5 Changes between 0.9.8v and 0.9.8w [xx XXX xxxx]
9 Changes between 0.9.8u and 0.9.8v [19 Apr 2012]
11 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
12 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
13 in CRYPTO_realloc_clean.
15 Thanks to Tavis Ormandy, Google Security Team, for discovering this
16 issue and to Adam Langley <agl@chromium.org> for fixing it.
18 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
20 Changes between 0.9.8t and 0.9.8u [12 Mar 2012]
22 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
23 in CMS and PKCS7 code. When RSA decryption fails use a random key for
24 content decryption and always return the same error. Note: this attack
25 needs on average 2^20 messages so it only affects automated senders. The
26 old behaviour can be reenabled in the CMS code by setting the
27 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
28 an MMA defence is not necessary.
29 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
30 this issue. (CVE-2012-0884)
33 *) Fix CVE-2011-4619: make sure we really are receiving a
34 client hello before rejecting multiple SGC restarts. Thanks to
35 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
38 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
40 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
41 Thanks to Antonio Martin, Enterprise Secure Access Research and
42 Development, Cisco Systems, Inc. for discovering this bug and
43 preparing a fix. (CVE-2012-0050)
46 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
48 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
49 of the Vaudenay padding oracle attack on CBC mode encryption
50 which enables an efficient plaintext recovery attack against
51 the OpenSSL implementation of DTLS. Their attack exploits timing
52 differences arising during decryption processing. A research
53 paper describing this attack can be found at:
54 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
55 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
56 Security Group at Royal Holloway, University of London
57 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
58 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
59 for preparing the fix. (CVE-2011-4108)
60 [Robin Seggelmann, Michael Tuexen]
62 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
63 [Ben Laurie, Kasper <ekasper@google.com>]
65 *) Clear bytes used for block padding of SSL 3.0 records.
67 [Adam Langley (Google)]
69 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
70 Kadianakis <desnacked@gmail.com> for discovering this issue and
71 Adam Langley for preparing the fix. (CVE-2011-4619)
72 [Adam Langley (Google)]
74 *) Prevent malformed RFC3779 data triggering an assertion failure.
75 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
76 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
77 [Rob Austein <sra@hactrn.net>]
79 *) Fix ssl_ciph.c set-up race.
80 [Adam Langley (Google)]
82 *) Fix spurious failures in ecdsatest.c.
83 [Emilia Käsper (Google)]
85 *) Fix the BIO_f_buffer() implementation (which was mixing different
86 interpretations of the '..._len' fields).
87 [Adam Langley (Google)]
89 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
90 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
91 threads won't reuse the same blinding coefficients.
93 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
94 lock to call BN_BLINDING_invert_ex, and avoids one use of
95 BN_BLINDING_update for each BN_BLINDING structure (previously,
96 the last update always remained unused).
97 [Emilia Käsper (Google)]
99 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
100 for multi-threaded use of ECDH.
101 [Adam Langley (Google)]
103 *) Fix x509_name_ex_d2i memory leak on bad inputs.
106 *) Add protection against ECDSA timing attacks as mentioned in the paper
107 by Billy Bob Brumley and Nicola Tuveri, see:
109 http://eprint.iacr.org/2011/232.pdf
111 [Billy Bob Brumley and Nicola Tuveri]
113 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
115 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
116 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
118 *) Fix bug in string printing code: if *any* escaping is enabled we must
119 escape the escape character (backslash) or the resulting string is
123 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
125 *) Disable code workaround for ancient and obsolete Netscape browsers
126 and servers: an attacker can use it in a ciphersuite downgrade attack.
127 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
130 *) Fixed J-PAKE implementation error, originally discovered by
131 Sebastien Martini, further info and confirmation from Stefan
132 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
135 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
137 *) Fix extension code to avoid race conditions which can result in a buffer
138 overrun vulnerability: resumed sessions must not be modified as they can
139 be shared by multiple threads. CVE-2010-3864
142 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
145 *) Don't reencode certificate when calculating signature: cache and use
146 the original encoding instead. This makes signature verification of
147 some broken encodings work correctly.
150 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
151 is also one of the inputs.
152 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
154 *) Don't repeatedly append PBE algorithms to table if they already exist.
155 Sort table on each new add. This effectively makes the table read only
156 after all algorithms are added and subsequent calls to PKCS12_pbe_add
160 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
162 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
165 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
166 access or freeing data twice (CVE-2010-0742)
167 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
169 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
170 common in certificates and some applications which only call
171 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
175 Reduce copying into .apps and .test in makevms.com
176 Don't try to use blank CA certificate in CA.com
177 Allow use of C files from original directories in maketests.com
178 [Steven M. Schweda" <sms@antinode.info>]
180 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
182 *) When rejecting SSL/TLS records due to an incorrect version number, never
183 update s->server with a new major version number. As of
184 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
185 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
186 the previous behavior could result in a read attempt at NULL when
187 receiving specific incorrect SSL/TLS records once record payload
188 protection is active. (CVE-2010-0740)
189 [Bodo Moeller, Adam Langley <agl@chromium.org>]
191 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
192 could be crashed if the relevant tables were not present (e.g. chrooted).
193 [Tomas Hoger <thoger@redhat.com>]
195 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
197 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
198 [Martin Olsson, Neel Mehta]
200 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
201 accommodate for stack sorting, always a write lock!).
204 *) On some versions of WIN32 Heap32Next is very slow. This can cause
205 excessive delays in the RAND_poll(): over a minute. As a workaround
206 include a time check in the inner Heap32Next loop too.
209 *) The code that handled flushing of data in SSL/TLS originally used the
210 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
211 the problem outlined in PR#1949. The fix suggested there however can
212 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
213 of Apache). So instead simplify the code to flush unconditionally.
214 This should be fine since flushing with no data to flush is a no op.
217 *) Handle TLS versions 2.0 and later properly and correctly use the
218 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
219 off ancient servers have a habit of sticking around for a while...
222 *) Modify compression code so it frees up structures without using the
223 ex_data callbacks. This works around a problem where some applications
224 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
225 restarting) then use compression (e.g. SSL with compression) later.
226 This results in significant per-connection memory leaks and
227 has caused some security issues including CVE-2008-1678 and
231 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
232 change when encrypting or decrypting.
235 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
236 connect and renegotiate with servers which do not support RI.
237 Until RI is more widely deployed this option is enabled by default.
240 *) Add "missing" ssl ctrls to clear options and mode.
243 *) If client attempts to renegotiate and doesn't support RI respond with
244 a no_renegotiation alert as required by RFC5746. Some renegotiating
245 TLS clients will continue a connection gracefully when they receive
246 the alert. Unfortunately OpenSSL mishandled this alert and would hang
247 waiting for a server hello which it will never receive. Now we treat a
248 received no_renegotiation alert as a fatal error. This is because
249 applications requesting a renegotiation might well expect it to succeed
250 and would have no code in place to handle the server denying it so the
251 only safe thing to do is to terminate the connection.
254 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
255 peer supports secure renegotiation and 0 otherwise. Print out peer
256 renegotiation support in s_client/s_server.
259 *) Replace the highly broken and deprecated SPKAC certification method with
260 the updated NID creation version. This should correctly handle UTF8.
263 *) Implement RFC5746. Re-enable renegotiation but require the extension
264 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
265 turns out to be a bad idea. It has been replaced by
266 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
267 SSL_CTX_set_options(). This is really not recommended unless you
268 know what you are doing.
269 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
271 *) Fixes to stateless session resumption handling. Use initial_ctx when
272 issuing and attempting to decrypt tickets in case it has changed during
273 servername handling. Use a non-zero length session ID when attempting
274 stateless session resumption: this makes it possible to determine if
275 a resumption has occurred immediately after receiving server hello
276 (several places in OpenSSL subtly assume this) instead of later in
280 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
281 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
282 fixes for a few places where the return code is not checked
284 [Julia Lawall <julia@diku.dk>]
286 *) Add --strict-warnings option to Configure script to include devteam
287 warnings in other configurations.
290 *) Add support for --libdir option and LIBDIR variable in makefiles. This
291 makes it possible to install openssl libraries in locations which
292 have names other than "lib", for example "/usr/lib64" which some
294 [Steve Henson, based on patch from Jeremy Utley]
296 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
297 X690 8.9.12 and can produce some misleading textual output of OIDs.
298 [Steve Henson, reported by Dan Kaminsky]
300 *) Delete MD2 from algorithm tables. This follows the recommendation in
301 several standards that it is not used in new applications due to
302 several cryptographic weaknesses. For binary compatibility reasons
303 the MD2 API is still compiled in by default.
306 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
310 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
311 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
313 [Guenter <lists@gknw.net>]
315 *) Fix the server certificate chain building code to use X509_verify_cert(),
316 it used to have an ad-hoc builder which was unable to cope with anything
317 other than a simple chain.
318 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
320 *) Don't check self signed certificate signatures in X509_verify_cert()
321 by default (a flag can override this): it just wastes time without
322 adding any security. As a useful side effect self signed root CAs
323 with non-FIPS digests are now usable in FIPS mode.
326 *) In dtls1_process_out_of_seq_message() the check if the current message
327 is already buffered was missing. For every new message was memory
328 allocated, allowing an attacker to perform an denial of service attack
329 with sending out of seq handshake messages until there is no memory
330 left. Additionally every future messege was buffered, even if the
331 sequence number made no sense and would be part of another handshake.
332 So only messages with sequence numbers less than 10 in advance will be
333 buffered. (CVE-2009-1378)
334 [Robin Seggelmann, discovered by Daniel Mentz]
336 *) Records are buffered if they arrive with a future epoch to be
337 processed after finishing the corresponding handshake. There is
338 currently no limitation to this buffer allowing an attacker to perform
339 a DOS attack with sending records with future epochs until there is no
340 memory left. This patch adds the pqueue_size() function to detemine
341 the size of a buffer and limits the record buffer to 100 entries.
343 [Robin Seggelmann, discovered by Daniel Mentz]
345 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
346 parent structure is freed. (CVE-2009-1379)
349 *) Handle non-blocking I/O properly in SSL_shutdown() call.
350 [Darryl Miles <darryl-mailinglists@netbauds.net>]
353 [Ilya O. <vrghost@gmail.com>]
355 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
357 *) Disable renegotiation completely - this fixes a severe security
358 problem (CVE-2009-3555) at the cost of breaking all
359 renegotiation. Renegotiation can be re-enabled by setting
360 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
361 run-time. This is really not recommended unless you know what
365 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
367 *) Don't set val to NULL when freeing up structures, it is freed up by
368 underlying code. If sizeof(void *) > sizeof(long) this can result in
369 zeroing past the valid field. (CVE-2009-0789)
370 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
372 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
373 checked correctly. This would allow some invalid signed attributes to
374 appear to verify correctly. (CVE-2009-0591)
375 [Ivan Nestlerode <inestlerode@us.ibm.com>]
377 *) Reject UniversalString and BMPString types with invalid lengths. This
378 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
379 a legal length. (CVE-2009-0590)
382 *) Set S/MIME signing as the default purpose rather than setting it
383 unconditionally. This allows applications to override it at the store
387 *) Permit restricted recursion of ASN1 strings. This is needed in practice
388 to handle some structures.
391 *) Improve efficiency of mem_gets: don't search whole buffer each time
393 [Jeremy Shapiro <jnshapir@us.ibm.com>]
395 *) New -hex option for openssl rand.
398 *) Print out UTF8String and NumericString when parsing ASN1.
401 *) Support NumericString type for name components.
404 *) Allow CC in the environment to override the automatically chosen
405 compiler. Note that nothing is done to ensure flags work with the
409 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
411 *) Properly check EVP_VerifyFinal() and similar return values
413 [Ben Laurie, Bodo Moeller, Google Security Team]
415 *) Enable TLS extensions by default.
418 *) Allow the CHIL engine to be loaded, whether the application is
419 multithreaded or not. (This does not release the developer from the
420 obligation to set up the dynamic locking callbacks.)
421 [Sander Temme <sander@temme.net>]
423 *) Use correct exit code if there is an error in dgst command.
424 [Steve Henson; problem pointed out by Roland Dirlewanger]
426 *) Tweak Configure so that you need to say "experimental-jpake" to enable
427 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
430 *) Add experimental JPAKE support, including demo authentication in
431 s_client and s_server.
434 *) Set the comparison function in v3_addr_canonize().
435 [Rob Austein <sra@hactrn.net>]
437 *) Add support for XMPP STARTTLS in s_client.
438 [Philip Paeps <philip@freebsd.org>]
440 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
441 to ensure that even with this option, only ciphersuites in the
442 server's preference list will be accepted. (Note that the option
443 applies only when resuming a session, so the earlier behavior was
444 just about the algorithm choice for symmetric cryptography.)
447 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
449 *) Fix NULL pointer dereference if a DTLS server received
450 ChangeCipherSpec as first record (CVE-2009-1386).
453 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
454 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
457 *) The fix in 0.9.8c that supposedly got rid of unsafe
458 double-checked locking was incomplete for RSA blinding,
459 addressing just one layer of what turns out to have been
460 doubly unsafe triple-checked locking.
462 So now fix this for real by retiring the MONT_HELPER macro
463 in crypto/rsa/rsa_eay.c.
465 [Bodo Moeller; problem pointed out by Marius Schilder]
467 *) Various precautionary measures:
469 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
471 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
472 (NB: This would require knowledge of the secret session ticket key
473 to exploit, in which case you'd be SOL either way.)
475 - Change bn_nist.c so that it will properly handle input BIGNUMs
476 outside the expected range.
478 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
481 [Neel Mehta, Bodo Moeller]
483 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
484 the load fails. Useful for distros.
485 [Ben Laurie and the FreeBSD team]
487 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
490 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
493 *) Expand ENGINE to support engine supplied SSL client certificate functions.
495 This work was sponsored by Logica.
498 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
499 keystores. Support for SSL/TLS client authentication too.
500 Not compiled unless enable-capieng specified to Configure.
502 This work was sponsored by Logica.
505 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
506 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
507 attribute creation routines such as certifcate requests and PKCS#12
511 Changes between 0.9.8g and 0.9.8h [28 May 2008]
513 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
514 handshake which could lead to a cilent crash as found using the
515 Codenomicon TLS test suite (CVE-2008-1672)
516 [Steve Henson, Mark Cox]
518 *) Fix double free in TLS server name extensions which could lead to
519 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
522 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
524 Clear the error queue to ensure that error entries left from
525 older function calls do not interfere with the correct operation.
526 [Lutz Jaenicke, Erik de Castro Lopo]
528 *) Remove root CA certificates of commercial CAs:
530 The OpenSSL project does not recommend any specific CA and does not
531 have any policy with respect to including or excluding any CA.
532 Therefore it does not make any sense to ship an arbitrary selection
533 of root CA certificates with the OpenSSL software.
536 *) RSA OAEP patches to fix two separate invalid memory reads.
537 The first one involves inputs when 'lzero' is greater than
538 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
539 before the beginning of from). The second one involves inputs where
540 the 'db' section contains nothing but zeroes (there is a one-byte
541 invalid read after the end of 'db').
542 [Ivan Nestlerode <inestlerode@us.ibm.com>]
544 *) Partial backport from 0.9.9-dev:
546 Introduce bn_mul_mont (dedicated Montgomery multiplication
547 procedure) as a candidate for BIGNUM assembler implementation.
548 While 0.9.9-dev uses assembler for various architectures, only
549 x86_64 is available by default here in the 0.9.8 branch, and
550 32-bit x86 is available through a compile-time setting.
552 To try the 32-bit x86 assembler implementation, use Configure
553 option "enable-montasm" (which exists only for this backport).
555 As "enable-montasm" for 32-bit x86 disclaims code stability
556 anyway, in this constellation we activate additional code
557 backported from 0.9.9-dev for further performance improvements,
558 namely BN_from_montgomery_word. (To enable this otherwise,
559 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
561 [Andy Polyakov (backport partially by Bodo Moeller)]
563 *) Add TLS session ticket callback. This allows an application to set
564 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
565 values. This is useful for key rollover for example where several key
566 sets may exist with different names.
569 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
570 This was broken until now in 0.9.8 releases, such that the only way
571 a registered ENGINE could be used (assuming it initialises
572 successfully on the host) was to explicitly set it as the default
573 for the relevant algorithms. This is in contradiction with 0.9.7
574 behaviour and the documentation. With this fix, when an ENGINE is
575 registered into a given algorithm's table of implementations, the
576 'uptodate' flag is reset so that auto-discovery will be used next
577 time a new context for that algorithm attempts to select an
579 [Ian Lister (tweaked by Geoff Thorpe)]
581 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
582 implemention in the following ways:
584 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
587 Lack of BER streaming support means one pass streaming processing is
588 only supported if data is detached: setting the streaming flag is
589 ignored for embedded content.
591 CMS support is disabled by default and must be explicitly enabled
592 with the enable-cms configuration option.
595 *) Update the GMP engine glue to do direct copies between BIGNUM and
596 mpz_t when openssl and GMP use the same limb size. Otherwise the
597 existing "conversion via a text string export" trick is still used.
598 [Paul Sheer <paulsheer@gmail.com>]
600 *) Zlib compression BIO. This is a filter BIO which compressed and
601 uncompresses any data passed through it.
604 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
605 RFC3394 compatible AES key wrapping.
608 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
609 sets string data without copying. X509_ALGOR_set0() and
610 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
611 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
612 from an X509_ATTRIBUTE structure optionally checking it occurs only
613 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
617 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
618 to get the expected BN_FLG_CONSTTIME behavior.
619 [Bodo Moeller (Google)]
623 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
624 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
625 - added some more tests to do_tests.pl
626 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
627 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
628 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
629 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
630 - various changes to netware.pl to enable gcc-cross builds on Win32
632 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
633 - various changes to fix missing prototype warnings
634 - fixed x86nasm.pl to create correct asm files for NASM COFF output
635 - added AES, WHIRLPOOL and CPUID assembler code to build files
636 - added missing AES assembler make rules to mk1mf.pl
637 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
638 [Guenter Knauf <eflash@gmx.net>]
640 *) Implement certificate status request TLS extension defined in RFC3546.
641 A client can set the appropriate parameters and receive the encoded
642 OCSP response via a callback. A server can query the supplied parameters
643 and set the encoded OCSP response in the callback. Add simplified examples
644 to s_client and s_server.
647 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
650 + Binary incompatibility of ssl_ctx_st structure
651 + DTLS interoperation with non-compliant servers
652 + Don't call get_session_cb() without proposed session
653 + Fix ia64 assembler code
654 [Andy Polyakov, Steve Henson]
656 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
658 *) DTLS Handshake overhaul. There were longstanding issues with
659 OpenSSL DTLS implementation, which were making it impossible for
660 RFC 4347 compliant client to communicate with OpenSSL server.
661 Unfortunately just fixing these incompatibilities would "cut off"
662 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
663 server keeps tolerating non RFC compliant syntax. The opposite is
664 not true, 0.9.8f client can not communicate with earlier server.
665 This update even addresses CVE-2007-4995.
668 *) Changes to avoid need for function casts in OpenSSL: some compilers
669 (gcc 4.2 and later) reject their use.
670 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
673 *) Add RFC4507 support to OpenSSL. This includes the corrections in
674 RFC4507bis. The encrypted ticket format is an encrypted encoded
675 SSL_SESSION structure, that way new session features are automatically
678 If a client application caches session in an SSL_SESSION structure
679 support is transparent because tickets are now stored in the encoded
682 The SSL_CTX structure automatically generates keys for ticket
683 protection in servers so again support should be possible
684 with no application modification.
686 If a client or server wishes to disable RFC4507 support then the option
687 SSL_OP_NO_TICKET can be set.
689 Add a TLS extension debugging callback to allow the contents of any client
690 or server extensions to be examined.
692 This work was sponsored by Google.
695 *) Add initial support for TLS extensions, specifically for the server_name
696 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
697 have new members for a host name. The SSL data structure has an
698 additional member SSL_CTX *initial_ctx so that new sessions can be
699 stored in that context to allow for session resumption, even after the
700 SSL has been switched to a new SSL_CTX in reaction to a client's
701 server_name extension.
