalmers.se>]
7597 *) Accept any -xxx and +xxx compiler options in Configure.
7600 *) Update HPUX configuration.
7603 *) Add missing sk_<type>_unshift() function to safestack.h
7604 [Ralf S. Engelschall]
7606 *) New function SSL_CTX_use_certificate_chain_file that sets the
7607 "extra_cert"s in addition to the certificate. (This makes sense
7608 only for "PEM" format files, as chains as a whole are not
7612 *) Support verify_depth from the SSL API.
7613 x509_vfy.c had what can be considered an off-by-one-error:
7614 Its depth (which was not part of the external interface)
7615 was actually counting the number of certificates in a chain;
7616 now it really counts the depth.
7619 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
7620 instead of X509err, which often resulted in confusing error
7621 messages since the error codes are not globally unique
7622 (e.g. an alleged error in ssl3_accept when a certificate
7623 didn't match the private key).
7625 *) New function SSL_CTX_set_session_id_context that allows to set a default
7626 value (so that you don't need SSL_set_session_id_context for each
7627 connection using the SSL_CTX).
7630 *) OAEP decoding bug fix.
7633 *) Support INSTALL_PREFIX for package builders, as proposed by
7637 *) New Configure options "threads" and "no-threads". For systems
7638 where the proper compiler options are known (currently Solaris
7639 and Linux), "threads" is the default.
7642 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
7645 *) Install various scripts to $(OPENSSLDIR)/misc, not to
7646 $(INSTALLTOP)/bin -- they shouldn't clutter directories
7647 such as /usr/local/bin.
7650 *) "make linux-shared" to build shared libraries.
7651 [Niels Poppe <niels@netbox.org>]
7653 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
7656 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
7657 extension adding in x509 utility.
7660 *) Remove NOPROTO sections and error code comments.
7663 *) Partial rewrite of the DEF file generator to now parse the ANSI
7667 *) New Configure options --prefix=DIR and --openssldir=DIR.
7670 *) Complete rewrite of the error code script(s). It is all now handled
7671 by one script at the top level which handles error code gathering,
7672 header rewriting and C source file generation. It should be much better
7673 than the old method: it now uses a modified version of Ulf's parser to
7674 read the ANSI prototypes in all header files (thus the old K&R definitions
7675 aren't needed for error creation any more) and do a better job of
7676 translating function codes into names. The old 'ASN1 error code imbedded
7677 in a comment' is no longer necessary and it doesn't use .err files which
7678 have now been deleted. Also the error code call doesn't have to appear all
7679 on one line (which resulted in some large lines...).
7682 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
7685 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
7686 0 (which usually indicates a closed connection), but continue reading.
7689 *) Fix some race conditions.
7692 *) Add support for CRL distribution points extension. Add Certificate
7693 Policies and CRL distribution points documentation.
7696 *) Move the autogenerated header file parts to crypto/opensslconf.h.
7699 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
7700 8 of keying material. Merlin has also confirmed interop with this fix
7701 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
7702 [Merlin Hughes <merlin@baltimore.ie>]
7704 *) Fix lots of warnings.
7705 [Richard Levitte <levitte@stacken.kth.se>]
7707 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
7708 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
7709 [Richard Levitte <levitte@stacken.kth.se>]
7711 *) Fix problems with sizeof(long) == 8.
7712 [Andy Polyakov <appro@fy.chalmers.se>]
7714 *) Change functions to ANSI C.
7717 *) Fix typos in error codes.
7718 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
7720 *) Remove defunct assembler files from Configure.
7723 *) SPARC v8 assembler BIGNUM implementation.
7724 [Andy Polyakov <appro@fy.chalmers.se>]
7726 *) Support for Certificate Policies extension: both print and set.
7727 Various additions to support the r2i method this uses.
7730 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
7731 return a const string when you are expecting an allocated buffer.
7734 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
7735 types DirectoryString and DisplayText.