703 New functions (subject to change):
706 SSL_get_servername_type()
709 New CTRL codes and macros (subject to change):
711 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
712 - SSL_CTX_set_tlsext_servername_callback()
713 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
714 - SSL_CTX_set_tlsext_servername_arg()
715 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
717 openssl s_client has a new '-servername ...' option.
719 openssl s_server has new options '-servername_host ...', '-cert2 ...',
720 '-key2 ...', '-servername_fatal' (subject to change). This allows
721 testing the HostName extension for a specific single host name ('-cert'
722 and '-key' remain fallbacks for handshakes without HostName
723 negotiation). If the unrecogninzed_name alert has to be sent, this by
724 default is a warning; it becomes fatal with the '-servername_fatal'
727 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
729 *) Add AES and SSE2 assembly language support to VC++ build.
732 *) Mitigate attack on final subtraction in Montgomery reduction.
735 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
736 (which previously caused an internal error).
739 *) Squeeze another 10% out of IGE mode when in != out.
742 *) AES IGE mode speedup.
743 [Dean Gaudet (Google)]
745 *) Add the Korean symmetric 128-bit cipher SEED (see
746 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
747 add SEED ciphersuites from RFC 4162:
749 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
750 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
751 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
752 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
754 To minimize changes between patchlevels in the OpenSSL 0.9.8
755 series, SEED remains excluded from compilation unless OpenSSL
756 is configured with 'enable-seed'.
759 *) Mitigate branch prediction attacks, which can be practical if a
760 single processor is shared, allowing a spy process to extract
761 information. For detailed background information, see
762 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
763 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
764 and Necessary Software Countermeasures"). The core of the change
765 are new versions BN_div_no_branch() and
766 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
767 respectively, which are slower, but avoid the security-relevant
768 conditional branches. These are automatically called by BN_div()
769 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
770 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
771 remove a conditional branch.
773 BN_FLG_CONSTTIME is the new name for the previous
774 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
775 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
776 in the exponent causes BN_mod_exp_mont() to use the alternative
777 implementation in BN_mod_exp_mont_consttime().) The old name
778 remains as a deprecated alias.
780 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
781 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
782 constant-time implementations for more than just exponentiation.
783 Here too the old name is kept as a deprecated alias.
785 BN_BLINDING_new() will now use BN_dup() for the modulus so that
786 the BN_BLINDING structure gets an independent copy of the
787 modulus. This means that the previous "BIGNUM *m" argument to
788 BN_BLINDING_new() and to BN_BLINDING_create_param() now
789 essentially becomes "const BIGNUM *m", although we can't actually
790 change this in the header file before 0.9.9. It allows
791 RSA_setup_blinding() to use BN_with_flags() on the modulus to
792 enable BN_FLG_CONSTTIME.
794 [Matthew D Wood (Intel Corp)]
796 *) In the SSL/TLS server implementation, be strict about session ID
797 context matching (which matters if an application uses a single
798 external cache for different purposes). Previously,
799 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
800 set. This did ensure strict client verification, but meant that,
801 with applications using a single external cache for quite
802 different requirements, clients could circumvent ciphersuite
803 restrictions for a given session ID context by starting a session
804 in a different context.
807 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
808 a ciphersuite string such as "DEFAULT:RSA" cannot enable
809 authentication-only ciphersuites.
812 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
813 not complete and could lead to a possible single byte overflow
814 (CVE-2007-5135) [Ben Laurie]
816 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
818 *) Since AES128 and AES256 (and similarly Camellia128 and
819 Camellia256) share a single mask bit in the logic of
820 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
821 kludge to work properly if AES128 is available and AES256 isn't
822 (or if Camellia128 is available and Camellia256 isn't).
825 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
826 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
827 When a point or a seed is encoded in a BIT STRING, we need to
828 prevent the removal of trailing zero bits to get the proper DER
829 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
830 of a NamedBitList, for which trailing 0 bits need to be removed.)
833 *) Have SSL/TLS server implementation tolerate "mismatched" record
834 protocol version while receiving ClientHello even if the
835 ClientHello is fragmented. (The server can't insist on the
836 particular protocol version it has chosen before the ServerHello
837 message has informed the client about his choice.)
840 *) Add RFC 3779 support.
841 [Rob Austein for ARIN, Ben Laurie]
843 *) Load error codes if they are not already present instead of using a
844 static variable. This allows them to be cleanly unloaded and reloaded.
845 Improve header file function name parsing.
848 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
849 or CAPABILITY handshake as required by RFCs.
852 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
854 *) Introduce limits to prevent malicious keys being able to
855 cause a denial of service. (CVE-2006-2940)
856 [Steve Henson, Bodo Moeller]
858 *) Fix ASN.1 parsing of certain invalid structures that can result
859 in a denial of service. (CVE-2006-2937) [Steve Henson]
861 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
862 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
864 *) Fix SSL client code which could crash if connecting to a
865 malicious SSLv2 server. (CVE-2006-4343)
866 [Tavis Ormandy and Will Drewry, Google Security Team]
868 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
869 match only those. Before that, "AES256-SHA" would be interpreted
870 as a pattern and match "AES128-SHA" too (since AES128-SHA got
871 the same strength classification in 0.9.7h) as we currently only
872 have a single AES bit in the ciphersuite description bitmap.
873 That change, however, also applied to ciphersuite strings such as
874 "RC4-MD5" that intentionally matched multiple ciphersuites --
875 namely, SSL 2.0 ciphersuites in addition to the more common ones
876 from SSL 3.0/TLS 1.0.
878 So we change the selection algorithm again: Naming an explicit
879 ciphersuite selects this one ciphersuite, and any other similar
880 ciphersuite (same bitmap) from *other* protocol versions.
881 Thus, "RC4-MD5" again will properly select both the SSL 2.0
882 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
884 Since SSL 2.0 does not have any ciphersuites for which the
885 128/256 bit distinction would be relevant, this works for now.
886 The proper fix will be to use different bits for AES128 and
887 AES256, which would have avoided the problems from the beginning;
888 however, bits are scarce, so we can only do this in a new release
889 (not just a patchlevel) when we can change the SSL_CIPHER
890 definition to split the single 'unsigned long mask' bitmap into
891 multiple values to extend the available space.
895 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
897 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
898 (CVE-2006-4339) [Ben Laurie and Google Security Team]
900 *) Add AES IGE and biIGE modes.
903 *) Change the Unix randomness entropy gathering to use poll() when
904 possible instead of select(), since the latter has some
905 undesirable limitations.
906 [Darryl Miles via Richard Levitte and Bodo Moeller]
908 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
909 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
910 cannot be implicitly activated as part of, e.g., the "AES" alias.
911 However, please upgrade to OpenSSL 0.9.9[-dev] for
912 non-experimental use of the ECC ciphersuites to get TLS extension
913 support, which is required for curve and point format negotiation
914 to avoid potential handshake problems.
917 *) Disable rogue ciphersuites:
919 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
920 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
921 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
923 The latter two were purportedly from
924 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
927 Also deactivate the remaining ciphersuites from
928 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
929 unofficial, and the ID has long expired.
932 *) Fix RSA blinding Heisenbug (problems sometimes occured on
933 dual-core machines) and other potential thread-safety issues.
936 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
937 versions), which is now available for royalty-free use
938 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
939 Also, add Camellia TLS ciphersuites from RFC 4132.
941 To minimize changes between patchlevels in the OpenSSL 0.9.8
942 series, Camellia remains excluded from compilation unless OpenSSL
943 is configured with 'enable-camellia'.
946 *) Disable the padding bug check when compression is in use. The padding
947 bug check assumes the first packet is of even length, this is not
948 necessarily true if compresssion is enabled and can result in false
949 positives causing handshake failure. The actual bug test is ancient
950 code so it is hoped that implementations will either have fixed it by
951 now or any which still have the bug do not support compression.
954 Changes between 0.9.8a and 0.9.8b [04 May 2006]
956 *) When applying a cipher rule check to see if string match is an explicit
957 cipher suite and only match that one cipher suite if it is.
960 *) Link in manifests for VC++ if needed.
961 [Austin Ziegler <halostatue@gmail.com>]
963 *) Update support for ECC-based TLS ciphersuites according to
964 draft-ietf-tls-ecc-12.txt with proposed changes (but without
965 TLS extensions, which are supported starting with the 0.9.9
966 branch, not in the OpenSSL 0.9.8 branch).
969 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
970 opaque EVP_CIPHER_CTX handling.
973 *) Fixes and enhancements to zlib compression code. We now only use
974 "zlib1.dll" and use the default __cdecl calling convention on Win32
975 to conform with the standards mentioned here:
976 http://www.zlib.net/DLL_FAQ.txt
977 Static zlib linking now works on Windows and the new --with-zlib-include
978 --with-zlib-lib options to Configure can be used to supply the location
979 of the headers and library. Gracefully handle case where zlib library
983 *) Several fixes and enhancements to the OID generation code. The old code
984 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
985 handle numbers larger than ULONG_MAX, truncated printing and had a
986 non standard OBJ_obj2txt() behaviour.
989 *) Add support for building of engines under engine/ as shared libraries
990 under VC++ build system.
993 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
994 Hopefully, we will not see any false combination of paths any more.
997 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
999 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1000 (part of SSL_OP_ALL). This option used to disable the
1001 countermeasure against man-in-the-middle protocol-version
1002 rollback in the SSL 2.0 server implementation, which is a bad
1003 idea. (CVE-2005-2969)
1005 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1006 for Information Security, National Institute of Advanced Industrial
1007 Science and Technology [AIST], Japan)]
1009 *) Add two function to clear and return the verify parameter flags.
1012 *) Keep cipherlists sorted in the source instead of sorting them at
1013 runtime, thus removing the need for a lock.
1016 *) Avoid some small subgroup attacks in Diffie-Hellman.
1017 [Nick Mathewson and Ben Laurie]
1019 *) Add functions for well-known primes.
1022 *) Extended Windows CE support.
1023 [Satoshi Nakamura and Andy Polyakov]
1025 *) Initialize SSL_METHOD structures at compile time instead of during
1026 runtime, thus removing the need for a lock.
1029 *) Make PKCS7_decrypt() work even if no certificate is supplied by
1030 attempting to decrypt each encrypted key in turn. Add support to
1034 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
1036 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
1039 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
1042 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
1043 key into the same file any more.
1046 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
1049 *) Add -utf8 command line and config file option to 'ca'.
1050 [Stefan <stf@udoma.org]
1052 *) Removed the macro des_crypt(), as it seems to conflict with some
1053 libraries. Use DES_crypt().
1056 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
1057 involves renaming the source and generated shared-libs for
1058 both. The engines will accept the corrected or legacy ids
1059 ('ncipher' and '4758_cca' respectively) when binding. NB,
1060 this only applies when building 'shared'.
1061 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
1063 *) Add attribute functions to EVP_PKEY structure. Modify
1064 PKCS12_create() to recognize a CSP name attribute and
1065 use it. Make -CSP option work again in pkcs12 utility.
1068 *) Add new functionality to the bn blinding code:
1069 - automatic re-creation of the BN_BLINDING parameters after
1070 a fixed number of uses (currently 32)
1071 - add new function for parameter creation
1072 - introduce flags to control the update behaviour of the
1073 BN_BLINDING parameters
1074 - hide BN_BLINDING structure
1075 Add a second BN_BLINDING slot to the RSA structure to improve
1076 performance when a single RSA object is shared among several
1080 *) Add support for DTLS.
1081 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
1083 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
1084 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
1087 *) Remove buggy and incompletet DH cert support from
1088 ssl/ssl_rsa.c and ssl/s3_both.c
1091 *) Use SHA-1 instead of MD5 as the default digest algorithm for
1092 the apps/openssl applications.
1095 *) Compile clean with "-Wall -Wmissing-prototypes
1096 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
1097 DEBUG_SAFESTACK must also be set.
1100 *) Change ./Configure so that certain algorithms can be disabled by default.
1101 The new counterpiece to "no-xxx" is "enable-xxx".
1103 The patented RC5 and MDC2 algorithms will now be disabled unless
1104 "enable-rc5" and "enable-mdc2", respectively, are specified.
1106 (IDEA remains enabled despite being patented. This is because IDEA
1107 is frequently required for interoperability, and there is no license
1108 fee for non-commercial use. As before, "no-idea" can be used to
1109 avoid this algorithm.)
1113 *) Add processing of proxy certificates (see RFC 3820). This work was
1114 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
1115 EGEE (Enabling Grids for E-science in Europe).
1118 *) RC4 performance overhaul on modern architectures/implementations, such
1119 as Intel P4, IA-64 and AMD64.
1122 *) New utility extract-section.pl. This can be used specify an alternative
1123 section number in a pod file instead of having to treat each file as
1124 a separate case in Makefile. This can be done by adding two lines to the
1127 =for comment openssl_section:XXX
1129 The blank line is mandatory.
1133 *) New arguments -certform, -keyform and -pass for s_client and s_server
1134 to allow alternative format key and certificate files and passphrase
1138 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
1139 update associated structures and add various utility functions.
1141 Add new policy related verify parameters, include policy checking in
1142 standard verify code. Enhance 'smime' application with extra parameters
1143 to support policy checking and print out.
1146 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
1147 Nehemiah processors. These extensions support AES encryption in hardware
1148 as well as RNG (though RNG support is currently disabled).
1149 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
1151 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
1154 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
1155 [Andy Polyakov and a number of other people]
1157 *) Improved PowerPC platform support. Most notably BIGNUM assembler
1158 implementation contributed by IBM.
1159 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
1161 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
1162 exponent rather than 'unsigned long'. There is a corresponding change to
1163 the new 'rsa_keygen' element of the RSA_METHOD structure.
1164 [Jelte Jansen, Geoff Thorpe]
1166 *) Functionality for creating the initial serial number file is now
1167 moved from CA.pl to the 'ca' utility with a new option -create_serial.
1169 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
1170 number file to 1, which is bound to cause problems. To avoid
1171 the problems while respecting compatibility between different 0.9.7
1172 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
1173 CA.pl for serial number initialization. With the new release 0.9.8,
1174 we can fix the problem directly in the 'ca' utility.)
1177 *) Reduced header interdepencies by declaring more opaque objects in
1178 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
1179 give fewer recursive includes, which could break lazy source code - so
1180 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
1181 developers should define this symbol when building and using openssl to
1182 ensure they track the recommended behaviour, interfaces, [etc], but
1183 backwards-compatible behaviour prevails when this isn't defined.
1186 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
1189 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
1190 This will generate a random key of the appropriate length based on the
1191 cipher context. The EVP_CIPHER can provide its own random key generation
1192 routine to support keys of a specific form. This is used in the des and
1193 3des routines to generate a key of the correct parity. Update S/MIME
1194 code to use new functions and hence generate correct parity DES keys.
1195 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
1196 valid (weak or incorrect parity).
1199 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
1200 as looking them up. This is useful when the verified structure may contain
1201 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
1202 present unless the new PKCS7_NO_CRL flag is asserted.
1205 *) Extend ASN1 oid configuration module. It now additionally accepts the
1208 shortName = some long name, 1.2.3.4
1211 *) Reimplemented the BN_CTX implementation. There is now no more static
1212 limitation on the number of variables it can handle nor the depth of the
1213 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
1214 information can now expand as required, and rather than having a single
1215 static array of bignums, BN_CTX now uses a linked-list of such arrays
1216 allowing it to expand on demand whilst maintaining the usefulness of
1217 BN_CTX's "bundling".
1220 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
1221 to allow all RSA operations to function using a single BN_CTX.
1224 *) Preliminary support for certificate policy evaluation and checking. This
1225 is initially intended to pass the tests outlined in "Conformance Testing
1226 of Relying Party Client Certificate Path Processing Logic" v1.07.
1229 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
1230 remained unused and not that useful. A variety of other little bignum
1231 tweaks and fixes have also been made continuing on from the audit (see
1235 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
1236 associated ASN1, EVP and SSL functions and old ASN1 macros.
1239 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
1240 and this should never fail. So the return value from the use of
1241 BN_set_word() (which can fail due to needless expansion) is now deprecated;
1242 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
1245 *) BN_CTX_get() should return zero-valued bignums, providing the same
1246 initialised value as BN_new().
1247 [Geoff Thorpe, suggested by Ulf Möller]
1249 *) Support for inhibitAnyPolicy certificate extension.
1252 *) An audit of the BIGNUM code is underway, for which debugging code is
1253 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
1254 is considered valid when processing BIGNUMs, and causes execution to
1255 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
1256 further steps are taken to deliberately pollute unused data in BIGNUM
1257 structures to try and expose faulty code further on. For now, openssl will
1258 (in its default mode of operation) continue to tolerate the inconsistent
1259 forms that it has tolerated in the past, but authors and packagers should
1260 consider trying openssl and their own applications when compiled with
1261 these debugging symbols defined. It will help highlight potential bugs in
1262 their own code, and will improve the test coverage for OpenSSL itself. At
1263 some point, these tighter rules will become openssl's default to improve
1264 maintainability, though the assert()s and other overheads will remain only
1265 in debugging configurations. See bn.h for more details.
1266 [Geoff Thorpe, Nils Larsch, Ulf Möller]
1268 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
1269 that can only be obtained through BN_CTX_new() (which implicitly
1270 initialises it). The presence of this function only made it possible
1271 to overwrite an existing structure (and cause memory leaks).
1274 *) Because of the callback-based approach for implementing LHASH as a
1275 template type, lh_insert() adds opaque objects to hash-tables and
1276 lh_doall() or lh_doall_arg() are typically used with a destructor callback
1277 to clean up those corresponding objects before destroying the hash table
1278 (and losing the object pointers). So some over-zealous constifications in
1279 LHASH have been relaxed so that lh_insert() does not take (nor store) the
1280 objects as "const" and the lh_doall[_arg] callback wrappers are not
1281 prototyped to have "const" restrictions on the object pointers they are
1282 given (and so aren't required to cast them away any more).
1285 *) The tmdiff.h API was so ugly and minimal that our own timing utility
1286 (speed) prefers to use its own implementation. The two implementations
1287 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
1288 its object type properly exposed (MS_TM) instead of casting to/from "char
1289 *". This may still change yet if someone realises MS_TM and "ms_time_***"
1290 aren't necessarily the greatest nomenclatures - but this is what was used
1291 internally to the implementation so I've used that for now.
1294 *) Ensure that deprecated functions do not get compiled when
1295 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
1296 the self-tests were still using deprecated key-generation functions so
1297 these have been updated also.
1300 *) Reorganise PKCS#7 code to separate the digest location functionality
1301 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
1302 New function PKCS7_set_digest() to set the digest type for PKCS#7
1303 digestedData type. Add additional code to correctly generate the
1304 digestedData type and add support for this type in PKCS7 initialization
1308 *) New function PKCS7_set0_type_other() this initializes a PKCS7
1309 structure of type "other".
1312 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
1313 sure the loop does correctly stop and breaking ("division by zero")
1314 modulus operations are not performed. The (pre-generated) prime
1315 table crypto/bn/bn_prime.h was already correct, but it could not be
1316 re-generated on some platforms because of the "division by zero"
1317 situation in the script.
1318 [Ralf S. Engelschall]
1320 *) Update support for ECC-based TLS ciphersuites according to
1321 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
1322 SHA-1 now is only used for "small" curves (where the
1323 representation of a field element takes up to 24 bytes); for
1324 larger curves, the field element resulting from ECDH is directly
1325 used as premaster secret.
1326 [Douglas Stebila (Sun Microsystems Laboratories)]
1328 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
1329 curve secp160r1 to the tests.
1330 [Douglas Stebila (Sun Microsystems Laboratories)]
1332 *) Add the possibility to load symbols globally with DSO.
1333 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
1335 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
1336 control of the error stack.
1339 *) Add support for STORE in ENGINE.
1342 *) Add the STORE type. The intention is to provide a common interface
1343 to certificate and key stores, be they simple file-based stores, or
1344 HSM-type store, or LDAP stores, or...
1345 NOTE: The code is currently UNTESTED and isn't really used anywhere.
1348 *) Add a generic structure called OPENSSL_ITEM. This can be used to
1349 pass a list of arguments to any function as well as provide a way
1350 for a function to pass data back to the caller.
1353 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
1354 works like BUF_strdup() but can be used to duplicate a portion of
1355 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
1359 *) Add the function sk_find_ex() which works like sk_find(), but will
1360 return an index to an element even if an exact match couldn't be
1361 found. The index is guaranteed to point at the element where the
1362 searched-for key would be inserted to preserve sorting order.
1365 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
1366 takes an extra flags argument for optional functionality. Currently,
1367 the following flags are defined:
1369 OBJ_BSEARCH_VALUE_ON_NOMATCH
1370 This one gets OBJ_bsearch_ex() to return a pointer to the first
1371 element where the comparing function returns a negative or zero
1374 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
1375 This one gets OBJ_bsearch_ex() to return a pointer to the first
1376 element where the comparing function returns zero. This is useful
1377 if there are more than one element where the comparing function
1381 *) Make it possible to create self-signed certificates with 'openssl ca'
1382 in such a way that the self-signed certificate becomes part of the
1383 CA database and uses the same mechanisms for serial number generation
1384 as all other certificate signing. The new flag '-selfsign' enables
1385 this functionality. Adapt CA.sh and CA.pl.in.
1388 *) Add functionality to check the public key of a certificate request
1389 against a given private. This is useful to check that a certificate
1390 request can be signed by that key (self-signing).
1393 *) Make it possible to have multiple active certificates with the same
1394 subject in the CA index file. This is done only if the keyword
1395 'unique_subject' is set to 'no' in the main CA section (default
1396 if 'CA_default') of the configuration file. The value is saved
1397 with the database itself in a separate index attribute file,
1398 named like the index file with '.attr' appended to the name.
1401 *) Generate muti valued AVAs using '+' notation in config files for
1405 *) Support for nameConstraints certificate extension.
1408 *) Support for policyConstraints certificate extension.
1411 *) Support for policyMappings certificate extension.
1414 *) Make sure the default DSA_METHOD implementation only uses its
1415 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
1416 and change its own handlers to be NULL so as to remove unnecessary
1417 indirection. This lets alternative implementations fallback to the
1418 default implementation more easily.
1421 *) Support for directoryName in GeneralName related extensions
1425 *) Make it possible to link applications using Makefile.shared.
1426 Make that possible even when linking against static libraries!
1429 *) Support for single pass processing for S/MIME signing. This now
1430 means that S/MIME signing can be done from a pipe, in addition
1431 cleartext signing (multipart/signed type) is effectively streaming
1432 and the signed data does not need to be all held in memory.
1434 This is done with a new flag PKCS7_STREAM. When this flag is set
1435 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
1436 is done after the data is output (and digests calculated) in
1437 SMIME_write_PKCS7().
1440 *) Add full support for -rpath/-R, both in shared libraries and
1441 applications, at least on the platforms where it's known how
1445 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
1446 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
1447 will now compute a table of multiples of the generator that
1448 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
1449 faster (notably in the case of a single point multiplication,
1450 scalar * generator).
1451 [Nils Larsch, Bodo Moeller]
1453 *) IPv6 support for certificate extensions. The various extensions
1454 which use the IP:a.b.c.d can now take IPv6 addresses using the
1455 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
1459 *) Added an ENGINE that implements RSA by performing private key
1460 exponentiations with the GMP library. The conversions to and from
1461 GMP's mpz_t format aren't optimised nor are any montgomery forms
1462 cached, and on x86 it appears OpenSSL's own performance has caught up.
1463 However there are likely to be other architectures where GMP could
1464 provide a boost. This ENGINE is not built in by default, but it can be
1465 specified at Configure time and should be accompanied by the necessary
1466 linker additions, eg;
1467 ./config -DOPENSSL_USE_GMP -lgmp
1470 *) "openssl engine" will not display ENGINE/DSO load failure errors when
1471 testing availability of engines with "-t" - the old behaviour is
1472 produced by increasing the feature's verbosity with "-tt".
1475 *) ECDSA routines: under certain error conditions uninitialized BN objects
1476 could be freed. Solution: make sure initialization is performed early
1477 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
1481 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
1482 and DH_METHOD (eg. by ENGINE implementations) to override the normal
1483 software implementations. For DSA and DH, parameter generation can
1484 also be overriden by providing the appropriate method callbacks.
1487 *) Change the "progress" mechanism used in key-generation and
1488 primality testing to functions that take a new BN_GENCB pointer in
1489 place of callback/argument pairs. The new API functions have "_ex"
1490 postfixes and the older functions are reimplemented as wrappers for
1491 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
1492 declarations of the old functions to help (graceful) attempts to
1493 migrate to the new functions. Also, the new key-generation API
1494 functions operate on a caller-supplied key-structure and return
1495 success/failure rather than returning a key or NULL - this is to
1496 help make "keygen" another member function of RSA_METHOD etc.
1498 Example for using the new callback interface:
1500 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
1504 BN_GENCB_set(&my_cb, my_callback, my_arg);
1506 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
1507 /* For the meaning of a, b in calls to my_callback(), see the
1508 * documentation of the function that calls the callback.
1509 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
1510 * my_callback should return 1 if it wants BN_is_prime_ex()
1511 * to continue, or 0 to stop.
1516 *) Change the ZLIB compression method to be stateful, and make it
1517 available to TLS with the number defined in
1518 draft-ietf-tls-compression-04.txt.
1521 *) Add the ASN.1 structures and functions for CertificatePair, which
1522 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
1524 CertificatePair ::= SEQUENCE {
1525 forward [0] Certificate OPTIONAL,
1526 reverse [1] Certificate OPTIONAL,
1527 -- at least one of the pair shall be present -- }
1529 Also implement the PEM functions to read and write certificate
1530 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
1532 This needed to be defined, mostly for the sake of the LDAP
1533 attribute crossCertificatePair, but may prove useful elsewhere as
1537 *) Make it possible to inhibit symlinking of shared libraries in
1538 Makefile.shared, for Cygwin's sake.
1541 *) Extend the BIGNUM API by creating a function
1542 void BN_set_negative(BIGNUM *a, int neg);
1543 and a macro that behave like
1544 int BN_is_negative(const BIGNUM *a);
1546 to avoid the need to access 'a->neg' directly in applications.
1549 *) Implement fast modular reduction for pseudo-Mersenne primes
1550 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
1551 EC_GROUP_new_curve_GFp() will now automatically use this
1553 [Nils Larsch <nla@trustcenter.de>]
1555 *) Add new lock type (CRYPTO_LOCK_BN).
1558 *) Change the ENGINE framework to automatically load engines
1559 dynamically from specific directories unless they could be
1560 found to already be built in or loaded. Move all the
1561 current engines except for the cryptodev one to a new
1563 The engines in engines/ are built as shared libraries if
1564 the "shared" options was given to ./Configure or ./config.
1565 Otherwise, they are inserted in libcrypto.a.
1566 /usr/local/ssl/engines is the default directory for dynamic
1567 engines, but that can be overriden at configure time through
1568 the usual use of --prefix and/or --openssldir, and at run
1569 time with the environment variable OPENSSL_ENGINES.
1570 [Geoff Thorpe and Richard Levitte]
1572 *) Add Makefile.shared, a helper makefile to build shared
1573 libraries. Addapt Makefile.org.
1576 *) Add version info to Win32 DLLs.
1577 [Peter 'Luna' Runestig" <peter@runestig.com>]
1579 *) Add new 'medium level' PKCS#12 API. Certificates and keys
1580 can be added using this API to created arbitrary PKCS#12
1581 files while avoiding the low level API.
1583 New options to PKCS12_create(), key or cert can be NULL and
1584 will then be omitted from the output file. The encryption
1585 algorithm NIDs can be set to -1 for no encryption, the mac
1586 iteration count can be set to 0 to omit the mac.
1588 Enhance pkcs12 utility by making the -nokeys and -nocerts
1589 options work when creating a PKCS#12 file. New option -nomac
1590 to omit the mac, NONE can be set for an encryption algorithm.
1591 New code is modified to use the enhanced PKCS12_create()
1592 instead of the low level API.
1595 *) Extend ASN1 encoder to support indefinite length constructed
1596 encoding. This can output sequences tags and octet strings in
1597 this form. Modify pk7_asn1.c to support indefinite length
1598 encoding. This is experimental and needs additional code to
1599 be useful, such as an ASN1 bio and some enhanced streaming
1602 Extend template encode functionality so that tagging is passed
1603 down to the template encoder.
1606 *) Let 'openssl req' fail if an argument to '-newkey' is not
1607 recognized instead of using RSA as a default.
1610 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
1611 As these are not official, they are not included in "ALL";
1612 the "ECCdraft" ciphersuite group alias can be used to select them.
1613 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
1615 *) Add ECDH engine support.
1616 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
1618 *) Add ECDH in new directory crypto/ecdh/.
1619 [Douglas Stebila (Sun Microsystems Laboratories)]
1621 *) Let BN_rand_range() abort with an error after 100 iterations
1622 without success (which indicates a broken PRNG).
1625 *) Change BN_mod_sqrt() so that it verifies that the input value
1626 is really the square of the return value. (Previously,
1627 BN_mod_sqrt would show GIGO behaviour.)
1630 *) Add named elliptic curves over binary fields from X9.62, SECG,
1631 and WAP/WTLS; add OIDs that were still missing.
1633 [Sheueling Chang Shantz and Douglas Stebila
1634 (Sun Microsystems Laboratories)]
1636 *) Extend the EC library for elliptic curves over binary fields
1637 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
1640 EC_GF2m_simple_method
1644 EC_GROUP_new_curve_GF2m
1645 EC_GROUP_set_curve_GF2m
1646 EC_GROUP_get_curve_GF2m
1647 EC_POINT_set_affine_coordinates_GF2m
1648 EC_POINT_get_affine_coordinates_GF2m
1649 EC_POINT_set_compressed_coordinates_GF2m
1651 Point compression for binary fields is disabled by default for
1652 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
1655 As binary polynomials are represented as BIGNUMs, various members
1656 of the EC_GROUP and EC_POINT data structures can be shared
1657 between the implementations for prime fields and binary fields;
1658 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
1659 are essentially identical to their ..._GFp counterparts.
1660 (For simplicity, the '..._GFp' prefix has been dropped from
1661 various internal method names.)
1663 An internal 'field_div' method (similar to 'field_mul' and
1664 'field_sqr') has been added; this is used only for binary fields.
1666 [Sheueling Chang Shantz and Douglas Stebila
1667 (Sun Microsystems Laboratories)]
1669 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
1670 through methods ('mul', 'precompute_mult').
1672 The generic implementations (now internally called 'ec_wNAF_mul'
1673 and 'ec_wNAF_precomputed_mult') remain the default if these
1674 methods are undefined.
1676 [Sheueling Chang Shantz and Douglas Stebila
1677 (Sun Microsystems Laboratories)]
1679 *) New function EC_GROUP_get_degree, which is defined through
1680 EC_METHOD. For curves over prime fields, this returns the bit
1681 length of the modulus.
1683 [Sheueling Chang Shantz and Douglas Stebila
1684 (Sun Microsystems Laboratories)]
1686 *) New functions EC_GROUP_dup, EC_POINT_dup.
1687 (These simply call ..._new and ..._copy).
1689 [Sheueling Chang Shantz and Douglas Stebila
1690 (Sun Microsystems Laboratories)]
1692 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
1693 Polynomials are represented as BIGNUMs (where the sign bit is not
1694 used) in the following functions [macros]:
1697 BN_GF2m_sub [= BN_GF2m_add]
1698 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
1699 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
1700 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
1702 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
1703 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
1704 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
1705 BN_GF2m_cmp [= BN_ucmp]
1707 (Note that only the 'mod' functions are actually for fields GF(2^m).
1708 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
1710 For some functions, an the irreducible polynomial defining a
1711 field can be given as an 'unsigned int[]' with strictly
1712 decreasing elements giving the indices of those bits that are set;
1713 i.e., p[] represents the polynomial
1714 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
1716 p[0] > p[1] > ... > p[k] = 0.
1717 This applies to the following functions:
1722 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
1723 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
1725 BN_GF2m_mod_sqrt_arr
1726 BN_GF2m_mod_solve_quad_arr
1730 Conversion can be performed by the following functions:
1735 bntest.c has additional tests for binary polynomial arithmetic.
1737 Two implementations for BN_GF2m_mod_div() are available.
1738 The default algorithm simply uses BN_GF2m_mod_inv() and
1739 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
1740 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
1741 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
1743 [Sheueling Chang Shantz and Douglas Stebila
1744 (Sun Microsystems Laboratories)]
1746 *) Add new error code 'ERR_R_DISABLED' that can be used when some
1747 functionality is disabled at compile-time.
1748 [Douglas Stebila <douglas.stebila@sun.com>]
1750 *) Change default behaviour of 'openssl asn1parse' so that more
1751 information is visible when viewing, e.g., a certificate:
1753 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
1754 mode the content of non-printable OCTET STRINGs is output in a
1755 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
1756 avoid the appearance of a printable string.
1757 [Nils Larsch <nla@trustcenter.de>]
1759 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
1761 EC_GROUP_set_asn1_flag()
1762 EC_GROUP_get_asn1_flag()
1763 EC_GROUP_set_point_conversion_form()
1764 EC_GROUP_get_point_conversion_form()
1765 These control ASN1 encoding details:
1766 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
1767 has been set to OPENSSL_EC_NAMED_CURVE.
1768 - Points are encoded in uncompressed form by default; options for
1769 asn1_for are as for point2oct, namely
1770 POINT_CONVERSION_COMPRESSED
1771 POINT_CONVERSION_UNCOMPRESSED
1772 POINT_CONVERSION_HYBRID
1774 Also add 'seed' and 'seed_len' members to EC_GROUP with access
1777 EC_GROUP_get0_seed()
1778 EC_GROUP_get_seed_len()
1779 This is used only for ASN1 purposes (so far).
1780 [Nils Larsch <nla@trustcenter.de>]
1782 *) Add 'field_type' member to EC_METHOD, which holds the NID
1783 of the appropriate field type OID. The new function
1784 EC_METHOD_get_field_type() returns this value.
1785 [Nils Larsch <nla@trustcenter.de>]
1790 EC_POINT_point2hex()
1791 EC_POINT_hex2point()
1792 providing useful interfaces to EC_POINT_point2oct() and
1793 EC_POINT_oct2point().
1794 [Nils Larsch <nla@trustcenter.de>]
1796 *) Change internals of the EC library so that the functions
1797 EC_GROUP_set_generator()
1798 EC_GROUP_get_generator()
1799 EC_GROUP_get_order()
1800 EC_GROUP_get_cofactor()
1801 are implemented directly in crypto/ec/ec_lib.c and not dispatched
1802 to methods, which would lead to unnecessary code duplication when
1803 adding different types of curves.
1804 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
1806 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
1807 arithmetic, and such that modified wNAFs are generated
1808 (which avoid length expansion in many cases).
1811 *) Add a function EC_GROUP_check_discriminant() (defined via
1812 EC_METHOD) that verifies that the curve discriminant is non-zero.
1814 Add a function EC_GROUP_check() that makes some sanity tests
1815 on a EC_GROUP, its generator and order. This includes
1816 EC_GROUP_check_discriminant().
1817 [Nils Larsch <nla@trustcenter.de>]
1819 *) Add ECDSA in new directory crypto/ecdsa/.
1821 Add applications 'openssl ecparam' and 'openssl ecdsa'
1822 (these are based on 'openssl dsaparam' and 'openssl dsa').
1824 ECDSA support is also included in various other files across the
1825 library. Most notably,
1826 - 'openssl req' now has a '-newkey ecdsa:file' option;
1827 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
1828 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
1829 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
1830 them suitable for ECDSA where domain parameters must be
1831 extracted before the specific public key;
1832 - ECDSA engine support has been added.
1833 [Nils Larsch <nla@trustcenter.de>]
1835 *) Include some named elliptic curves, and add OIDs from X9.62,
1836 SECG, and WAP/WTLS. Each curve can be obtained from the new
1838 EC_GROUP_new_by_curve_name(),
1839 and the list of available named curves can be obtained with
1840 EC_get_builtin_curves().
1841 Also add a 'curve_name' member to EC_GROUP objects, which can be
1843 EC_GROUP_set_curve_name()
1844 EC_GROUP_get_curve_name()
1845 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
1847 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1848 was actually never needed) and in BN_mul(). The removal in BN_mul()
1849 required a small change in bn_mul_part_recursive() and the addition
1850 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1851 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1852 bn_sub_words() and bn_add_words() except they take arrays with
1856 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
1858 *) Cleanse PEM buffers before freeing them since they may contain
1860 [Benjamin Bennett <ben@psc.edu>]
1862 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1863 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1864 authentication-only ciphersuites.
1867 *) Since AES128 and AES256 share a single mask bit in the logic of
1868 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1869 kludge to work properly if AES128 is available and AES256 isn't.
1872 *) Expand security boundary to match 1.1.1 module.
1875 *) Remove redundant features: hash file source, editing of test vectors
1876 modify fipsld to use external fips_premain.c signature.
1879 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
1880 run algorithm test programs.
1883 *) Make algorithm test programs more tolerant of whitespace.
1886 *) Have SSL/TLS server implementation tolerate "mismatched" record
1887 protocol version while receiving ClientHello even if the
1888 ClientHello is fragmented. (The server can't insist on the
1889 particular protocol version it has chosen before the ServerHello
1890 message has informed the client about his choice.)
1893 *) Load error codes if they are not already present instead of using a
1894 static variable. This allows them to be cleanly unloaded and reloaded.
1897 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
1899 *) Introduce limits to prevent malicious keys being able to
1900 cause a denial of service. (CVE-2006-2940)
1901 [Steve Henson, Bodo Moeller]
1903 *) Fix ASN.1 parsing of certain invalid structures that can result
1904 in a denial of service. (CVE-2006-2937) [Steve Henson]
1906 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1907 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1909 *) Fix SSL client code which could crash if connecting to a
1910 malicious SSLv2 server. (CVE-2006-4343)
1911 [Tavis Ormandy and Will Drewry, Google Security Team]
1913 *) Change ciphersuite string processing so that an explicit
1914 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
1915 will no longer include "AES128-SHA"), and any other similar
1916 ciphersuite (same bitmap) from *other* protocol versions (so that
1917 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
1918 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
1919 changes from 0.9.8b and 0.9.8d.
1922 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
1924 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1925 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1927 *) Change the Unix randomness entropy gathering to use poll() when
1928 possible instead of select(), since the latter has some
1929 undesirable limitations.
1930 [Darryl Miles via Richard Levitte and Bodo Moeller]
1932 *) Disable rogue ciphersuites:
1934 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1935 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1936 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1938 The latter two were purportedly from
1939 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1942 Also deactive the remaining ciphersuites from
1943 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1944 unofficial, and the ID has long expired.
1947 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1948 dual-core machines) and other potential thread-safety issues.
1951 Changes between 0.9.7i and 0.9.7j [04 May 2006]
1953 *) Adapt fipsld and the build system to link against the validated FIPS
1954 module in FIPS mode.
1957 *) Fixes for VC++ 2005 build under Windows.
1960 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
1961 from a Windows bash shell such as MSYS. It is autodetected from the
1962 "config" script when run from a VC++ environment. Modify standard VC++
1963 build to use fipscanister.o from the GNU make build.
1966 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
1968 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
1969 The value now differs depending on if you build for FIPS or not.
1970 BEWARE! A program linked with a shared FIPSed libcrypto can't be
1971 safely run with a non-FIPSed libcrypto, as it may crash because of
1972 the difference induced by this change.