7738 *) Add code to allow r2i extensions to access the configuration database,
7739 add an LHASH database driver and add several ctx helper functions.
7742 *) Fix an evil bug in bn_expand2() which caused various BN functions to
7743 fail when they extended the size of a BIGNUM.
7746 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
7747 support typesafe stack.
7750 *) Fix typo in SSL_[gs]et_options().
7751 [Nils Frostberg <nils@medcom.se>]
7753 *) Delete various functions and files that belonged to the (now obsolete)
7754 old X509V3 handling code.
7757 *) New Configure option "rsaref".
7760 *) Don't auto-generate pem.h.
7763 *) Introduce type-safe ASN.1 SETs.
7766 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
7767 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
7769 *) Introduce type-safe STACKs. This will almost certainly break lots of code
7770 that links with OpenSSL (well at least cause lots of warnings), but fear
7771 not: the conversion is trivial, and it eliminates loads of evil casts. A
7772 few STACKed things have been converted already. Feel free to convert more.
7773 In the fullness of time, I'll do away with the STACK type altogether.
7776 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
7777 specified in <certfile> by updating the entry in the index.txt file.
7778 This way one no longer has to edit the index.txt file manually for
7779 revoking a certificate. The -revoke option does the gory details now.
7780 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
7782 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
7783 `-text' option at all and this way the `-noout -text' combination was
7784 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
7785 [Ralf S. Engelschall]
7787 *) Make sure a corresponding plain text error message exists for the
7788 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
7789 verify callback function determined that a certificate was revoked.
7790 [Ralf S. Engelschall]
7792 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
7793 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
7794 all available cipers including rc5, which was forgotten until now.
7795 In order to let the testing shell script know which algorithms
7796 are available, a new (up to now undocumented) command
7797 "openssl list-cipher-commands" is used.
7800 *) Bugfix: s_client occasionally would sleep in select() when
7801 it should have checked SSL_pending() first.
7804 *) New functions DSA_do_sign and DSA_do_verify to provide access to
7805 the raw DSA values prior to ASN.1 encoding.
7808 *) Tweaks to Configure
7809 [Niels Poppe <niels@netbox.org>]
7811 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
7815 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
7818 *) New config option to avoid instructions that are illegal on the 80386.
7819 The default code is faster, but requires at least a 486.
7822 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
7823 SSL2_SERVER_VERSION (not used at all) macros, which are now the
7824 same as SSL2_VERSION anyway.
7827 *) New "-showcerts" option for s_client.
7830 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
7831 application. Various cleanups and fixes.
7834 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
7835 modify error routines to work internally. Add error codes and PBE init
7836 to library startup routines.
7839 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
7840 packing functions to asn1 and evp. Changed function names and error
7841 codes along the way.
7844 *) PKCS12 integration: and so it begins... First of several patches to
7845 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
7846 objects to objects.h
7849 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
7850 and display support for Thawte strong extranet extension.
7853 *) Add LinuxPPC support.
7854 [Jeff Dubrule <igor@pobox.org>]
7856 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
7857 bn_div_words in alpha.s.
7858 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
7860 *) Make sure the RSA OAEP test is skipped under -DRSAref because
7861 OAEP isn't supported when OpenSSL is built with RSAref.
7862 [Ulf Moeller <ulf@fitug.de>]
7864 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
7865 so they no longer are missing under -DNOPROTO.
7866 [Soren S. Jorvang <soren@t.dk>]
7869 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
7871 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
7872 doesn't work when the session is reused. Coming soon!
7875 *) Fix a security hole, that allows sessions to be reused in the wrong
7876 context thus bypassing client cert protection! All software that uses
7877 client certs and session caches in multiple contexts NEEDS PATCHING to
7878 allow session reuse! A fuller solution is in the works.