1975 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
1977 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1978 (part of SSL_OP_ALL). This option used to disable the
1979 countermeasure against man-in-the-middle protocol-version
1980 rollback in the SSL 2.0 server implementation, which is a bad
1981 idea. (CVE-2005-2969)
1983 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1984 for Information Security, National Institute of Advanced Industrial
1985 Science and Technology [AIST], Japan)]
1987 *) Minimal support for X9.31 signatures and PSS padding modes. This is
1988 mainly for FIPS compliance and not fully integrated at this stage.
1991 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
1992 the exponentiation using a fixed-length exponent. (Otherwise,
1993 the information leaked through timing could expose the secret key
1994 after many signatures; cf. Bleichenbacher's attack on DSA with
1998 *) Make a new fixed-window mod_exp implementation the default for
1999 RSA, DSA, and DH private-key operations so that the sequence of
2000 squares and multiplies and the memory access pattern are
2001 independent of the particular secret key. This will mitigate
2002 cache-timing and potential related attacks.
2004 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
2005 and this is automatically used by BN_mod_exp_mont() if the new flag
2006 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
2007 will use this BN flag for private exponents unless the flag
2008 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
2009 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
2011 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
2013 *) Change the client implementation for SSLv23_method() and
2014 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
2015 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
2016 (Previously, the SSL 2.0 backwards compatible Client Hello
2017 message format would be used even with SSL_OP_NO_SSLv2.)
2020 *) Add support for smime-type MIME parameter in S/MIME messages which some
2024 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
2025 a threadsafe manner. Modify rsa code to use new function and add calls
2026 to dsa and dh code (which had race conditions before).
2029 *) Include the fixed error library code in the C error file definitions
2030 instead of fixing them up at runtime. This keeps the error code
2031 structures constant.
2034 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
2036 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
2039 *) Fixes for newer kerberos headers. NB: the casts are needed because
2040 the 'length' field is signed on one version and unsigned on another
2041 with no (?) obvious way to tell the difference, without these VC++
2042 complains. Also the "definition" of FAR (blank) is no longer included
2043 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
2044 some needed definitions.
2047 *) Undo Cygwin change.
2050 *) Added support for proxy certificates according to RFC 3820.
2051 Because they may be a security thread to unaware applications,
2052 they must be explicitely allowed in run-time. See
2053 docs/HOWTO/proxy_certificates.txt for further information.
2056 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
2058 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
2059 server and client random values. Previously
2060 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
2061 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
2063 This change has negligible security impact because:
2065 1. Server and client random values still have 24 bytes of pseudo random
2068 2. Server and client random values are sent in the clear in the initial
2071 3. The master secret is derived using the premaster secret (48 bytes in
2072 size for static RSA ciphersuites) as well as client server and random
2075 The OpenSSL team would like to thank the UK NISCC for bringing this issue
2078 [Stephen Henson, reported by UK NISCC]
2080 *) Use Windows randomness collection on Cygwin.
2083 *) Fix hang in EGD/PRNGD query when communication socket is closed
2084 prematurely by EGD/PRNGD.
2085 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
2087 *) Prompt for pass phrases when appropriate for PKCS12 input format.
2090 *) Back-port of selected performance improvements from development
2091 branch, as well as improved support for PowerPC platforms.
2094 *) Add lots of checks for memory allocation failure, error codes to indicate
2095 failure and freeing up memory if a failure occurs.
2096 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
2098 *) Add new -passin argument to dgst.
2101 *) Perform some character comparisons of different types in X509_NAME_cmp:
2102 this is needed for some certificates that reencode DNs into UTF8Strings
2103 (in violation of RFC3280) and can't or wont issue name rollover
2107 *) Make an explicit check during certificate validation to see that
2108 the CA setting in each certificate on the chain is correct. As a
2109 side effect always do the following basic checks on extensions,
2110 not just when there's an associated purpose to the check:
2112 - if there is an unhandled critical extension (unless the user
2113 has chosen to ignore this fault)
2114 - if the path length has been exceeded (if one is set at all)
2115 - that certain extensions fit the associated purpose (if one has
2119 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
2121 *) Avoid a race condition when CRLs are checked in a multi threaded
2122 environment. This would happen due to the reordering of the revoked
2123 entries during signature checking and serial number lookup. Now the
2124 encoding is cached and the serial number sort performed under a lock.
2125 Add new STACK function sk_is_sorted().
2128 *) Add Delta CRL to the extension code.
2131 *) Various fixes to s3_pkt.c so alerts are sent properly.
2132 [David Holmes <d.holmes@f5.com>]
2134 *) Reduce the chances of duplicate issuer name and serial numbers (in
2135 violation of RFC3280) using the OpenSSL certificate creation utilities.
2136 This is done by creating a random 64 bit value for the initial serial
2137 number when a serial number file is created or when a self signed
2138 certificate is created using 'openssl req -x509'. The initial serial
2139 number file is created using 'openssl x509 -next_serial' in CA.pl
2140 rather than being initialized to 1.
2143 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
2145 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
2146 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
2147 [Joe Orton, Steve Henson]
2149 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
2151 [Joe Orton, Steve Henson]
2153 *) Make it possible to have multiple active certificates with the same
2154 subject in the CA index file. This is done only if the keyword
2155 'unique_subject' is set to 'no' in the main CA section (default
2156 if 'CA_default') of the configuration file. The value is saved
2157 with the database itself in a separate index attribute file,
2158 named like the index file with '.attr' appended to the name.
2161 *) X509 verify fixes. Disable broken certificate workarounds when
2162 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
2163 keyUsage extension present. Don't accept CRLs with unhandled critical
2164 extensions: since verify currently doesn't process CRL extensions this
2165 rejects a CRL with *any* critical extensions. Add new verify error codes
2169 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
2170 A clarification of RFC2560 will require the use of OCTET STRINGs and
2171 some implementations cannot handle the current raw format. Since OpenSSL
2172 copies and compares OCSP nonces as opaque blobs without any attempt at
2173 parsing them this should not create any compatibility issues.
2176 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
2177 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
2178 this HMAC (and other) operations are several times slower than OpenSSL
2182 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
2183 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2185 *) Use the correct content when signing type "other".
2188 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
2190 *) Fix various bugs revealed by running the NISCC test suite:
2192 Stop out of bounds reads in the ASN1 code when presented with
2193 invalid tags (CVE-2003-0543 and CVE-2003-0544).
2195 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
2197 If verify callback ignores invalid public key errors don't try to check
2198 certificate signature with the NULL public key.
2202 *) New -ignore_err option in ocsp application to stop the server
2203 exiting on the first error in a request.
2206 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
2207 if the server requested one: as stated in TLS 1.0 and SSL 3.0
2211 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
2212 extra data after the compression methods not only for TLS 1.0
2213 but also for SSL 3.0 (as required by the specification).
2214 [Bodo Moeller; problem pointed out by Matthias Loepfe]
2216 *) Change X509_certificate_type() to mark the key as exported/exportable
2217 when it's 512 *bits* long, not 512 bytes.
2220 *) Change AES_cbc_encrypt() so it outputs exact multiple of
2221 blocks during encryption.
2224 *) Various fixes to base64 BIO and non blocking I/O. On write
2225 flushes were not handled properly if the BIO retried. On read
2226 data was not being buffered properly and had various logic bugs.
2227 This also affects blocking I/O when the data being decoded is a
2231 *) Various S/MIME bugfixes and compatibility changes:
2232 output correct application/pkcs7 MIME type if
2233 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
2234 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
2235 of files as .eml work). Correctly handle very long lines in MIME
2239 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
2241 *) Countermeasure against the Klima-Pokorny-Rosa extension of
2242 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
2243 a protocol version number mismatch like a decryption error
2244 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
2247 *) Turn on RSA blinding by default in the default implementation
2248 to avoid a timing attack. Applications that don't want it can call
2249 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
2250 They would be ill-advised to do so in most cases.
2251 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
2253 *) Change RSA blinding code so that it works when the PRNG is not
2254 seeded (in this case, the secret RSA exponent is abused as
2255 an unpredictable seed -- if it is not unpredictable, there
2256 is no point in blinding anyway). Make RSA blinding thread-safe
2257 by remembering the creator's thread ID in rsa->blinding and
2258 having all other threads use local one-time blinding factors
2259 (this requires more computation than sharing rsa->blinding, but
2260 avoids excessive locking; and if an RSA object is not shared
2261 between threads, blinding will still be very fast).
2264 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
2265 ENGINE as defaults for all supported algorithms irrespective of
2266 the 'flags' parameter. 'flags' is now honoured, so applications
2267 should make sure they are passing it correctly.
2270 *) Target "mingw" now allows native Windows code to be generated in
2271 the Cygwin environment as well as with the MinGW compiler.
2274 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
2276 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
2277 via timing by performing a MAC computation even if incorrrect
2278 block cipher padding has been found. This is a countermeasure
2279 against active attacks where the attacker has to distinguish
2280 between bad padding and a MAC verification error. (CVE-2003-0078)
2282 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
2283 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
2284 Martin Vuagnoux (EPFL, Ilion)]
2286 *) Make the no-err option work as intended. The intention with no-err
2287 is not to have the whole error stack handling routines removed from
2288 libcrypto, it's only intended to remove all the function name and
2289 reason texts, thereby removing some of the footprint that may not
2290 be interesting if those errors aren't displayed anyway.
2292 NOTE: it's still possible for any application or module to have it's
2293 own set of error texts inserted. The routines are there, just not
2294 used by default when no-err is given.
2297 *) Add support for FreeBSD on IA64.
2298 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
2300 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
2301 Kerberos function mit_des_cbc_cksum(). Before this change,
2302 the value returned by DES_cbc_cksum() was like the one from
2303 mit_des_cbc_cksum(), except the bytes were swapped.
2304 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
2306 *) Allow an application to disable the automatic SSL chain building.
2307 Before this a rather primitive chain build was always performed in
2308 ssl3_output_cert_chain(): an application had no way to send the
2309 correct chain if the automatic operation produced an incorrect result.
2311 Now the chain builder is disabled if either:
2313 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
2315 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
2317 The reasoning behind this is that an application would not want the
2318 auto chain building to take place if extra chain certificates are
2319 present and it might also want a means of sending no additional
2320 certificates (for example the chain has two certificates and the
2324 *) Add the possibility to build without the ENGINE framework.
2325 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2327 *) Under Win32 gmtime() can return NULL: check return value in
2328 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
2331 *) DSA routines: under certain error conditions uninitialized BN objects
2332 could be freed. Solution: make sure initialization is performed early
2333 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
2334 Nils Larsch <nla@trustcenter.de> via PR#459)
2337 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
2338 checked on reconnect on the client side, therefore session resumption
2339 could still fail with a "ssl session id is different" error. This
2340 behaviour is masked when SSL_OP_ALL is used due to
2341 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
2342 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2343 followup to PR #377.
2346 *) IA-32 assembler support enhancements: unified ELF targets, support
2347 for SCO/Caldera platforms, fix for Cygwin shared build.
2350 *) Add support for FreeBSD on sparc64. As a consequence, support for
2351 FreeBSD on non-x86 processors is separate from x86 processors on
2352 the config script, much like the NetBSD support.
2353 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
2355 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
2357 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
2360 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
2361 code (06) was taken as the first octet of the session ID and the last
2362 octet was ignored consequently. As a result SSLv2 client side session
2363 caching could not have worked due to the session ID mismatch between
2365 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2369 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
2370 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
2374 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
2375 seems that in spite of existing for more than a year, many application
2376 author have done nothing to provide the necessary callbacks, which
2377 means that this particular engine will not work properly anywhere.
2378 This is a very unfortunate situation which forces us, in the name
2379 of usability, to give the hw_ncipher.c a static lock, which is part
2381 NOTE: This is for the 0.9.7 series ONLY. This hack will never
2382 appear in 0.9.8 or later. We EXPECT application authors to have
2383 dealt properly with this when 0.9.8 is released (unless we actually
2384 make such changes in the libcrypto locking code that changes will
2385 have to be made anyway).
2388 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
2389 octets have been read, EOF or an error occurs. Without this change
2390 some truncated ASN1 structures will not produce an error.
2393 *) Disable Heimdal support, since it hasn't been fully implemented.
2394 Still give the possibility to force the use of Heimdal, but with
2395 warnings and a request that patches get sent to openssl-dev.
2398 *) Add the VC-CE target, introduce the WINCE sysname, and add
2399 INSTALL.WCE and appropriate conditionals to make it build.
2400 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2402 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
2403 cygssl-x.y.z.dll, where x, y and z are the major, minor and
2404 edit numbers of the version.
2405 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2407 *) Introduce safe string copy and catenation functions
2408 (BUF_strlcpy() and BUF_strlcat()).
2409 [Ben Laurie (CHATS) and Richard Levitte]
2411 *) Avoid using fixed-size buffers for one-line DNs.
2412 [Ben Laurie (CHATS)]
2414 *) Add BUF_MEM_grow_clean() to avoid information leakage when
2415 resizing buffers containing secrets, and use where appropriate.
2416 [Ben Laurie (CHATS)]
2418 *) Avoid using fixed size buffers for configuration file location.
2419 [Ben Laurie (CHATS)]
2421 *) Avoid filename truncation for various CA files.
2422 [Ben Laurie (CHATS)]
2424 *) Use sizeof in preference to magic numbers.
2425 [Ben Laurie (CHATS)]
2427 *) Avoid filename truncation in cert requests.
2428 [Ben Laurie (CHATS)]
2430 *) Add assertions to check for (supposedly impossible) buffer
2432 [Ben Laurie (CHATS)]
2434 *) Don't cache truncated DNS entries in the local cache (this could
2435 potentially lead to a spoofing attack).
2436 [Ben Laurie (CHATS)]
2438 *) Fix various buffers to be large enough for hex/decimal
2439 representations in a platform independent manner.
2440 [Ben Laurie (CHATS)]
2442 *) Add CRYPTO_realloc_clean() to avoid information leakage when
2443 resizing buffers containing secrets, and use where appropriate.
2444 [Ben Laurie (CHATS)]
2446 *) Add BIO_indent() to avoid much slightly worrying code to do
2448 [Ben Laurie (CHATS)]
2450 *) Convert sprintf()/BIO_puts() to BIO_printf().
2451 [Ben Laurie (CHATS)]
2453 *) buffer_gets() could terminate with the buffer only half
2455 [Ben Laurie (CHATS)]
2457 *) Add assertions to prevent user-supplied crypto functions from
2458 overflowing internal buffers by having large block sizes, etc.
2459 [Ben Laurie (CHATS)]
2461 *) New OPENSSL_assert() macro (similar to assert(), but enabled
2463 [Ben Laurie (CHATS)]
2465 *) Eliminate unused copy of key in RC4.
2466 [Ben Laurie (CHATS)]
2468 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
2469 [Ben Laurie (CHATS)]
2471 *) Fix off-by-one error in EGD path.
2472 [Ben Laurie (CHATS)]
2474 *) If RANDFILE path is too long, ignore instead of truncating.
2475 [Ben Laurie (CHATS)]
2477 *) Eliminate unused and incorrectly sized X.509 structure
2479 [Ben Laurie (CHATS)]
2481 *) Eliminate unused and dangerous function knumber().
2482 [Ben Laurie (CHATS)]
2484 *) Eliminate unused and dangerous structure, KSSL_ERR.
2485 [Ben Laurie (CHATS)]
2487 *) Protect against overlong session ID context length in an encoded
2488 session object. Since these are local, this does not appear to be
2490 [Ben Laurie (CHATS)]
2492 *) Change from security patch (see 0.9.6e below) that did not affect
2493 the 0.9.6 release series:
2495 Remote buffer overflow in SSL3 protocol - an attacker could
2496 supply an oversized master key in Kerberos-enabled versions.
2498 [Ben Laurie (CHATS)]
2500 *) Change the SSL kerb5 codes to match RFC 2712.
2503 *) Make -nameopt work fully for req and add -reqopt switch.
2504 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
2506 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
2507 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
2509 *) Make sure tests can be performed even if the corresponding algorithms
2510 have been removed entirely. This was also the last step to make
2511 OpenSSL compilable with DJGPP under all reasonable conditions.
2512 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
2514 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
2515 to allow version independent disabling of normally unselected ciphers,
2516 which may be activated as a side-effect of selecting a single cipher.
2518 (E.g., cipher list string "RSA" enables ciphersuites that are left
2519 out of "ALL" because they do not provide symmetric encryption.
2520 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
2521 [Lutz Jaenicke, Bodo Moeller]
2523 *) Add appropriate support for separate platform-dependent build
2524 directories. The recommended way to make a platform-dependent
2525 build directory is the following (tested on Linux), maybe with
2528 # Place yourself outside of the OpenSSL source tree. In
2529 # this example, the environment variable OPENSSL_SOURCE
2530 # is assumed to contain the absolute OpenSSL source directory.
2531 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
2532 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
2533 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
2534 mkdir -p `dirname $F`
2535 ln -s $OPENSSL_SOURCE/$F $F
2538 To be absolutely sure not to disturb the source tree, a "make clean"
2539 is a good thing. If it isn't successfull, don't worry about it,
2540 it probably means the source directory is very clean.
2543 *) Make sure any ENGINE control commands make local copies of string
2544 pointers passed to them whenever necessary. Otherwise it is possible
2545 the caller may have overwritten (or deallocated) the original string
2546 data when a later ENGINE operation tries to use the stored values.
2547 [Götz Babin-Ebell <babinebell@trustcenter.de>]
2549 *) Improve diagnostics in file reading and command-line digests.
2550 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
2552 *) Add AES modes CFB and OFB to the object database. Correct an
2553 error in AES-CFB decryption.
2556 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
2557 allows existing EVP_CIPHER_CTX structures to be reused after
2558 calling EVP_*Final(). This behaviour is used by encryption
2559 BIOs and some applications. This has the side effect that
2560 applications must explicitly clean up cipher contexts with
2561 EVP_CIPHER_CTX_cleanup() or they will leak memory.
2564 *) Check the values of dna and dnb in bn_mul_recursive before calling
2565 bn_mul_comba (a non zero value means the a or b arrays do not contain
2566 n2 elements) and fallback to bn_mul_normal if either is not zero.
2569 *) Fix escaping of non-ASCII characters when using the -subj option
2570 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
2573 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
2574 form for "surname", serialNumber has no short form.
2575 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
2576 therefore remove "mail" short name for "internet 7".
2577 The OID for unique identifiers in X509 certificates is
2578 x500UniqueIdentifier, not uniqueIdentifier.
2579 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
2582 *) Add an "init" command to the ENGINE config module and auto initialize
2583 ENGINEs. Without any "init" command the ENGINE will be initialized
2584 after all ctrl commands have been executed on it. If init=1 the
2585 ENGINE is initailized at that point (ctrls before that point are run
2586 on the uninitialized ENGINE and after on the initialized one). If
2587 init=0 then the ENGINE will not be iniatialized at all.
2590 *) Fix the 'app_verify_callback' interface so that the user-defined
2591 argument is actually passed to the callback: In the
2592 SSL_CTX_set_cert_verify_callback() prototype, the callback
2593 declaration has been changed from
2596 int (*cb)(X509_STORE_CTX *,void *);
2597 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
2598 i=s->ctx->app_verify_callback(&ctx)
2599 has been changed into
2600 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
2602 To update applications using SSL_CTX_set_cert_verify_callback(),
2603 a dummy argument can be added to their callback functions.
2604 [D. K. Smetters <smetters@parc.xerox.com>]
2606 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
2607 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
2609 *) Add and OPENSSL_LOAD_CONF define which will cause
2610 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
2611 This allows older applications to transparently support certain
2612 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
2613 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
2614 load the config file and OPENSSL_add_all_algorithms_conf() which will
2615 always load it have also been added.
2618 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
2619 Adjust NIDs and EVP layer.
2620 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2622 *) Config modules support in openssl utility.
2624 Most commands now load modules from the config file,
2625 though in a few (such as version) this isn't done
2626 because it couldn't be used for anything.