7879 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
7881 *) Some more source tree cleanups (removed obsolete files
7882 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
7883 permission on "config" script to be executable) and a fix for the INSTALL
7885 [Ulf Moeller <ulf@fitug.de>]
7887 *) Remove some legacy and erroneous uses of malloc, free instead of
7889 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
7891 *) Make rsa_oaep_test return non-zero on error.
7892 [Ulf Moeller <ulf@fitug.de>]
7894 *) Add support for native Solaris shared libraries. Configure
7895 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
7896 if someone would make that last step automatic.
7897 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
7899 *) ctx_size was not built with the right compiler during "make links". Fixed.
7902 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
7903 except NULL ciphers". This means the default cipher list will no longer
7904 enable NULL ciphers. They need to be specifically enabled e.g. with
7905 the string "DEFAULT:eNULL".
7908 *) Fix to RSA private encryption routines: if p < q then it would
7909 occasionally produce an invalid result. This will only happen with
7910 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
7913 *) Be less restrictive and allow also `perl util/perlpath.pl
7914 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
7915 because this way one can also use an interpreter named `perl5' (which is
7916 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
7917 installed as `perl').
7918 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7920 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
7921 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7923 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
7924 advapi32.lib to Win32 build and change the pem test comparision
7925 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
7926 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
7927 and crypto/des/ede_cbcm_enc.c.
7930 *) DES quad checksum was broken on big-endian architectures. Fixed.
7933 *) Comment out two functions in bio.h that aren't implemented. Fix up the
7934 Win32 test batch file so it (might) work again. The Win32 test batch file
7935 is horrible: I feel ill....
7938 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
7939 in e_os.h. Audit of header files to check ANSI and non ANSI
7940 sections: 10 functions were absent from non ANSI section and not exported
7941 from Windows DLLs. Fixed up libeay.num for new functions.
7944 *) Make `openssl version' output lines consistent.
7945 [Ralf S. Engelschall]
7947 *) Fix Win32 symbol export lists for BIO functions: Added
7948 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
7949 to ms/libeay{16,32}.def.
7950 [Ralf S. Engelschall]
7952 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
7953 fine under Unix and passes some trivial tests I've now added. But the
7954 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
7955 added to make sure no one expects that this stuff really works in the
7956 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
7957 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
7959 [Ralf S. Engelschall]
7961 *) Fix the generation of two part addresses in perl.
7962 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
7964 *) Add config entry for Linux on MIPS.
7965 [John Tobey <jtobey@channel1.com>]
7967 *) Make links whenever Configure is run, unless we are on Windoze.
7970 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
7971 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
7975 *) Add a useful kludge to allow package maintainers to specify compiler and
7976 other platforms details on the command line without having to patch the
7977 Configure script everytime: One now can use ``perl Configure
7978 <id>:<details>'', i.e. platform ids are allowed to have details appended
7979 to them (seperated by colons). This is treated as there would be a static
7980 pre-configured entry in Configure's %table under key <id> with value
7981 <details> and ``perl Configure <id>'' is called. So, when you want to
7982 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
7983 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
7984 now, which overrides the FreeBSD-elf entry on-the-fly.
7985 [Ralf S. Engelschall]
7987 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
7990 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
7991 on the `perl Configure ...' command line. This way one can compile
7992 OpenSSL libraries with Position Independent Code (PIC) which is needed
7993 for linking it into DSOs.
7994 [Ralf S. Engelschall]
7996 *) Remarkably, export ciphers were totally broken and no-one had noticed!
8000 *) Cleaned up the LICENSE document: The official contact for any license
8001 questions now is the OpenSSL core team under openssl-core@openssl.org.
8002 And add a paragraph about the dual-license situation to make sure people
8003 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
8004 to the OpenSSL toolkit.
8005 [Ralf S. Engelschall]
8007 *) General source tree makefile cleanups: Made `making xxx in yyy...'
8008 display consistent in the source tree and replaced `/bin/rm' by `rm'.