2628 In the case of ca and req the config file used is
2629 the same as the utility itself: that is the -config
2630 command line option can be used to specify an
2634 *) Move default behaviour from OPENSSL_config(). If appname is NULL
2635 use "openssl_conf" if filename is NULL use default openssl config file.
2638 *) Add an argument to OPENSSL_config() to allow the use of an alternative
2639 config section name. Add a new flag to tolerate a missing config file
2640 and move code to CONF_modules_load_file().
2643 *) Support for crypto accelerator cards from Accelerated Encryption
2644 Processing, www.aep.ie. (Use engine 'aep')
2645 The support was copied from 0.9.6c [engine] and adapted/corrected
2646 to work with the new engine framework.
2647 [AEP Inc. and Richard Levitte]
2649 *) Support for SureWare crypto accelerator cards from Baltimore
2650 Technologies. (Use engine 'sureware')
2651 The support was copied from 0.9.6c [engine] and adapted
2652 to work with the new engine framework.
2655 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
2656 make the newer ENGINE framework commands for the CHIL engine work.
2657 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
2659 *) Make it possible to produce shared libraries on ReliantUNIX.
2660 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
2662 *) Add the configuration target debug-linux-ppro.
2663 Make 'openssl rsa' use the general key loading routines
2664 implemented in apps.c, and make those routines able to
2665 handle the key format FORMAT_NETSCAPE and the variant
2667 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2669 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2670 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2672 *) Add -keyform to rsautl, and document -engine.
2673 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
2675 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
2676 BIO_R_NO_SUCH_FILE error code rather than the generic
2677 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
2680 *) Add new functions
2682 ERR_peek_last_error_line
2683 ERR_peek_last_error_line_data.
2684 These are similar to
2687 ERR_peek_error_line_data,
2688 but report on the latest error recorded rather than the first one
2689 still in the error queue.
2690 [Ben Laurie, Bodo Moeller]
2692 *) default_algorithms option in ENGINE config module. This allows things
2694 default_algorithms = ALL
2695 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
2698 *) Prelminary ENGINE config module.
2701 *) New experimental application configuration code.
2704 *) Change the AES code to follow the same name structure as all other
2705 symmetric ciphers, and behave the same way. Move everything to
2706 the directory crypto/aes, thereby obsoleting crypto/rijndael.
2707 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2709 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
2710 [Ben Laurie and Theo de Raadt]
2712 *) Add option to output public keys in req command.
2713 [Massimiliano Pala madwolf@openca.org]
2715 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
2716 (up to about 10% better than before for P-192 and P-224).
2719 *) New functions/macros
2721 SSL_CTX_set_msg_callback(ctx, cb)
2722 SSL_CTX_set_msg_callback_arg(ctx, arg)
2723 SSL_set_msg_callback(ssl, cb)
2724 SSL_set_msg_callback_arg(ssl, arg)
2726 to request calling a callback function
2728 void cb(int write_p, int version, int content_type,
2729 const void *buf, size_t len, SSL *ssl, void *arg)
2731 whenever a protocol message has been completely received
2732 (write_p == 0) or sent (write_p == 1). Here 'version' is the
2733 protocol version according to which the SSL library interprets
2734 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
2735 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
2736 the content type as defined in the SSL 3.0/TLS 1.0 protocol
2737 specification (change_cipher_spec(20), alert(21), handshake(22)).
2738 'buf' and 'len' point to the actual message, 'ssl' to the
2739 SSL object, and 'arg' is the application-defined value set by
2740 SSL[_CTX]_set_msg_callback_arg().
2742 'openssl s_client' and 'openssl s_server' have new '-msg' options
2743 to enable a callback that displays all protocol messages.
2746 *) Change the shared library support so shared libraries are built as
2747 soon as the corresponding static library is finished, and thereby get
2748 openssl and the test programs linked against the shared library.
2749 This still only happens when the keyword "shard" has been given to
2750 the configuration scripts.
2752 NOTE: shared library support is still an experimental thing, and
2753 backward binary compatibility is still not guaranteed.
2754 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
2756 *) Add support for Subject Information Access extension.
2757 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2759 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
2760 additional bytes when new memory had to be allocated, not just
2761 when reusing an existing buffer.
2764 *) New command line and configuration option 'utf8' for the req command.
2765 This allows field values to be specified as UTF8 strings.
2768 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
2769 runs for the former and machine-readable output for the latter.
2772 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
2773 of the e-mail address in the DN (i.e., it will go into a certificate
2774 extension only). The new configuration file option 'email_in_dn = no'
2775 has the same effect.
2776 [Massimiliano Pala madwolf@openca.org]
2778 *) Change all functions with names starting with des_ to be starting
2779 with DES_ instead. Add wrappers that are compatible with libdes,
2780 but are named _ossl_old_des_*. Finally, add macros that map the
2781 des_* symbols to the corresponding _ossl_old_des_* if libdes
2782 compatibility is desired. If OpenSSL 0.9.6c compatibility is
2783 desired, the des_* symbols will be mapped to DES_*, with one
2786 Since we provide two compatibility mappings, the user needs to
2787 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
2788 compatibility is desired. The default (i.e., when that macro
2789 isn't defined) is OpenSSL 0.9.6c compatibility.
2791 There are also macros that enable and disable the support of old
2792 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
2793 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
2794 are defined, the default will apply: to support the old des routines.
2796 In either case, one must include openssl/des.h to get the correct
2797 definitions. Do not try to just include openssl/des_old.h, that
2800 NOTE: This is a major break of an old API into a new one. Software
2801 authors are encouraged to switch to the DES_ style functions. Some
2802 time in the future, des_old.h and the libdes compatibility functions
2803 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
2804 default), and then completely removed.
2807 *) Test for certificates which contain unsupported critical extensions.
2808 If such a certificate is found during a verify operation it is
2809 rejected by default: this behaviour can be overridden by either
2810 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
2811 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
2812 X509_supported_extension() has also been added which returns 1 if a
2813 particular extension is supported.
2816 *) Modify the behaviour of EVP cipher functions in similar way to digests
2817 to retain compatibility with existing code.
2820 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
2821 compatibility with existing code. In particular the 'ctx' parameter does
2822 not have to be to be initialized before the call to EVP_DigestInit() and
2823 it is tidied up after a call to EVP_DigestFinal(). New function
2824 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
2825 EVP_MD_CTX_copy() changed to not require the destination to be
2826 initialized valid and new function EVP_MD_CTX_copy_ex() added which
2827 requires the destination to be valid.
2829 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
2830 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
2833 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
2834 so that complete 'Handshake' protocol structures are kept in memory
2835 instead of overwriting 'msg_type' and 'length' with 'body' data.
2838 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
2839 [Massimo Santin via Richard Levitte]
2841 *) Major restructuring to the underlying ENGINE code. This includes
2842 reduction of linker bloat, separation of pure "ENGINE" manipulation
2843 (initialisation, etc) from functionality dealing with implementations
2844 of specific crypto iterfaces. This change also introduces integrated
2845 support for symmetric ciphers and digest implementations - so ENGINEs
2846 can now accelerate these by providing EVP_CIPHER and EVP_MD
2847 implementations of their own. This is detailed in crypto/engine/README
2848 as it couldn't be adequately described here. However, there are a few
2849 API changes worth noting - some RSA, DSA, DH, and RAND functions that
2850 were changed in the original introduction of ENGINE code have now
2851 reverted back - the hooking from this code to ENGINE is now a good
2852 deal more passive and at run-time, operations deal directly with
2853 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
2854 dereferencing through an ENGINE pointer any more. Also, the ENGINE
2855 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
2856 they were not being used by the framework as there is no concept of a
2857 BIGNUM_METHOD and they could not be generalised to the new
2858 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
2859 ENGINE_cpy() has been removed as it cannot be consistently defined in
2863 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
2866 *) Change mkdef.pl to sort symbols that get the same entry number,
2867 and make sure the automatically generated functions ERR_load_*
2868 become part of libeay.num as well.
2871 *) New function SSL_renegotiate_pending(). This returns true once
2872 renegotiation has been requested (either SSL_renegotiate() call
2873 or HelloRequest/ClientHello receveived from the peer) and becomes
2874 false once a handshake has been completed.
2875 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
2876 sends a HelloRequest, but does not ensure that a handshake takes
2877 place. SSL_renegotiate_pending() is useful for checking if the
2878 client has followed the request.)
2881 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
2882 By default, clients may request session resumption even during
2883 renegotiation (if session ID contexts permit); with this option,
2884 session resumption is possible only in the first handshake.
2886 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
2887 more bits available for options that should not be part of
2888 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
2891 *) Add some demos for certificate and certificate request creation.
2894 *) Make maximum certificate chain size accepted from the peer application
2895 settable (SSL*_get/set_max_cert_list()), as proposed by
2896 "Douglas E. Engert" <deengert@anl.gov>.
2899 *) Add support for shared libraries for Unixware-7
2900 (Boyd Lynn Gerber <gerberb@zenez.com>).
2903 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
2904 be done prior to destruction. Use this to unload error strings from
2905 ENGINEs that load their own error strings. NB: This adds two new API
2906 functions to "get" and "set" this destroy handler in an ENGINE.
2909 *) Alter all existing ENGINE implementations (except "openssl" and
2910 "openbsd") to dynamically instantiate their own error strings. This
2911 makes them more flexible to be built both as statically-linked ENGINEs
2912 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
2913 Also, add stub code to each that makes building them as self-contained
2914 shared-libraries easier (see README.ENGINE).
2917 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
2918 implementations into applications that are completely implemented in
2919 self-contained shared-libraries. The "dynamic" ENGINE exposes control
2920 commands that can be used to configure what shared-library to load and
2921 to control aspects of the way it is handled. Also, made an update to
2922 the README.ENGINE file that brings its information up-to-date and
2923 provides some information and instructions on the "dynamic" ENGINE
2924 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
2927 *) Make it possible to unload ranges of ERR strings with a new
2928 "ERR_unload_strings" function.
2931 *) Add a copy() function to EVP_MD.
2934 *) Make EVP_MD routines take a context pointer instead of just the
2935 md_data void pointer.
2938 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
2939 that the digest can only process a single chunk of data
2940 (typically because it is provided by a piece of
2941 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
2942 is only going to provide a single chunk of data, and hence the
2943 framework needn't accumulate the data for oneshot drivers.
2946 *) As with "ERR", make it possible to replace the underlying "ex_data"
2947 functions. This change also alters the storage and management of global
2948 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
2949 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
2950 index counters. The API functions that use this state have been changed
2951 to take a "class_index" rather than pointers to the class's local STACK
2952 and counter, and there is now an API function to dynamically create new
2953 classes. This centralisation allows us to (a) plug a lot of the
2954 thread-safety problems that existed, and (b) makes it possible to clean
2955 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
2956 such data would previously have always leaked in application code and
2957 workarounds were in place to make the memory debugging turn a blind eye
2958 to it. Application code that doesn't use this new function will still
2959 leak as before, but their memory debugging output will announce it now
2960 rather than letting it slide.
2962 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
2963 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
2964 has a return value to indicate success or failure.
2967 *) Make it possible to replace the underlying "ERR" functions such that the
2968 global state (2 LHASH tables and 2 locks) is only used by the "default"
2969 implementation. This change also adds two functions to "get" and "set"
2970 the implementation prior to it being automatically set the first time
2971 any other ERR function takes place. Ie. an application can call "get",
2972 pass the return value to a module it has just loaded, and that module
2973 can call its own "set" function using that value. This means the
2974 module's "ERR" operations will use (and modify) the error state in the
2975 application and not in its own statically linked copy of OpenSSL code.
2978 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
2979 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
2980 the operation, and provides a more encapsulated way for external code
2981 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
2982 to use these functions rather than manually incrementing the counts.
2984 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
2987 *) Add EVP test program.
2990 *) Add symmetric cipher support to ENGINE. Expect the API to change!
2993 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
2994 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
2995 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
2996 These allow a CRL to be built without having to access X509_CRL fields
2997 directly. Modify 'ca' application to use new functions.
3000 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
3001 bug workarounds. Rollback attack detection is a security feature.
3002 The problem will only arise on OpenSSL servers when TLSv1 is not
3003 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
3004 Software authors not wanting to support TLSv1 will have special reasons
3005 for their choice and can explicitly enable this option.
3006 [Bodo Moeller, Lutz Jaenicke]
3008 *) Rationalise EVP so it can be extended: don't include a union of
3009 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
3010 (similar to those existing for EVP_CIPHER_CTX).
3015 EVP_MD_CTX_init(&md); /* new function call */
3016 EVP_DigestInit(&md, EVP_sha1());
3017 EVP_DigestUpdate(&md, in, len);
3018 EVP_DigestFinal(&md, out, NULL);
3019 EVP_MD_CTX_cleanup(&md); /* new function call */
3023 *) Make DES key schedule conform to the usual scheme, as well as
3024 correcting its structure. This means that calls to DES functions
3025 now have to pass a pointer to a des_key_schedule instead of a
3026 plain des_key_schedule (which was actually always a pointer
3029 des_key_schedule ks;
3031 des_set_key_checked(..., &ks);
3032 des_ncbc_encrypt(..., &ks, ...);
3034 (Note that a later change renames 'des_...' into 'DES_...'.)
3037 *) Initial reduction of linker bloat: the use of some functions, such as
3038 PEM causes large amounts of unused functions to be linked in due to
3039 poor organisation. For example pem_all.c contains every PEM function
3040 which has a knock on effect of linking in large amounts of (unused)
3041 ASN1 code. Grouping together similar functions and splitting unrelated
3042 functions prevents this.
3045 *) Cleanup of EVP macros.
3048 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
3049 correct _ecb suffix.
3052 *) Add initial OCSP responder support to ocsp application. The
3053 revocation information is handled using the text based index
3054 use by the ca application. The responder can either handle
3055 requests generated internally, supplied in files (for example
3056 via a CGI script) or using an internal minimal server.
3059 *) Add configuration choices to get zlib compression for TLS.
3062 *) Changes to Kerberos SSL for RFC 2712 compliance:
3063 1. Implemented real KerberosWrapper, instead of just using
3064 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
3065 2. Implemented optional authenticator field of KerberosWrapper.
3067 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
3068 and authenticator structs; see crypto/krb5/.
3070 Generalized Kerberos calls to support multiple Kerberos libraries.
3071 [Vern Staats <staatsvr@asc.hpc.mil>,
3072 Jeffrey Altman <jaltman@columbia.edu>
3073 via Richard Levitte]
3075 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
3076 already does with RSA. testdsa.h now has 'priv_key/pub_key'
3077 values for each of the key sizes rather than having just
3078 parameters (and 'speed' generating keys each time).
3081 *) Speed up EVP routines.
3084 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
3085 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
3086 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
3087 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
3089 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
3090 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
3091 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
3094 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
3096 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
3099 *) Added the OS2-EMX target.
3100 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
3102 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
3103 to support NCONF routines in extension code. New function CONF_set_nconf()
3104 to allow functions which take an NCONF to also handle the old LHASH
3105 structure: this means that the old CONF compatible routines can be
3106 retained (in particular wrt extensions) without having to duplicate the
3107 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
3110 *) Enhance the general user interface with mechanisms for inner control
3111 and with possibilities to have yes/no kind of prompts.
3114 *) Change all calls to low level digest routines in the library and
3115 applications to use EVP. Add missing calls to HMAC_cleanup() and
3116 don't assume HMAC_CTX can be copied using memcpy().
3117 [Verdon Walker <VWalker@novell.com>, Steve Henson]
3119 *) Add the possibility to control engines through control names but with
3120 arbitrary arguments instead of just a string.
3121 Change the key loaders to take a UI_METHOD instead of a callback
3122 function pointer. NOTE: this breaks binary compatibility with earlier
3123 versions of OpenSSL [engine].
3124 Adapt the nCipher code for these new conditions and add a card insertion
3128 *) Enhance the general user interface with mechanisms to better support
3129 dialog box interfaces, application-defined prompts, the possibility
3130 to use defaults (for example default passwords from somewhere else)
3131 and interrupts/cancellations.
3134 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
3135 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
3138 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
3139 tidy up some unnecessarily weird code in 'sk_new()').
3140 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
3142 *) Change the key loading routines for ENGINEs to use the same kind
3143 callback (pem_password_cb) as all other routines that need this
3147 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
3148 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
3149 than this minimum value is recommended.
3152 *) New random seeder for OpenVMS, using the system process statistics
3153 that are easily reachable.
3156 *) Windows apparently can't transparently handle global
3157 variables defined in DLLs. Initialisations such as:
3159 const ASN1_ITEM *it = &ASN1_INTEGER_it;
3161 wont compile. This is used by the any applications that need to
3162 declare their own ASN1 modules. This was fixed by adding the option
3163 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
3164 needed for static libraries under Win32.
3167 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
3168 setting of purpose and trust fields. New X509_STORE trust and
3169 purpose functions and tidy up setting in other SSL functions.
3172 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
3173 structure. These are inherited by X509_STORE_CTX when it is
3174 initialised. This allows various defaults to be set in the
3175 X509_STORE structure (such as flags for CRL checking and custom
3176 purpose or trust settings) for functions which only use X509_STORE_CTX
3177 internally such as S/MIME.
3179 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
3180 trust settings if they are not set in X509_STORE. This allows X509_STORE
3181 purposes and trust (in S/MIME for example) to override any set by default.
3183 Add command line options for CRL checking to smime, s_client and s_server
3187 *) Initial CRL based revocation checking. If the CRL checking flag(s)
3188 are set then the CRL is looked up in the X509_STORE structure and
3189 its validity and signature checked, then if the certificate is found
3190 in the CRL the verify fails with a revoked error.
3192 Various new CRL related callbacks added to X509_STORE_CTX structure.
3194 Command line options added to 'verify' application to support this.
3196 This needs some additional work, such as being able to handle multiple
3197 CRLs with different times, extension based lookup (rather than just
3198 by subject name) and ultimately more complete V2 CRL extension
3202 *) Add a general user interface API (crypto/ui/). This is designed
3203 to replace things like des_read_password and friends (backward
3204 compatibility functions using this new API are provided).
3205 The purpose is to remove prompting functions from the DES code
3206 section as well as provide for prompting through dialog boxes in
3207 a window system and the like.
3210 *) Add "ex_data" support to ENGINE so implementations can add state at a
3211 per-structure level rather than having to store it globally.
3214 *) Make it possible for ENGINE structures to be copied when retrieved by
3215 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
3216 This causes the "original" ENGINE structure to act like a template,
3217 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
3218 operational state can be localised to each ENGINE structure, despite the
3219 fact they all share the same "methods". New ENGINE structures returned in
3220 this case have no functional references and the return value is the single
3221 structural reference. This matches the single structural reference returned
3222 by ENGINE_by_id() normally, when it is incremented on the pre-existing
3226 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
3227 needs to match any other type at all we need to manually clear the
3231 *) Changes to the "openssl engine" utility to include;
3232 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
3233 about an ENGINE's available control commands.
3234 - executing control commands from command line arguments using the
3235 '-pre' and '-post' switches. '-post' is only used if '-t' is
3236 specified and the ENGINE is successfully initialised. The syntax for
3237 the individual commands are colon-separated, for example;
3238 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
3241 *) New dynamic control command support for ENGINEs. ENGINEs can now
3242 declare their own commands (numbers), names (strings), descriptions,
3243 and input types for run-time discovery by calling applications. A
3244 subset of these commands are implicitly classed as "executable"
3245 depending on their input type, and only these can be invoked through
3246 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
3247 can be based on user input, config files, etc). The distinction is
3248 that "executable" commands cannot return anything other than a boolean
3249 result and can only support numeric or string input, whereas some
3250 discoverable commands may only be for direct use through
3251 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
3252 pointers, or other custom uses. The "executable" commands are to
3253 support parameterisations of ENGINE behaviour that can be
3254 unambiguously defined by ENGINEs and used consistently across any
3255 OpenSSL-based application. Commands have been added to all the
3256 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
3257 control over shared-library paths without source code alterations.