8009 Additonally cleaned up the `make links' target: Remove unnecessary
8010 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
8011 to speed processing and no longer clutter the display with confusing
8012 stuff. Instead only the actually done links are displayed.
8013 [Ralf S. Engelschall]
8015 *) Permit null encryption ciphersuites, used for authentication only. It used
8016 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
8017 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
8021 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
8022 signed attributes when verifying signatures (this would break them),
8023 the detached data encoding was wrong and public keys obtained using
8024 X509_get_pubkey() weren't freed.
8027 *) Add text documentation for the BUFFER functions. Also added a work around
8028 to a Win95 console bug. This was triggered by the password read stuff: the
8029 last character typed gets carried over to the next fread(). If you were
8030 generating a new cert request using 'req' for example then the last
8031 character of the passphrase would be CR which would then enter the first
8035 *) Added the new `Includes OpenSSL Cryptography Software' button as
8036 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
8037 button and can be used by applications based on OpenSSL to show the
8038 relationship to the OpenSSL project.
8039 [Ralf S. Engelschall]
8041 *) Remove confusing variables in function signatures in files
8042 ssl/ssl_lib.c and ssl/ssl.h.
8043 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8045 *) Don't install bss_file.c under PREFIX/include/
8046 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8048 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
8049 functions that return function pointers and has support for NT specific
8050 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
8051 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
8052 unsigned to signed types: this was killing the Win32 compile.
8055 *) Add new certificate file to stack functions,
8056 SSL_add_dir_cert_subjects_to_stack() and
8057 SSL_add_file_cert_subjects_to_stack(). These largely supplant
8058 SSL_load_client_CA_file(), and can be used to add multiple certs easily
8059 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
8060 This means that Apache-SSL and similar packages don't have to mess around
8061 to add as many CAs as they want to the preferred list.
8064 *) Experiment with doxygen documentation. Currently only partially applied to
8066 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
8067 openssl.doxy as the configuration file.
8070 *) Get rid of remaining C++-style comments which strict C compilers hate.
8071 [Ralf S. Engelschall, pointed out by Carlos Amengual]
8073 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
8074 compiled in by default: it has problems with large keys.
8077 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
8078 DH private keys and/or callback functions which directly correspond to
8079 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
8080 is needed for applications which have to configure certificates on a
8081 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
8083 For the RSA certificate situation is makes no difference, but
8084 for the DSA certificate situation this fixes the "no shared cipher"
8085 problem where the OpenSSL cipher selection procedure failed because the
8086 temporary keys were not overtaken from the context and the API provided
8087 no way to reconfigure them.
8088 The new functions now let applications reconfigure the stuff and they
8089 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
8090 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
8091 non-public-API function ssl_cert_instantiate() is used as a helper
8092 function and also to reduce code redundancy inside ssl_rsa.c.
8093 [Ralf S. Engelschall]
8095 *) Move s_server -dcert and -dkey options out of the undocumented feature
8096 area because they are useful for the DSA situation and should be
8097 recognized by the users.
8098 [Ralf S. Engelschall]
8100 *) Fix the cipher decision scheme for export ciphers: the export bits are
8101 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
8102 SSL_EXP_MASK. So, the original variable has to be used instead of the
8103 already masked variable.
8104 [Richard Levitte <levitte@stacken.kth.se>]
8106 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
8107 [Richard Levitte <levitte@stacken.kth.se>]
8109 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
8110 from `int' to `unsigned int' because it's a length and initialized by
8111 EVP_DigestFinal() which expects an `unsigned int *'.
8112 [Richard Levitte <levitte@stacken.kth.se>]
8114 *) Don't hard-code path to Perl interpreter on shebang line of Configure
8115 script. Instead use the usual Shell->Perl transition trick.