3260 *) Changed all ENGINE implementations to dynamically allocate their
3261 ENGINEs rather than declaring them statically. Apart from this being
3262 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
3263 this also allows the implementations to compile without using the
3264 internal engine_int.h header.
3267 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
3268 'const' value. Any code that should be able to modify a RAND_METHOD
3269 should already have non-const pointers to it (ie. they should only
3270 modify their own ones).
3273 *) Made a variety of little tweaks to the ENGINE code.
3274 - "atalla" and "ubsec" string definitions were moved from header files
3275 to C code. "nuron" string definitions were placed in variables
3276 rather than hard-coded - allowing parameterisation of these values
3277 later on via ctrl() commands.
3278 - Removed unused "#if 0"'d code.
3279 - Fixed engine list iteration code so it uses ENGINE_free() to release
3280 structural references.
3281 - Constified the RAND_METHOD element of ENGINE structures.
3282 - Constified various get/set functions as appropriate and added
3283 missing functions (including a catch-all ENGINE_cpy that duplicates
3284 all ENGINE values onto a new ENGINE except reference counts/state).
3285 - Removed NULL parameter checks in get/set functions. Setting a method
3286 or function to NULL is a way of cancelling out a previously set
3287 value. Passing a NULL ENGINE parameter is just plain stupid anyway
3288 and doesn't justify the extra error symbols and code.
3289 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
3290 flags from engine_int.h to engine.h.
3291 - Changed prototypes for ENGINE handler functions (init(), finish(),
3292 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
3295 *) Implement binary inversion algorithm for BN_mod_inverse in addition
3296 to the algorithm using long division. The binary algorithm can be
3297 used only if the modulus is odd. On 32-bit systems, it is faster
3298 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
3299 roughly 5-15% for 256-bit moduli), so we use it only for moduli
3300 up to 450 bits. In 64-bit environments, the binary algorithm
3301 appears to be advantageous for much longer moduli; here we use it
3302 for moduli up to 2048 bits.
3305 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
3306 could not support the combine flag in choice fields.
3309 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
3310 extensions from a certificate request to the certificate.
3313 *) Allow multiple 'certopt' and 'nameopt' options to be separated
3314 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
3315 file: this allows the display of the certificate about to be
3316 signed to be customised, to allow certain fields to be included
3317 or excluded and extension details. The old system didn't display
3318 multicharacter strings properly, omitted fields not in the policy
3319 and couldn't display additional details such as extensions.
3322 *) Function EC_POINTs_mul for multiple scalar multiplication
3323 of an arbitrary number of elliptic curve points
3324 \sum scalars[i]*points[i],
3325 optionally including the generator defined for the EC_GROUP:
3326 scalar*generator + \sum scalars[i]*points[i].
3328 EC_POINT_mul is a simple wrapper function for the typical case
3329 that the point list has just one item (besides the optional
3333 *) First EC_METHODs for curves over GF(p):
3335 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
3336 operations and provides various method functions that can also
3337 operate with faster implementations of modular arithmetic.
3339 EC_GFp_mont_method() reuses most functions that are part of
3340 EC_GFp_simple_method, but uses Montgomery arithmetic.
3342 [Bodo Moeller; point addition and point doubling
3343 implementation directly derived from source code provided by
3344 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
3346 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
3347 crypto/ec/ec_lib.c):
3349 Curves are EC_GROUP objects (with an optional group generator)
3350 based on EC_METHODs that are built into the library.
3352 Points are EC_POINT objects based on EC_GROUP objects.
3354 Most of the framework would be able to handle curves over arbitrary
3355 finite fields, but as there are no obvious types for fields other
3356 than GF(p), some functions are limited to that for now.
3359 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
3360 that the file contains a complete HTTP response.
3363 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
3364 change the def and num file printf format specifier from "%-40sXXX"
3365 to "%-39s XXX". The latter will always guarantee a space after the
3366 field while the former will cause them to run together if the field
3367 is 40 of more characters long.
3370 *) Constify the cipher and digest 'method' functions and structures
3371 and modify related functions to take constant EVP_MD and EVP_CIPHER
3375 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
3376 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
3379 *) Modify EVP_Digest*() routines so they now return values. Although the
3380 internal software routines can never fail additional hardware versions
3384 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
3386 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
3387 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
3390 ERR_R_NESTED_ASN1_ERROR
3392 ERR_R_MISSING_ASN1_EOS
3393 were 4 .. 9, conflicting with
3394 ERR_LIB_RSA (= ERR_R_RSA_LIB)
3396 ERR_LIB_PEM (= ERR_R_PEM_LIB).
3397 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
3399 Add new error code 'ERR_R_INTERNAL_ERROR'.
3402 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
3406 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
3407 sets the subject name for a new request or supersedes the
3408 subject name in a given request. Formats that can be parsed are
3409 'CN=Some Name, OU=myOU, C=IT'
3411 'CN=Some Name/OU=myOU/C=IT'.
3413 Add options '-batch' and '-verbose' to 'openssl req'.
3414 [Massimiliano Pala <madwolf@hackmasters.net>]
3416 *) Introduce the possibility to access global variables through
3417 functions on platform were that's the best way to handle exporting
3418 global variables in shared libraries. To enable this functionality,
3419 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
3420 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
3421 is normally done by Configure or something similar).
3423 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
3424 in the source file (foo.c) like this:
3426 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
3427 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
3429 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
3430 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
3432 OPENSSL_DECLARE_GLOBAL(int,foo);
3433 #define foo OPENSSL_GLOBAL_REF(foo)
3434 OPENSSL_DECLARE_GLOBAL(double,bar);
3435 #define bar OPENSSL_GLOBAL_REF(bar)
3437 The #defines are very important, and therefore so is including the
3438 header file everywhere where the defined globals are used.
3440 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
3441 of ASN.1 items, but that structure is a bit different.
3443 The largest change is in util/mkdef.pl which has been enhanced with
3444 better and easier to understand logic to choose which symbols should
3445 go into the Windows .def files as well as a number of fixes and code
3446 cleanup (among others, algorithm keywords are now sorted
3447 lexicographically to avoid constant rewrites).
3450 *) In BN_div() keep a copy of the sign of 'num' before writing the
3451 result to 'rm' because if rm==num the value will be overwritten
3452 and produce the wrong result if 'num' is negative: this caused
3453 problems with BN_mod() and BN_nnmod().
3456 *) Function OCSP_request_verify(). This checks the signature on an
3457 OCSP request and verifies the signer certificate. The signer
3458 certificate is just checked for a generic purpose and OCSP request
3462 *) Add OCSP_check_validity() function to check the validity of OCSP
3463 responses. OCSP responses are prepared in real time and may only
3464 be a few seconds old. Simply checking that the current time lies
3465 between thisUpdate and nextUpdate max reject otherwise valid responses
3466 caused by either OCSP responder or client clock inaccuracy. Instead
3467 we allow thisUpdate and nextUpdate to fall within a certain period of
3468 the current time. The age of the response can also optionally be
3469 checked. Two new options -validity_period and -status_age added to
3473 *) If signature or public key algorithm is unrecognized print out its
3474 OID rather that just UNKNOWN.
3477 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
3478 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
3479 ID to be generated from the issuer certificate alone which can then be
3480 passed to OCSP_id_issuer_cmp().
3483 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
3484 ASN1 modules to export functions returning ASN1_ITEM pointers
3485 instead of the ASN1_ITEM structures themselves. This adds several
3486 new macros which allow the underlying ASN1 function/structure to
3487 be accessed transparently. As a result code should not use ASN1_ITEM
3488 references directly (such as &X509_it) but instead use the relevant
3489 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
3490 use of the new ASN1 code on platforms where exporting structures
3491 is problematical (for example in shared libraries) but exporting
3492 functions returning pointers to structures is not.
3495 *) Add support for overriding the generation of SSL/TLS session IDs.
3496 These callbacks can be registered either in an SSL_CTX or per SSL.
3497 The purpose of this is to allow applications to control, if they wish,
3498 the arbitrary values chosen for use as session IDs, particularly as it
3499 can be useful for session caching in multiple-server environments. A
3500 command-line switch for testing this (and any client code that wishes
3501 to use such a feature) has been added to "s_server".
3502 [Geoff Thorpe, Lutz Jaenicke]
3504 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
3505 of the form '#if defined(...) || defined(...) || ...' and
3506 '#if !defined(...) && !defined(...) && ...'. This also avoids
3507 the growing number of special cases it was previously handling.
3510 *) Make all configuration macros available for application by making
3511 sure they are available in opensslconf.h, by giving them names starting
3512 with "OPENSSL_" to avoid conflicts with other packages and by making
3513 sure e_os2.h will cover all platform-specific cases together with
3515 Additionally, it is now possible to define configuration/platform-
3516 specific names (called "system identities"). In the C code, these
3517 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
3518 macro with the name beginning with "OPENSSL_SYS_", which is determined
3519 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
3523 *) New option -set_serial to 'req' and 'x509' this allows the serial
3524 number to use to be specified on the command line. Previously self
3525 signed certificates were hard coded with serial number 0 and the
3526 CA options of 'x509' had to use a serial number in a file which was
3530 *) New options to 'ca' utility to support V2 CRL entry extensions.
3531 Currently CRL reason, invalidity date and hold instruction are
3532 supported. Add new CRL extensions to V3 code and some new objects.
3535 *) New function EVP_CIPHER_CTX_set_padding() this is used to
3536 disable standard block padding (aka PKCS#5 padding) in the EVP
3537 API, which was previously mandatory. This means that the data is
3538 not padded in any way and so the total length much be a multiple
3539 of the block size, otherwise an error occurs.
3542 *) Initial (incomplete) OCSP SSL support.
3545 *) New function OCSP_parse_url(). This splits up a URL into its host,
3546 port and path components: primarily to parse OCSP URLs. New -url
3547 option to ocsp utility.
3550 *) New nonce behavior. The return value of OCSP_check_nonce() now
3551 reflects the various checks performed. Applications can decide
3552 whether to tolerate certain situations such as an absent nonce
3553 in a response when one was present in a request: the ocsp application
3554 just prints out a warning. New function OCSP_add1_basic_nonce()
3555 this is to allow responders to include a nonce in a response even if
3556 the request is nonce-less.
3559 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
3560 skipped when using openssl x509 multiple times on a single input file,
3561 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
3564 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
3565 set string type: to handle setting ASN1_TIME structures. Fix ca
3566 utility to correctly initialize revocation date of CRLs.
3569 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
3570 the clients preferred ciphersuites and rather use its own preferences.
3571 Should help to work around M$ SGC (Server Gated Cryptography) bug in
3572 Internet Explorer by ensuring unchanged hash method during stepup.
3573 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
3576 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
3577 to aes and add a new 'exist' option to print out symbols that don't
3581 *) Additional options to ocsp utility to allow flags to be set and
3582 additional certificates supplied.
3585 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
3586 OCSP client a number of certificate to only verify the response
3590 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
3591 handle the new API. Currently only ECB, CBC modes supported. Add new
3594 Add TLS AES ciphersuites as described in RFC3268, "Advanced
3595 Encryption Standard (AES) Ciphersuites for Transport Layer
3596 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
3597 not enabled by default and were not part of the "ALL" ciphersuite
3598 alias because they were not yet official; they could be
3599 explicitly requested by specifying the "AESdraft" ciphersuite
3600 group alias. In the final release of OpenSSL 0.9.7, the group
3601 alias is called "AES" and is part of "ALL".)
3602 [Ben Laurie, Steve Henson, Bodo Moeller]
3604 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
3605 request to response.
3608 *) Functions for OCSP responders. OCSP_request_onereq_count(),
3609 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
3610 extract information from a certificate request. OCSP_response_create()
3611 creates a response and optionally adds a basic response structure.
3612 OCSP_basic_add1_status() adds a complete single response to a basic
3613 response and returns the OCSP_SINGLERESP structure just added (to allow
3614 extensions to be included for example). OCSP_basic_add1_cert() adds a
3615 certificate to a basic response and OCSP_basic_sign() signs a basic
3616 response with various flags. New helper functions ASN1_TIME_check()
3617 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
3618 (converts ASN1_TIME to GeneralizedTime).
3621 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
3622 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
3623 structure from a certificate. X509_pubkey_digest() digests the public_key
3624 contents: this is used in various key identifiers.
3627 *) Make sk_sort() tolerate a NULL argument.
3628 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
3630 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
3631 passed by the function are trusted implicitly. If any of them signed the
3632 response then it is assumed to be valid and is not verified.
3635 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
3636 to data. This was previously part of the PKCS7 ASN1 code. This
3637 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
3638 [Steve Henson, reported by Kenneth R. Robinette
3639 <support@securenetterm.com>]
3641 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
3642 routines: without these tracing memory leaks is very painful.
3643 Fix leaks in PKCS12 and PKCS7 routines.
3646 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
3647 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
3648 effectively meant GeneralizedTime would never be used. Now it
3649 is initialised to -1 but X509_time_adj() now has to check the value
3650 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
3651 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
3652 [Steve Henson, reported by Kenneth R. Robinette
3653 <support@securenetterm.com>]
3655 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
3656 result in a zero length in the ASN1_INTEGER structure which was
3657 not consistent with the structure when d2i_ASN1_INTEGER() was used
3658 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
3659 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
3660 where it did not print out a minus for negative ASN1_INTEGER.
3663 *) Add summary printout to ocsp utility. The various functions which
3664 convert status values to strings have been renamed to:
3665 OCSP_response_status_str(), OCSP_cert_status_str() and
3666 OCSP_crl_reason_str() and are no longer static. New options
3667 to verify nonce values and to disable verification. OCSP response
3668 printout format cleaned up.
3671 *) Add additional OCSP certificate checks. These are those specified
3672 in RFC2560. This consists of two separate checks: the CA of the
3673 certificate being checked must either be the OCSP signer certificate
3674 or the issuer of the OCSP signer certificate. In the latter case the
3675 OCSP signer certificate must contain the OCSP signing extended key
3676 usage. This check is performed by attempting to match the OCSP
3677 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
3678 in the OCSP_CERTID structures of the response.
3681 *) Initial OCSP certificate verification added to OCSP_basic_verify()
3682 and related routines. This uses the standard OpenSSL certificate
3683 verify routines to perform initial checks (just CA validity) and
3684 to obtain the certificate chain. Then additional checks will be
3685 performed on the chain. Currently the root CA is checked to see
3686 if it is explicitly trusted for OCSP signing. This is used to set
3687 a root CA as a global signing root: that is any certificate that
3688 chains to that CA is an acceptable OCSP signing certificate.
3691 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
3692 extensions from a separate configuration file.
3693 As when reading extensions from the main configuration file,
3694 the '-extensions ...' option may be used for specifying the
3696 [Massimiliano Pala <madwolf@comune.modena.it>]
3698 *) New OCSP utility. Allows OCSP requests to be generated or
3699 read. The request can be sent to a responder and the output
3700 parsed, outputed or printed in text form. Not complete yet:
3701 still needs to check the OCSP response validity.
3704 *) New subcommands for 'openssl ca':
3705 'openssl ca -status <serial>' prints the status of the cert with
3706 the given serial number (according to the index file).
3707 'openssl ca -updatedb' updates the expiry status of certificates
3709 [Massimiliano Pala <madwolf@comune.modena.it>]
3711 *) New '-newreq-nodes' command option to CA.pl. This is like
3712 '-newreq', but calls 'openssl req' with the '-nodes' option
3713 so that the resulting key is not encrypted.
3714 [Damien Miller <djm@mindrot.org>]
3716 *) New configuration for the GNU Hurd.
3717 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
3719 *) Initial code to implement OCSP basic response verify. This
3720 is currently incomplete. Currently just finds the signer's
3721 certificate and verifies the signature on the response.
3724 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
3725 value of OPENSSLDIR. This is available via the new '-d' option
3726 to 'openssl version', and is also included in 'openssl version -a'.
3729 *) Allowing defining memory allocation callbacks that will be given
3730 file name and line number information in additional arguments
3731 (a const char* and an int). The basic functionality remains, as
3732 well as the original possibility to just replace malloc(),
3733 realloc() and free() by functions that do not know about these
3734 additional arguments. To register and find out the current
3735 settings for extended allocation functions, the following
3736 functions are provided:
3738 CRYPTO_set_mem_ex_functions
3739 CRYPTO_set_locked_mem_ex_functions
3740 CRYPTO_get_mem_ex_functions
3741 CRYPTO_get_locked_mem_ex_functions
3743 These work the same way as CRYPTO_set_mem_functions and friends.
3744 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
3745 extended allocation function is enabled.
3746 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
3747 a conventional allocation function is enabled.
3748 [Richard Levitte, Bodo Moeller]
3750 *) Finish off removing the remaining LHASH function pointer casts.
3751 There should no longer be any prototype-casting required when using
3752 the LHASH abstraction, and any casts that remain are "bugs". See
3753 the callback types and macros at the head of lhash.h for details
3754 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
3757 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
3758 If /dev/[u]random devices are not available or do not return enough
3759 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
3761 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
3762 /etc/entropy will be queried once each in this sequence, quering stops
3763 when enough entropy was collected without querying more sockets.
3766 *) Change the Unix RAND_poll() variant to be able to poll several
3767 random devices, as specified by DEVRANDOM, until a sufficient amount
3768 of data has been collected. We spend at most 10 ms on each file
3769 (select timeout) and read in non-blocking mode. DEVRANDOM now
3770 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
3771 (previously it was just the string "/dev/urandom"), so on typical
3772 platforms the 10 ms delay will never occur.
3773 Also separate out the Unix variant to its own file, rand_unix.c.
3774 For VMS, there's a currently-empty rand_vms.c.
3777 *) Move OCSP client related routines to ocsp_cl.c. These
3778 provide utility functions which an application needing
3779 to issue a request to an OCSP responder and analyse the
3780 response will typically need: as opposed to those which an
3781 OCSP responder itself would need which will be added later.
3783 OCSP_request_sign() signs an OCSP request with an API similar
3784 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
3785 response. OCSP_response_get1_basic() extracts basic response
3786 from response. OCSP_resp_find_status(): finds and extracts status
3787 information from an OCSP_CERTID structure (which will be created
3788 when the request structure is built). These are built from lower
3789 level functions which work on OCSP_SINGLERESP structures but
3790 wont normally be used unless the application wishes to examine
3791 extensions in the OCSP response for example.
3793 Replace nonce routines with a pair of functions.
3794 OCSP_request_add1_nonce() adds a nonce value and optionally
3795 generates a random value. OCSP_check_nonce() checks the
3796 validity of the nonce in an OCSP response.
3799 *) Change function OCSP_request_add() to OCSP_request_add0_id().
3800 This doesn't copy the supplied OCSP_CERTID and avoids the
3801 need to free up the newly created id. Change return type
3802 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
3803 This can then be used to add extensions to the request.
3804 Deleted OCSP_request_new(), since most of its functionality
3805 is now in OCSP_REQUEST_new() (and the case insensitive name
3806 clash) apart from the ability to set the request name which
3807 will be added elsewhere.
3810 *) Update OCSP API. Remove obsolete extensions argument from
3811 various functions. Extensions are now handled using the new
3812 OCSP extension code. New simple OCSP HTTP function which
3813 can be used to send requests and parse the response.
3816 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
3817 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
3818 uses the special reorder version of SET OF to sort the attributes
3819 and reorder them to match the encoded order. This resolves a long
3820 standing problem: a verify on a PKCS7 structure just after signing
3821 it used to fail because the attribute order did not match the
3822 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
3823 it uses the received order. This is necessary to tolerate some broken
3824 software that does not order SET OF. This is handled by encoding
3825 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
3826 to produce the required SET OF.
3829 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
3830 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
3831 files to get correct declarations of the ASN.1 item variables.
3834 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
3835 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
3836 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
3837 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
3838 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
3839 ASN1_ITEM and no wrapper functions.
3842 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
3843 replace the old function pointer based I/O routines. Change most of
3844 the *_d2i_bio() and *_d2i_fp() functions to use these.
3847 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
3848 lines, recognice more "algorithms" that can be deselected, and make
3849 it complain about algorithm deselection that isn't recognised.
3852 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
3853 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
3854 to use new functions. Add NO_ASN1_OLD which can be set to remove
3855 some old style ASN1 functions: this can be used to determine if old
3856 code will still work when these eventually go away.
3859 *) New extension functions for OCSP structures, these follow the
3860 same conventions as certificates and CRLs.
3863 *) New function X509V3_add1_i2d(). This automatically encodes and
3864 adds an extension. Its behaviour can be customised with various
3865 flags to append, replace or delete. Various wrappers added for
3866 certifcates and CRLs.
3869 *) Fix to avoid calling the underlying ASN1 print routine when
3870 an extension cannot be parsed. Correct a typo in the
3871 OCSP_SERVICELOC extension. Tidy up print OCSP format.
3874 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
3875 entries for variables.
3878 *) Add functionality to apps/openssl.c for detecting locking
3879 problems: As the program is single-threaded, all we have
3880 to do is register a locking callback using an array for
3881 storing which locks are currently held by the program.
3884 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
3885 SSL_get_ex_data_X509_STORE_idx(), which is used in
3886 ssl_verify_cert_chain() and thus can be called at any time
3887 during TLS/SSL handshakes so that thread-safety is essential.
3888 Unfortunately, the ex_data design is not at all suited
3889 for multi-threaded use, so it probably should be abolished.
3892 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
3893 [Broadcom, tweaked and integrated by Geoff Thorpe]
3895 *) Move common extension printing code to new function
3896 X509V3_print_extensions(). Reorganise OCSP print routines and
3897 implement some needed OCSP ASN1 functions. Add OCSP extensions.
3900 *) New function X509_signature_print() to remove duplication in some
3904 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
3905 set (this was treated exactly the same as SET OF previously). This
3906 is used to reorder the STACK representing the structure to match the
3907 encoding. This will be used to get round a problem where a PKCS7
3908 structure which was signed could not be verified because the STACK
3909 order did not reflect the encoded order.
3912 *) Reimplement the OCSP ASN1 module using the new code.
3915 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
3916 for its ASN1 operations. The old style function pointers still exist
3917 for now but they will eventually go away.
3920 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
3921 completely replaces the old ASN1 functionality with a table driven
3922 encoder and decoder which interprets an ASN1_ITEM structure describing
3923 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
3924 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
3925 has also been converted to the new form.
3928 *) Change BN_mod_exp_recp so that negative moduli are tolerated
3929 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
3930 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
3931 for negative moduli.
3934 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
3935 of not touching the result's sign bit.
3938 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
3942 *) Changed the LHASH code to use prototypes for callbacks, and created
3943 macros to declare and implement thin (optionally static) functions
3944 that provide type-safety and avoid function pointer casting for the
3945 type-specific callbacks.
3948 *) Added Kerberos Cipher Suites to be used with TLS, as written in
3950 [Veers Staats <staatsvr@asc.hpc.mil>,
3951 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
3953 *) Reformat the FAQ so the different questions and answers can be divided
3954 in sections depending on the subject.
3957 *) Have the zlib compression code load ZLIB.DLL dynamically under
3961 *) New function BN_mod_sqrt for computing square roots modulo a prime
3962 (using the probabilistic Tonelli-Shanks algorithm unless
3963 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
3964 be handled deterministically).
3965 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3967 *) Make BN_mod_inverse faster by explicitly handling small quotients
3968 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
3969 512 bits], about 30% for larger ones [1024 or 2048 bits].)
3972 *) New function BN_kronecker.
3975 *) Fix BN_gcd so that it works on negative inputs; the result is
3976 positive unless both parameters are zero.
3977 Previously something reasonably close to an infinite loop was
3978 possible because numbers could be growing instead of shrinking
3979 in the implementation of Euclid's algorithm.
3982 *) Fix BN_is_word() and BN_is_one() macros to take into account the
3983 sign of the number in question.
3985 Fix BN_is_word(a,w) to work correctly for w == 0.
3987 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
3988 because its test if the absolute value of 'a' equals 'w'.
3989 Note that BN_abs_is_word does *not* handle w == 0 reliably;
3990 it exists mostly for use in the implementations of BN_is_zero(),
3991 BN_is_one(), and BN_is_word().
3994 *) New function BN_swap.
3997 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
3998 the exponentiation functions are more likely to produce reasonable
3999 results on negative inputs.
4002 *) Change BN_mod_mul so that the result is always non-negative.
4003 Previously, it could be negative if one of the factors was negative;
4004 I don't think anyone really wanted that behaviour.
4007 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
4008 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
4009 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
4010 and add new functions:
4019 BN_mod_lshift1_quick
4023 These functions always generate non-negative results.
4025 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
4026 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
4028 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
4029 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
4030 be reduced modulo m.
4031 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
4034 The following entry accidentily appeared in the CHANGES file
4035 distributed with OpenSSL 0.9.7. The modifications described in
4036 it do *not* apply to OpenSSL 0.9.7.
4038 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4039 was actually never needed) and in BN_mul(). The removal in BN_mul()
4040 required a small change in bn_mul_part_recursive() and the addition
4041 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4042 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4043 bn_sub_words() and bn_add_words() except they take arrays with
4048 *) In 'openssl passwd', verify passwords read from the terminal
4049 unless the '-salt' option is used (which usually means that
4050 verification would just waste user's time since the resulting
4051 hash is going to be compared with some given password hash)
4052 or the new '-noverify' option is used.
4054 This is an incompatible change, but it does not affect
4055 non-interactive use of 'openssl passwd' (passwords on the command
4056 line, '-stdin' option, '-in ...' option) and thus should not
4060 *) Remove all references to RSAref, since there's no more need for it.
4063 *) Make DSO load along a path given through an environment variable
4064 (SHLIB_PATH) with shl_load().
4067 *) Constify the ENGINE code as a result of BIGNUM constification.
4068 Also constify the RSA code and most things related to it. In a
4069 few places, most notable in the depth of the ASN.1 code, ugly
4070 casts back to non-const were required (to be solved at a later
4074 *) Make it so the openssl application has all engines loaded by default.
4077 *) Constify the BIGNUM routines a little more.
4080 *) Add the following functions:
4082 ENGINE_load_cswift()
4084 ENGINE_load_atalla()
4086 ENGINE_load_builtin_engines()
4088 That way, an application can itself choose if external engines that
4089 are built-in in OpenSSL shall ever be used or not. The benefit is
4090 that applications won't have to be linked with libdl or other dso
4091 libraries unless it's really needed.
4093 Changed 'openssl engine' to load all engines on demand.
4094 Changed the engine header files to avoid the duplication of some
4095 declarations (they differed!).
4098 *) 'openssl engine' can now list capabilities.
4101 *) Better error reporting in 'openssl engine'.
4104 *) Never call load_dh_param(NULL) in s_server.
4107 *) Add engine application. It can currently list engines by name and
4108 identity, and test if they are actually available.
4111 *) Improve RPM specification file by forcing symbolic linking and making
4112 sure the installed documentation is also owned by root.root.
4113 [Damien Miller <djm@mindrot.org>]
4115 *) Give the OpenSSL applications more possibilities to make use of
4116 keys (public as well as private) handled by engines.
4119 *) Add OCSP code that comes from CertCo.
4122 *) Add VMS support for the Rijndael code.
4125 *) Added untested support for Nuron crypto accelerator.
4128 *) Add support for external cryptographic devices. This code was
4129 previously distributed separately as the "engine" branch.
4130 [Geoff Thorpe, Richard Levitte]
4132 *) Rework the filename-translation in the DSO code. It is now possible to
4133 have far greater control over how a "name" is turned into a filename
4134 depending on the operating environment and any oddities about the
4135 different shared library filenames on each system.
4138 *) Support threads on FreeBSD-elf in Configure.
4141 *) Fix for SHA1 assembly problem with MASM: it produces
4142 warnings about corrupt line number information when assembling
4143 with debugging information. This is caused by the overlapping
4145 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
4148 NCONF_get_number() has no error checking at all. As a replacement,
4149 NCONF_get_number_e() is defined (_e for "error checking") and is
4150 promoted strongly. The old NCONF_get_number is kept around for
4151 binary backward compatibility.
4152 Make it possible for methods to load from something other than a BIO,
4153 by providing a function pointer that is given a name instead of a BIO.
4154 For example, this could be used to load configuration data from an
4158 *) Fix for non blocking accept BIOs. Added new I/O special reason
4159 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
4160 with non blocking I/O was not possible because no retry code was
4161 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
4165 *) Added the beginnings of Rijndael support.
4168 *) Fix for bug in DirectoryString mask setting. Add support for
4169 X509_NAME_print_ex() in 'req' and X509_print_ex() function
4170 to allow certificate printing to more controllable, additional
4171 'certopt' option to 'x509' to allow new printing options to be
4175 *) Clean old EAY MD5 hack from e_os.h.
4178 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
4180 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4181 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4182 [Joe Orton, Steve Henson]
4184 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
4186 *) Fix additional bug revealed by the NISCC test suite:
4188 Stop bug triggering large recursion when presented with
4189 certain ASN.1 tags (CVE-2003-0851)
4192 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
4194 *) Fix various bugs revealed by running the NISCC test suite:
4196 Stop out of bounds reads in the ASN1 code when presented with
4197 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4199 If verify callback ignores invalid public key errors don't try to check
4200 certificate signature with the NULL public key.
4204 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4205 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4209 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4210 extra data after the compression methods not only for TLS 1.0
4211 but also for SSL 3.0 (as required by the specification).
4212 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4214 *) Change X509_certificate_type() to mark the key as exported/exportable
4215 when it's 512 *bits* long, not 512 bytes.
4218 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
4220 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4221 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4222 a protocol version number mismatch like a decryption error
4223 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4226 *) Turn on RSA blinding by default in the default implementation
4227 to avoid a timing attack. Applications that don't want it can call
4228 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4229 They would be ill-advised to do so in most cases.
4230 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4232 *) Change RSA blinding code so that it works when the PRNG is not
4233 seeded (in this case, the secret RSA exponent is abused as
4234 an unpredictable seed -- if it is not unpredictable, there
4235 is no point in blinding anyway). Make RSA blinding thread-safe
4236 by remembering the creator's thread ID in rsa->blinding and
4237 having all other threads use local one-time blinding factors
4238 (this requires more computation than sharing rsa->blinding, but
4239 avoids excessive locking; and if an RSA object is not shared
4240 between threads, blinding will still be very fast).
4243 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
4245 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4246 via timing by performing a MAC computation even if incorrrect
4247 block cipher padding has been found. This is a countermeasure
4248 against active attacks where the attacker has to distinguish
4249 between bad padding and a MAC verification error. (CVE-2003-0078)
4251 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4252 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4253 Martin Vuagnoux (EPFL, Ilion)]
4255 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
4257 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
4258 memory from it's contents. This is done with a counter that will
4259 place alternating values in each byte. This can be used to solve
4260 two issues: 1) the removal of calls to memset() by highly optimizing
4261 compilers, and 2) cleansing with other values than 0, since those can
4262 be read through on certain media, for example a swap space on disk.
4265 *) Bugfix: client side session caching did not work with external caching,
4266 because the session->cipher setting was not restored when reloading
4267 from the external cache. This problem was masked, when
4268 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
4269 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
4272 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
4273 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
4274 [Zeev Lieber <zeev-l@yahoo.com>]
4276 *) Undo an undocumented change introduced in 0.9.6e which caused
4277 repeated calls to OpenSSL_add_all_ciphers() and
4278 OpenSSL_add_all_digests() to be ignored, even after calling
4282 *) Change the default configuration reader to deal with last line not
4283 being properly terminated.
4286 *) Change X509_NAME_cmp() so it applies the special rules on handling
4287 DN values that are of type PrintableString, as well as RDNs of type
4288 emailAddress where the value has the type ia5String.
4289 [stefank@valicert.com via Richard Levitte]
4291 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
4292 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
4293 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
4294 the bitwise-OR of the two for use by the majority of applications
4295 wanting this behaviour, and update the docs. The documented
4296 behaviour and actual behaviour were inconsistent and had been
4297 changing anyway, so this is more a bug-fix than a behavioural
4299 [Geoff Thorpe, diagnosed by Nadav Har'El]
4301 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
4302 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
4305 *) Fix initialization code race conditions in
4306 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
4307 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
4308 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
4309 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
4310 ssl2_get_cipher_by_char(),
4311 ssl3_get_cipher_by_char().
4312 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
4314 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
4315 the cached sessions are flushed, as the remove_cb() might use ex_data
4316 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
4317 (see [openssl.org #212]).
4318 [Geoff Thorpe, Lutz Jaenicke]
4320 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
4321 length, instead of the encoding length to d2i_ASN1_OBJECT.
4324 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
4326 *) [In 0.9.6g-engine release:]
4327 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
4328 [Lynn Gazis <lgazis@rainbow.com>]
4330 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
4332 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
4333 and get fix the header length calculation.
4334 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
4335 Alon Kantor <alonk@checkpoint.com> (and others),
4338 *) Use proper error handling instead of 'assertions' in buffer
4339 overflow checks added in 0.9.6e. This prevents DoS (the
4340 assertions could call abort()).
4341 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
4343 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
4345 *) Add various sanity checks to asn1_get_length() to reject
4346 the ASN1 length bytes if they exceed sizeof(long), will appear
4347 negative or the content length exceeds the length of the
4349 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4351 *) Fix cipher selection routines: ciphers without encryption had no flags
4352 for the cipher strength set and where therefore not handled correctly
4353 by the selection routines (PR #130).
4356 *) Fix EVP_dsa_sha macro.
4360 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
4361 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
4362 that was added in OpenSSL 0.9.6d.
4364 As the countermeasure turned out to be incompatible with some
4365 broken SSL implementations, the new option is part of SSL_OP_ALL.
4366 SSL_OP_ALL is usually employed when compatibility with weird SSL
4367 implementations is desired (e.g. '-bugs' option to 's_client' and
4368 's_server'), so the new option is automatically set in many
4372 *) Changes in security patch:
4374 Changes marked "(CHATS)" were sponsored by the Defense Advanced
4375 Research Projects Agency (DARPA) and Air Force Research Laboratory,
4376 Air Force Materiel Command, USAF, under agreement number
4379 *) Add various sanity checks to asn1_get_length() to reject
4380 the ASN1 length bytes if they exceed sizeof(long), will appear
4381 negative or the content length exceeds the length of the
4382 supplied buffer. (CVE-2002-0659)
4383 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4385 *) Assertions for various potential buffer overflows, not known to
4387 [Ben Laurie (CHATS)]
4389 *) Various temporary buffers to hold ASCII versions of integers were
4390 too small for 64 bit platforms. (CVE-2002-0655)
4391 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
4393 *) Remote buffer overflow in SSL3 protocol - an attacker could
4394 supply an oversized session ID to a client. (CVE-2002-0656)
4395 [Ben Laurie (CHATS)]
4397 *) Remote buffer overflow in SSL2 protocol - an attacker could
4398 supply an oversized client master key. (CVE-2002-0656)
4399 [Ben Laurie (CHATS)]
4401 Changes between 0.9.6c and 0.9.6d [9 May 2002]
4403 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
4404 encoded as NULL) with id-dsa-with-sha1.
4405 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
4407 *) Check various X509_...() return values in apps/req.c.
4408 [Nils Larsch <nla@trustcenter.de>]
4410 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
4411 an end-of-file condition would erronously be flagged, when the CRLF
4412 was just at the end of a processed block. The bug was discovered when
4413 processing data through a buffering memory BIO handing the data to a
4414 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
4415 <ptsekov@syntrex.com> and Nedelcho Stanev.
4418 *) Implement a countermeasure against a vulnerability recently found
4419 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
4420 before application data chunks to avoid the use of known IVs
4421 with data potentially chosen by the attacker.
4424 *) Fix length checks in ssl3_get_client_hello().
4427 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
4428 to prevent ssl3_read_internal() from incorrectly assuming that
4429 ssl3_read_bytes() found application data while handshake
4430 processing was enabled when in fact s->s3->in_read_app_data was
4431 merely automatically cleared during the initial handshake.
4432 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
4434 *) Fix object definitions for Private and Enterprise: they were not
4435 recognized in their shortname (=lowercase) representation. Extend
4436 obj_dat.pl to issue an error when using undefined keywords instead
4437 of silently ignoring the problem (Svenning Sorensen
4438 <sss@sss.dnsalias.net>).
4441 *) Fix DH_generate_parameters() so that it works for 'non-standard'
4442 generators, i.e. generators other than 2 and 5. (Previously, the
4443 code did not properly initialise the 'add' and 'rem' values to
4444 BN_generate_prime().)
4446 In the new general case, we do not insist that 'generator' is
4447 actually a primitive root: This requirement is rather pointless;
4448 a generator of the order-q subgroup is just as good, if not
4452 *) Map new X509 verification errors to alerts. Discovered and submitted by
4453 Tom Wu <tom@arcot.com>.
4456 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
4457 returning non-zero before the data has been completely received
4458 when using non-blocking I/O.
4459 [Bodo Moeller; problem pointed out by John Hughes]
4461 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
4462 [Ben Laurie, Lutz Jaenicke]
4464 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
4465 Yoram Zahavi <YoramZ@gilian.com>).
4468 *) Add information about CygWin 1.3 and on, and preserve proper
4469 configuration for the versions before that.
4470 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4472 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
4473 check whether we deal with a copy of a session and do not delete from
4474 the cache in this case. Problem reported by "Izhar Shoshani Levi"
4475 <izhar@checkpoint.com>.
4478 *) Do not store session data into the internal session cache, if it
4479 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
4480 flag is set). Proposed by Aslam <aslam@funk.com>.
4483 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
4487 *) [In 0.9.6d-engine release:]
4488 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4489 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4491 *) Add the configuration target linux-s390x.
4492 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
4494 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
4495 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
4496 variable as an indication that a ClientHello message has been
4497 received. As the flag value will be lost between multiple
4498 invocations of ssl3_accept when using non-blocking I/O, the
4499 function may not be aware that a handshake has actually taken
4500 place, thus preventing a new session from being added to the
4503 To avoid this problem, we now set s->new_session to 2 instead of
4504 using a local variable.
4505 [Lutz Jaenicke, Bodo Moeller]
4507 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
4508 if the SSL_R_LENGTH_MISMATCH error is detected.
4509 [Geoff Thorpe, Bodo Moeller]
4511 *) New 'shared_ldflag' column in Configure platform table.
4514 *) Fix EVP_CIPHER_mode macro.
4515 ["Dan S. Camper" <dan@bti.net>]
4517 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
4518 type, we must throw them away by setting rr->length to 0.
4519 [D P Chang <dpc@qualys.com>]
4521 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
4523 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
4524 <Dominikus.Scherkl@biodata.com>. (The previous implementation
4525 worked incorrectly for those cases where range = 10..._2 and
4526 3*range is two bits longer than range.)
4529 *) Only add signing time to PKCS7 structures if it is not already
4533 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
4534 OBJ_ld_ce should be OBJ_id_ce.
4535 Also some ip-pda OIDs in crypto/objects/objects.txt were
4536 incorrect (cf. RFC 3039).
4537 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
4539 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
4540 returns early because it has nothing to do.
4541 [Andy Schneider <andy.schneider@bjss.co.uk>]
4543 *) [In 0.9.6c-engine release:]
4544 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
4545 [Andy Schneider <andy.schneider@bjss.co.uk>]
4547 *) [In 0.9.6c-engine release:]
4548 Add support for Cryptographic Appliance's keyserver technology.
4549 (Use engine 'keyclient')
4550 [Cryptographic Appliances and Geoff Thorpe]
4552 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
4553 is called via tools/c89.sh because arguments have to be
4554 rearranged (all '-L' options must appear before the first object
4556 [Richard Shapiro <rshapiro@abinitio.com>]
4558 *) [In 0.9.6c-engine release:]
4559 Add support for Broadcom crypto accelerator cards, backported
4561 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
4563 *) [In 0.9.6c-engine release:]
4564 Add support for SureWare crypto accelerator cards from
4565 Baltimore Technologies. (Use engine 'sureware')
4566 [Baltimore Technologies and Mark Cox]
4568 *) [In 0.9.6c-engine release:]
4569 Add support for crypto accelerator cards from Accelerated
4570 Encryption Processing, www.aep.ie. (Use engine 'aep')
4571 [AEP Inc. and Mark Cox]
4573 *) Add a configuration entry for gcc on UnixWare.