8116 [Ralf S. Engelschall]
8118 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
8119 (in addition to RSA certificates) to match the behaviour of `openssl dsa
8120 -noout -modulus' as it's already the case for `openssl rsa -noout
8121 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
8122 currently the public key is printed (a decision which was already done by
8123 `openssl dsa -modulus' in the past) which serves a similar purpose.
8124 Additionally the NO_RSA no longer completely removes the whole -modulus
8125 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
8127 [Ralf S. Engelschall]
8129 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
8130 BIO. See the source (crypto/evp/bio_ok.c) for more info.
8131 [Arne Ansper <arne@ats.cyber.ee>]
8133 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
8134 to be added. Now both 'req' and 'ca' can use new objects defined in the
8138 *) Add cool BIO that does syslog (or event log on NT).
8139 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
8141 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
8142 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
8143 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
8144 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
8147 *) Add preliminary config info for new extension code.
8150 *) Make RSA_NO_PADDING really use no padding.
8151 [Ulf Moeller <ulf@fitug.de>]
8153 *) Generate errors when private/public key check is done.
8156 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
8157 for some CRL extensions and new objects added.
8160 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
8161 key usage extension and fuller support for authority key id.
8164 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
8165 padding method for RSA, which is recommended for new applications in PKCS
8166 #1 v2.0 (RFC 2437, October 1998).
8167 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
8168 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
8169 against Bleichbacher's attack on RSA.
8170 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
8173 *) Updates to the new SSL compression code
8174 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8176 *) Fix so that the version number in the master secret, when passed
8177 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
8178 (because the server will not accept higher), that the version number
8179 is 0x03,0x01, not 0x03,0x00
8180 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8182 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
8183 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
8184 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
8187 *) Support for RAW extensions where an arbitrary extension can be
8188 created by including its DER encoding. See apps/openssl.cnf for
8192 *) Make sure latest Perl versions don't interpret some generated C array
8193 code as Perl array code in the crypto/err/err_genc.pl script.
8194 [Lars Weber <3weber@informatik.uni-hamburg.de>]
8196 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
8197 not many people have the assembler. Various Win32 compilation fixes and
8198 update to the INSTALL.W32 file with (hopefully) more accurate Win32
8202 *) Modify configure script 'Configure' to automatically create crypto/date.h
8203 file under Win32 and also build pem.h from pem.org. New script
8204 util/mkfiles.pl to create the MINFO file on environments that can't do a
8205 'make files': perl util/mkfiles.pl >MINFO should work.
8208 *) Major rework of DES function declarations, in the pursuit of correctness
8209 and purity. As a result, many evil casts evaporated, and some weirdness,
8210 too. You may find this causes warnings in your code. Zapping your evil
8211 casts will probably fix them. Mostly.
8214 *) Fix for a typo in asn1.h. Bug fix to object creation script
8215 obj_dat.pl. It considered a zero in an object definition to mean
8216 "end of object": none of the objects in objects.h have any zeros
8217 so it wasn't spotted.
8218 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
8220 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
8221 Masking (CBCM). In the absence of test vectors, the best I have been able
8222 to do is check that the decrypt undoes the encrypt, so far. Send me test
8223 vectors if you have them.
8226 *) Correct calculation of key length for export ciphers (too much space was
8227 allocated for null ciphers). This has not been tested!
8230 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
8231 message is now correct (it understands "crypto" and "ssl" on its
8232 command line). There is also now an "update" option. This will update
8233 the util/ssleay.num and util/libeay.num files with any new functions.
8235 perl util/mkdef.pl crypto ssl update
8236 it will update them.
8239 *) Overhauled the Perl interface (perl/*):
8240 - ported BN stuff to OpenSSL's different BN library
8241 - made the perl/ source tree CVS-aware
8242 - renamed the package from SSLeay to OpenSSL (the files still contain
8243 their history because I've copied them in the repository)
8244 - removed obsolete files (the test scripts will be replaced
8245 by better Test::Harness variants in the future)
8246 [Ralf S. Engelschall]
8248 *) First cut for a very conservative source tree cleanup:
8249 1. merge various obsolete readme texts into doc/ssleay.txt
8250 where we collect the old documents and readme texts.
8251 2. remove the first part of files where I'm already sure that we no
8252 longer need them because of three reasons: either they are just temporary
8253 files which were left by Eric or they are preserved original files where
8254 I've verified that the diff is also available in the CVS via "cvs diff
8255 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
8256 the crypto/md/ stuff).
8257 [Ralf S. Engelschall]
8259 *) More extension code. Incomplete support for subject and issuer alt
8260 name, issuer and authority key id. Change the i2v function parameters
8261 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
8262 what that's for :-) Fix to ASN1 macro which messed up
8263 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
8266 *) Preliminary support for ENUMERATED type. This is largely copied from the
8270 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
8271 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8273 *) Make sure `make rehash' target really finds the `openssl' program.
8274 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8276 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
8277 like to hear about it if this slows down other processors.
8280 *) Add CygWin32 platform information to Configure script.
8281 [Alan Batie <batie@aahz.jf.intel.com>]
8283 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
8284 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
8286 *) New program nseq to manipulate netscape certificate sequences
8289 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
8293 *) Fixes to BN code. Previously the default was to define BN_RECURSION
8294 but the BN code had some problems that would cause failures when
8295 doing certificate verification and some other functions.
8296 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8298 *) Add ASN1 and PEM code to support netscape certificate sequences.
8301 *) Add ASN1 and PEM code to support netscape certificate sequences.
8304 *) Add several PKIX and private extended key usage OIDs.
8307 *) Modify the 'ca' program to handle the new extension code. Modify
8308 openssl.cnf for new extension format, add comments.
8311 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
8312 and add a sample to openssl.cnf so req -x509 now adds appropriate
8316 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
8317 error code, add initial support to X509_print() and x509 application.
8320 *) Takes a deep breath and start addding X509 V3 extension support code. Add
8321 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
8322 stuff is currently isolated and isn't even compiled yet.
8325 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
8326 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
8327 Removed the versions check from X509 routines when loading extensions:
8328 this allows certain broken certificates that don't set the version
8329 properly to be processed.
8332 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
8333 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
8334 can still be regenerated with "make depend".
8337 *) Spelling mistake in C version of CAST-128.
8338 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
8340 *) Changes to the error generation code. The perl script err-code.pl
8341 now reads in the old error codes and retains the old numbers, only
8342 adding new ones if necessary. It also only changes the .err files if new
8343 codes are added. The makefiles have been modified to only insert errors
8344 when needed (to avoid needlessly modifying header files). This is done
8345 by only inserting errors if the .err file is newer than the auto generated
8346 C file. To rebuild all the error codes from scratch (the old behaviour)
8347 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
8348 or delete all the .err files.
8351 *) CAST-128 was incorrectly implemented for short keys. The C version has
8352 been fixed, but is untested. The assembler versions are also fixed, but
8353 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
8354 to regenerate it if needed.
8355 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
8356 Hagino <itojun@kame.net>]
8358 *) File was opened incorrectly in randfile.c.
8359 [Ulf Möller <ulf@fitug.de>]
8361 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
8362 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
8363 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
8364 al: it's just almost always a UTCTime. Note this patch adds new error
8365 codes so do a "make errors" if there are problems.
8368 *) Correct Linux 1 recognition in config.
8369 [Ulf Möller <ulf@fitug.de>]
8371 *) Remove pointless MD5 hash when using DSA keys in ca.
8372 [Anonymous <nobody@replay.com>]
8374 *) Generate an error if given an empty string as a cert directory. Also
8375 generate an error if handed NULL (previously returned 0 to indicate an
8376 error, but didn't set one).
8377 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
8379 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
8382 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
8383 parameters. This was causing a warning which killed off the Win32 compile.
8386 *) Remove C++ style comments from crypto/bn/bn_local.h.
8387 [Neil Costigan <neil.costigan@celocom.com>]
8389 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
8390 based on a text string, looking up short and long names and finally
8391 "dot" format. The "dot" format stuff didn't work. Added new function
8392 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
8393 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
8394 OID is not part of the table.
8397 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
8398 X509_LOOKUP_by_alias().
8401 *) Sort openssl functions by name.
8404 *) Get the gendsa program working (hopefully) and add it to app list. Remove
8405 encryption from sample DSA keys (in case anyone is interested the password
8409 *) Make _all_ *_free functions accept a NULL pointer.
8410 [Frans Heymans <fheymans@isaserver.be>]
8412 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
8414 [Anonymous <nobody@replay.com>]
8416 *) s_server should send the CAfile as acceptable CAs, not its own cert.
8417 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8419 *) Don't blow it for numeric -newkey arguments to apps/req.
8420 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8422 *) Temp key "for export" tests were wrong in s3_srvr.c.
8423 [Anonymous <nobody@replay.com>]
8425 *) Add prototype for temp key callback functions
8426 SSL_CTX_set_tmp_{rsa,dh}_callback().
8429 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
8430 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
8433 *) X509_name_add_entry() freed the wrong thing after an error.
8434 [Arne Ansper <arne@ats.cyber.ee>]
8436 *) rsa_eay.c would attempt to free a NULL context.
8437 [Arne Ansper <arne@ats.cyber.ee>]
8439 *) BIO_s_socket() had a broken should_retry() on Windoze.
8440 [Arne Ansper <arne@ats.cyber.ee>]
8442 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
8443 [Arne Ansper <arne@ats.cyber.ee>]
8445 *) Make sure the already existing X509_STORE->depth variable is initialized
8446 in X509_STORE_new(), but document the fact that this variable is still
8447 unused in the certificate verification process.
8448 [Ralf S. Engelschall]
8450 *) Fix the various library and apps files to free up pkeys obtained from
8451 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
8454 *) Fix reference counting in X509_PUBKEY_get(). This makes
8455 demos/maurice/example2.c work, amongst others, probably.
8456 [Steve Henson and Ben Laurie]
8458 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
8459 `openssl' and second, the shortcut symlinks for the `openssl <command>'
8460 are no longer created. This way we have a single and consistent command
8461 line interface `openssl <command>', similar to `cvs <command>'.
8462 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
8464 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
8465 BIT STRING wrapper always have zero unused bits.
8468 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
8471 *) Make the top-level INSTALL documentation easier to understand.
8474 *) Makefiles updated to exit if an error occurs in a sub-directory
8475 make (including if user presses ^C) [Paul Sutton]
8477 *) Make Montgomery context stuff explicit in RSA data structure.
8480 *) Fix build order of pem and err to allow for generated pem.h.
8483 *) Fix renumbering bug in X509_NAME_delete_entry().
8486 *) Enhanced the err-ins.pl script so it makes the error library number
8487 global and can add a library name. This is needed for external ASN1 and
8488 other error libraries.
8491 *) Fixed sk_insert which never worked properly.
8494 *) Fix ASN1 macros so they can handle indefinite length construted
8495 EXPLICIT tags. Some non standard certificates use these: they can now
8499 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
8500 into a single doc/ssleay.txt bundle. This way the information is still
8501 preserved but no longer messes up this directory. Now it's new room for
8502 the new set of documenation files.
8503 [Ralf S. Engelschall]
8505 *) SETs were incorrectly DER encoded. This was a major pain, because they
8506 shared code with SEQUENCEs, which aren't coded the same. This means that
8507 almost everything to do with SETs or SEQUENCEs has either changed name or
8508 number of arguments.
8509 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
8511 *) Fix test data to work with the above.
8514 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
8515 was already fixed by Eric for 0.9.1 it seems.
8516 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
8518 *) Autodetect FreeBSD3.
8521 *) Fix various bugs in Configure. This affects the following platforms:
8525 unixware-2.0-pentium
8529 *) Eliminate generated files from CVS. Reorder tests to regenerate files
8530 before they are needed.
8533 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
8537 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
8539 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
8540 changed SSLeay to OpenSSL in version strings.
8541 [Ralf S. Engelschall]
8543 *) Some fixups to the top-level documents.
8546 *) Fixed the nasty bug where rsaref.h was not found under compile-time
8547 because the symlink to include/ was missing.
8548 [Ralf S. Engelschall]
8550 *) Incorporated the popular no-RSA/DSA-only patches
8551 which allow to compile a RSA-free SSLeay.
8552 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
8554 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
8555 when "ssleay" is still not found.
8556 [Ralf S. Engelschall]
8558 *) Added more platforms to Configure: Cray T3E, HPUX 11,
8559 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
8561 *) Updated the README file.
8562 [Ralf S. Engelschall]
8564 *) Added various .cvsignore files in the CVS repository subdirs
8565 to make a "cvs update" really silent.
8566 [Ralf S. Engelschall]
8568 *) Recompiled the error-definition header files and added
8569 missing symbols to the Win32 linker tables.
8570 [Ralf S. Engelschall]
8572 *) Cleaned up the top-level documents;
8573 o new files: CHANGES and LICENSE
8574 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
8575 o merged COPYRIGHT into LICENSE
8576 o removed obsolete TODO file
8577 o renamed MICROSOFT to INSTALL.W32
8578 [Ralf S. Engelschall]
8580 *) Removed dummy files from the 0.9.1b source tree:
8581 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
8582 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
8583 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
8584 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
8585 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
8586 [Ralf S. Engelschall]
8588 *) Added various platform portability fixes.
8591 *) The Genesis of the OpenSSL rpject:
8592 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
8593 Young and Tim J. Hudson created while they were working for C2Net until
8595 [The OpenSSL Project]
8598 Changes between 0.9.0b and 0.9.1b [not released]
8600 *) Updated a few CA certificates under certs/
8603 *) Changed some BIGNUM api stuff.
8606 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
8607 DGUX x86, Linux Alpha, etc.
8610 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
8611 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
8615 *) Add -strparse option to asn1pars program which parses nested
8617 [Dr Stephen Henson <shenson@bigfoot.com>]
8619 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
8622 *) DSA fix for "ca" program.
8625 *) Added "-genkey" option to "dsaparam" program.
8628 *) Added RIPE MD160 (rmd160) message digest.
8631 *) Added -a (all) option to "ssleay version" command.
8634 *) Added PLATFORM define which is the id given to Configure.
8637 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
8640 *) Extended the ASN.1 parser routines.
8643 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
8646 *) Added a BN_CTX to the BN library.
8649 *) Fixed the weak key values in DES library
8652 *) Changed API in EVP library for cipher aliases.
8655 *) Added support for RC2/64bit cipher.
8658 *) Converted the lhash library to the crypto/mem.c functions.
8661 *) Added more recognized ASN.1 object ids.
8664 *) Added more RSA padding checks for SSL/TLS.
8667 *) Added BIO proxy/filter functionality.
8670 *) Added extra_certs to SSL_CTX which can be used
8671 send extra CA certificates to the client in the CA cert chain sending
8672 process. It can be configured with SSL_CTX_add_extra_chain_cert().
8675 *) Now Fortezza is denied in the authentication phase because
8676 this is key exchange mechanism is not supported by SSLeay at all.
8679 *) Additional PKCS1 checks.
8682 *) Support the string "TLSv1" for all TLS v1 ciphers.
8685 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
8686 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
8689 *) Fixed a few memory leaks.
8692 *) Fixed various code and comment typos.
8695 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
8696 bytes sent in the client random.
8697 [Edward Bishop <ebishop@spyglass.com>]