4574 [Gary Benson <gbenson@redhat.com>]
4576 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
4577 messages are stored in a single piece (fixed-length part and
4578 variable-length part combined) and fix various bugs found on the way.
4581 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
4582 instead. BIO_gethostbyname() does not know what timeouts are
4583 appropriate, so entries would stay in cache even when they have
4585 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
4587 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
4588 faced with a pathologically small ClientHello fragment that does
4589 not contain client_version: Instead of aborting with an error,
4590 simply choose the highest available protocol version (i.e.,
4591 TLS 1.0 unless it is disabled). In practice, ClientHello
4592 messages are never sent like this, but this change gives us
4593 strictly correct behaviour at least for TLS.
4596 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
4597 never resets s->method to s->ctx->method when called from within
4598 one of the SSL handshake functions.
4599 [Bodo Moeller; problem pointed out by Niko Baric]
4601 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
4602 (sent using the client's version number) if client_version is
4603 smaller than the protocol version in use. Also change
4604 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
4605 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
4606 the client will at least see that alert.
4609 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
4613 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
4614 client receives HelloRequest while in a handshake.
4615 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
4617 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
4618 should end in 'break', not 'goto end' which circuments various
4619 cleanups done in state SSL_ST_OK. But session related stuff
4620 must be disabled for SSL_ST_OK in the case that we just sent a
4623 Also avoid some overhead by not calling ssl_init_wbio_buffer()
4624 before just sending a HelloRequest.
4625 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
4627 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
4628 reveal whether illegal block cipher padding was found or a MAC
4629 verification error occured. (Neither SSLerr() codes nor alerts
4630 are directly visible to potential attackers, but the information
4631 may leak via logfiles.)
4633 Similar changes are not required for the SSL 2.0 implementation
4634 because the number of padding bytes is sent in clear for SSL 2.0,
4635 and the extra bytes are just ignored. However ssl/s2_pkt.c
4636 failed to verify that the purported number of padding bytes is in
4640 *) Add OpenUNIX-8 support including shared libraries
4641 (Boyd Lynn Gerber <gerberb@zenez.com>).
4644 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
4645 'wristwatch attack' using huge encoding parameters (cf.
4646 James H. Manger's CRYPTO 2001 paper). Note that the
4647 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
4648 encoding parameters and hence was not vulnerable.
4651 *) BN_sqr() bug fix.
4652 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
4654 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
4655 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
4656 followed by modular reduction.
4657 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
4659 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
4660 equivalent based on BN_pseudo_rand() instead of BN_rand().
4663 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
4664 This function was broken, as the check for a new client hello message
4665 to handle SGC did not allow these large messages.
4666 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
4669 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
4672 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
4673 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
4676 *) Rework the configuration and shared library support for Tru64 Unix.
4677 The configuration part makes use of modern compiler features and
4678 still retains old compiler behavior for those that run older versions
4679 of the OS. The shared library support part includes a variant that
4680 uses the RPATH feature, and is available through the special
4681 configuration target "alpha-cc-rpath", which will never be selected
4683 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
4685 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
4686 with the same message size as in ssl3_get_certificate_request().
4687 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
4688 messages might inadvertently be reject as too long.
4689 [Petr Lampa <lampa@fee.vutbr.cz>]
4691 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
4694 *) Modified SSL library such that the verify_callback that has been set
4695 specificly for an SSL object with SSL_set_verify() is actually being
4696 used. Before the change, a verify_callback set with this function was
4697 ignored and the verify_callback() set in the SSL_CTX at the time of
4698 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
4699 to allow the necessary settings.
4702 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
4703 explicitly to NULL, as at least on Solaris 8 this seems not always to be
4704 done automatically (in contradiction to the requirements of the C
4705 standard). This made problems when used from OpenSSH.
4708 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
4709 dh->length and always used
4711 BN_rand_range(priv_key, dh->p).
4713 BN_rand_range() is not necessary for Diffie-Hellman, and this
4714 specific range makes Diffie-Hellman unnecessarily inefficient if
4715 dh->length (recommended exponent length) is much smaller than the
4716 length of dh->p. We could use BN_rand_range() if the order of
4717 the subgroup was stored in the DH structure, but we only have
4722 BN_rand(priv_key, l, ...)
4724 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
4730 RSA_eay_public_encrypt
4731 RSA_eay_private_decrypt
4732 RSA_eay_private_encrypt (signing)
4733 RSA_eay_public_decrypt (signature verification)
4735 (default implementations for RSA_public_encrypt,
4736 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
4737 always reject numbers >= n.
4740 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
4741 to synchronize access to 'locking_thread'. This is necessary on
4742 systems where access to 'locking_thread' (an 'unsigned long'
4743 variable) is not atomic.
4746 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
4747 *before* setting the 'crypto_lock_rand' flag. The previous code had
4748 a race condition if 0 is a valid thread ID.
4749 [Travis Vitek <vitek@roguewave.com>]
4751 *) Add support for shared libraries under Irix.
4752 [Albert Chin-A-Young <china@thewrittenword.com>]
4754 *) Add configuration option to build on Linux on both big-endian and
4756 [Ralf Baechle <ralf@uni-koblenz.de>]
4758 *) Add the possibility to create shared libraries on HP-UX.
4761 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
4763 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
4764 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
4765 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
4766 PRNG state recovery was possible based on the output of
4767 one PRNG request appropriately sized to gain knowledge on
4768 'md' followed by enough consecutive 1-byte PRNG requests
4769 to traverse all of 'state'.
4771 1. When updating 'md_local' (the current thread's copy of 'md')
4772 during PRNG output generation, hash all of the previous
4773 'md_local' value, not just the half used for PRNG output.
4775 2. Make the number of bytes from 'state' included into the hash
4776 independent from the number of PRNG bytes requested.
4778 The first measure alone would be sufficient to avoid
4779 Markku-Juhani's attack. (Actually it had never occurred
4780 to me that the half of 'md_local' used for chaining was the
4781 half from which PRNG output bytes were taken -- I had always
4782 assumed that the secret half would be used.) The second
4783 measure makes sure that additional data from 'state' is never
4784 mixed into 'md_local' in small portions; this heuristically
4785 further strengthens the PRNG.
4788 *) Fix crypto/bn/asm/mips3.s.
4791 *) When only the key is given to "enc", the IV is undefined. Print out
4792 an error message in this case.
4795 *) Handle special case when X509_NAME is empty in X509 printing routines.
4798 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
4799 positive and less than q.
4802 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
4803 used: it isn't thread safe and the add_lock_callback should handle
4805 [Paul Rose <Paul.Rose@bridge.com>]
4807 *) Verify that incoming data obeys the block size in
4808 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
4812 [Ulf Möller, Bodo Möller]
4814 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
4815 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
4816 when fixing the server behaviour for backwards-compatible 'client
4817 hello' messages. (Note that the attack is impractical against
4818 SSL 3.0 and TLS 1.0 anyway because length and version checking
4819 means that the probability of guessing a valid ciphertext is
4820 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
4823 Before 0.9.5, the countermeasure (hide the error by generating a
4824 random 'decryption result') did not work properly because
4825 ERR_clear_error() was missing, meaning that SSL_get_error() would
4826 detect the supposedly ignored error.
4828 Both problems are now fixed.
4831 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
4832 (previously it was 1024).
4835 *) Fix for compatibility mode trust settings: ignore trust settings
4836 unless some valid trust or reject settings are present.
4839 *) Fix for blowfish EVP: its a variable length cipher.
4842 *) Fix various bugs related to DSA S/MIME verification. Handle missing
4843 parameters in DSA public key structures and return an error in the
4844 DSA routines if parameters are absent.
4847 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
4848 in the current directory if neither $RANDFILE nor $HOME was set.
4849 RAND_file_name() in 0.9.6a returned NULL in this case. This has
4850 caused some confusion to Windows users who haven't defined $HOME.
4851 Thus RAND_file_name() is changed again: e_os.h can define a
4852 DEFAULT_HOME, which will be used if $HOME is not set.
4853 For Windows, we use "C:"; on other platforms, we still require
4854 environment variables.
4856 *) Move 'if (!initialized) RAND_poll()' into regions protected by
4857 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
4858 having multiple threads call RAND_poll() concurrently.
4861 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
4862 combination of a flag and a thread ID variable.
4863 Otherwise while one thread is in ssleay_rand_bytes (which sets the
4864 flag), *other* threads can enter ssleay_add_bytes without obeying
4865 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
4866 that they do not hold after the first thread unsets add_do_not_lock).
4869 *) Change bctest again: '-x' expressions are not available in all
4873 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
4875 *) Fix a couple of memory leaks in PKCS7_dataDecode()
4876 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
4878 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
4879 the default extension for executables, if any. Also, make the perl
4880 scripts that use symlink() to test if it really exists and use "cp"
4881 if it doesn't. All this made OpenSSL compilable and installable in
4885 *) Fix for asn1_GetSequence() for indefinite length constructed data.
4886 If SEQUENCE is length is indefinite just set c->slen to the total
4887 amount of data available.
4888 [Steve Henson, reported by shige@FreeBSD.org]
4889 [This change does not apply to 0.9.7.]
4891 *) Change bctest to avoid here-documents inside command substitution
4892 (workaround for FreeBSD /bin/sh bug).
4893 For compatibility with Ultrix, avoid shell functions (introduced
4894 in the bctest version that searches along $PATH).
4897 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
4898 with des_encrypt() defined on some operating systems, like Solaris
4902 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
4903 On the Importance of Eliminating Errors in Cryptographic
4904 Computations, J. Cryptology 14 (2001) 2, 101-119,
4905 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
4908 *) MIPS assembler BIGNUM division bug fix.
4911 *) Disabled incorrect Alpha assembler code.
4914 *) Fix PKCS#7 decode routines so they correctly update the length
4915 after reading an EOC for the EXPLICIT tag.
4917 [This change does not apply to 0.9.7.]
4919 *) Fix bug in PKCS#12 key generation routines. This was triggered
4920 if a 3DES key was generated with a 0 initial byte. Include
4921 PKCS12_BROKEN_KEYGEN compilation option to retain the old
4922 (but broken) behaviour.
4925 *) Enhance bctest to search for a working bc along $PATH and print
4927 [Tim Rice <tim@multitalents.net> via Richard Levitte]
4929 *) Fix memory leaks in err.c: free err_data string if necessary;
4930 don't write to the wrong index in ERR_set_error_data.
4933 *) Implement ssl23_peek (analogous to ssl23_read), which previously
4937 *) Replace rdtsc with _emit statements for VC++ version 5.
4938 [Jeremy Cooper <jeremy@baymoo.org>]
4940 *) Make it possible to reuse SSLv2 sessions.
4943 *) In copy_email() check for >= 0 as a return value for
4944 X509_NAME_get_index_by_NID() since 0 is a valid index.
4945 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
4947 *) Avoid coredump with unsupported or invalid public keys by checking if
4948 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
4949 PKCS7_verify() fails with non detached data.
4952 *) Don't use getenv in library functions when run as setuid/setgid.
4953 New function OPENSSL_issetugid().
4956 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
4957 due to incorrect handling of multi-threading:
4959 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
4961 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
4963 3. Count how many times MemCheck_off() has been called so that
4964 nested use can be treated correctly. This also avoids
4965 inband-signalling in the previous code (which relied on the
4966 assumption that thread ID 0 is impossible).
4969 *) Add "-rand" option also to s_client and s_server.
4972 *) Fix CPU detection on Irix 6.x.
4973 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
4974 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4976 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
4979 [This change does not apply to 0.9.7.]
4981 *) Use the cached encoding of an X509_NAME structure rather than
4982 copying it. This is apparently the reason for the libsafe "errors"
4983 but the code is actually correct.
4986 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
4987 Bleichenbacher's DSA attack.
4988 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
4989 to be set and top=0 forces the highest bit to be set; top=-1 is new
4990 and leaves the highest bit random.
4991 [Ulf Moeller, Bodo Moeller]
4993 *) In the NCONF_...-based implementations for CONF_... queries
4994 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
4995 a temporary CONF structure with the data component set to NULL
4996 (which gives segmentation faults in lh_retrieve).
4997 Instead, use NULL for the CONF pointer in CONF_get_string and
4998 CONF_get_number (which may use environment variables) and directly
4999 return NULL from CONF_get_section.
5002 *) Fix potential buffer overrun for EBCDIC.
5005 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
5006 keyUsage if basicConstraints absent for a CA.
5009 *) Make SMIME_write_PKCS7() write mail header values with a format that
5010 is more generally accepted (no spaces before the semicolon), since
5011 some programs can't parse those values properly otherwise. Also make
5012 sure BIO's that break lines after each write do not create invalid
5016 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
5017 macros previously used would not encode an empty SEQUENCE OF
5018 and break the signature.
5020 [This change does not apply to 0.9.7.]
5022 *) Zero the premaster secret after deriving the master secret in
5026 *) Add some EVP_add_digest_alias registrations (as found in
5027 OpenSSL_add_all_digests()) to SSL_library_init()
5028 aka OpenSSL_add_ssl_algorithms(). This provides improved
5029 compatibility with peers using X.509 certificates
5030 with unconventional AlgorithmIdentifier OIDs.
5033 *) Fix for Irix with NO_ASM.
5034 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
5036 *) ./config script fixes.
5037 [Ulf Moeller, Richard Levitte]
5039 *) Fix 'openssl passwd -1'.
5042 *) Change PKCS12_key_gen_asc() so it can cope with non null
5043 terminated strings whose length is passed in the passlen
5044 parameter, for example from PEM callbacks. This was done
5045 by adding an extra length parameter to asc2uni().
5046 [Steve Henson, reported by <oddissey@samsung.co.kr>]
5048 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
5049 call failed, free the DSA structure.
5052 *) Fix to uni2asc() to cope with zero length Unicode strings.
5053 These are present in some PKCS#12 files.
5056 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
5057 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
5058 when writing a 32767 byte record.
5059 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
5061 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
5062 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
5064 (RSA objects have a reference count access to which is protected
5065 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
5066 so they are meant to be shared between threads.)
5067 [Bodo Moeller, Geoff Thorpe; original patch submitted by
5068 "Reddie, Steven" <Steven.Reddie@ca.com>]
5070 *) Fix a deadlock in CRYPTO_mem_leaks().
5073 *) Use better test patterns in bntest.
5076 *) rand_win.c fix for Borland C.
5079 *) BN_rshift bugfix for n == 0.
5082 *) Add a 'bctest' script that checks for some known 'bc' bugs
5083 so that 'make test' does not abort just because 'bc' is broken.
5086 *) Store verify_result within SSL_SESSION also for client side to
5087 avoid potential security hole. (Re-used sessions on the client side
5088 always resulted in verify_result==X509_V_OK, not using the original
5089 result of the server certificate verification.)
5092 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
5093 SSL3_RT_APPLICATION_DATA, return 0.
5094 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
5098 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
5099 releases, have been re-implemented by renaming the previous
5100 implementations of ssl2_read and ssl3_read to ssl2_read_internal
5101 and ssl3_read_internal, respectively, and adding 'peek' parameters
5102 to them. The new ssl[23]_{read,peek} functions are calls to
5103 ssl[23]_read_internal with the 'peek' flag set appropriately.
5104 A 'peek' parameter has also been added to ssl3_read_bytes, which
5105 does the actual work for ssl3_read_internal.
5108 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
5109 the method-specific "init()" handler. Also clean up ex_data after
5110 calling the method-specific "finish()" handler. Previously, this was
5111 happening the other way round.
5114 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
5115 The previous value, 12, was not always sufficient for BN_mod_exp().
5118 *) Make sure that shared libraries get the internal name engine with
5119 the full version number and not just 0. This should mark the
5120 shared libraries as not backward compatible. Of course, this should
5121 be changed again when we can guarantee backward binary compatibility.
5124 *) Fix typo in get_cert_by_subject() in by_dir.c
5125 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
5127 *) Rework the system to generate shared libraries:
5129 - Make note of the expected extension for the shared libraries and
5130 if there is a need for symbolic links from for example libcrypto.so.0
5131 to libcrypto.so.0.9.7. There is extended info in Configure for
5134 - Make as few rebuilds of the shared libraries as possible.
5136 - Still avoid linking the OpenSSL programs with the shared libraries.
5138 - When installing, install the shared libraries separately from the
5142 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
5144 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
5145 and not in SSL_clear because the latter is also used by the
5146 accept/connect functions; previously, the settings made by
5147 SSL_set_read_ahead would be lost during the handshake.
5148 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
5150 *) Correct util/mkdef.pl to be selective about disabled algorithms.
5151 Previously, it would create entries for disableed algorithms no
5155 *) Added several new manual pages for SSL_* function.
5158 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
5160 *) In ssl23_get_client_hello, generate an error message when faced
5161 with an initial SSL 3.0/TLS record that is too small to contain the
5162 first two bytes of the ClientHello message, i.e. client_version.
5163 (Note that this is a pathologic case that probably has never happened
5164 in real life.) The previous approach was to use the version number
5165 from the record header as a substitute; but our protocol choice
5166 should not depend on that one because it is not authenticated
5167 by the Finished messages.
5170 *) More robust randomness gathering functions for Windows.
5171 [Jeffrey Altman <jaltman@columbia.edu>]
5173 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
5174 not set then we don't setup the error code for issuer check errors
5175 to avoid possibly overwriting other errors which the callback does
5176 handle. If an application does set the flag then we assume it knows
5177 what it is doing and can handle the new informational codes
5181 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
5182 a general "ANY" type, as such it should be able to decode anything
5183 including tagged types. However it didn't check the class so it would
5184 wrongly interpret tagged types in the same way as their universal
5185 counterpart and unknown types were just rejected. Changed so that the
5186 tagged and unknown types are handled in the same way as a SEQUENCE:
5187 that is the encoding is stored intact. There is also a new type
5188 "V_ASN1_OTHER" which is used when the class is not universal, in this
5189 case we have no idea what the actual type is so we just lump them all
5193 *) On VMS, stdout may very well lead to a file that is written to
5194 in a record-oriented fashion. That means that every write() will
5195 write a separate record, which will be read separately by the
5196 programs trying to read from it. This can be very confusing.
5198 The solution is to put a BIO filter in the way that will buffer
5199 text until a linefeed is reached, and then write everything a
5200 line at a time, so every record written will be an actual line,
5201 not chunks of lines and not (usually doesn't happen, but I've
5202 seen it once) several lines in one record. BIO_f_linebuffer() is
5205 Currently, it's a VMS-only method, because that's where it has
5206 been tested well enough.
5209 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
5210 it can return incorrect results.
5211 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
5212 but it was in 0.9.6-beta[12].)
5215 *) Disable the check for content being present when verifying detached
5216 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
5217 include zero length content when signing messages.
5220 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
5221 BIO_ctrl (for BIO pairs).
5224 *) Add DSO method for VMS.
5227 *) Bug fix: Montgomery multiplication could produce results with the
5231 *) Add RPM specification openssl.spec and modify it to build three
5232 packages. The default package contains applications, application
5233 documentation and run-time libraries. The devel package contains
5234 include files, static libraries and function documentation. The
5235 doc package contains the contents of the doc directory. The original
5236 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
5239 *) Add a large number of documentation files for many SSL routines.
5240 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5242 *) Add a configuration entry for Sony News 4.
5243 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
5245 *) Don't set the two most significant bits to one when generating a
5246 random number < q in the DSA library.
5249 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
5250 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
5251 the underlying transport is blocking) if a handshake took place.
5252 (The default behaviour is needed by applications such as s_client
5253 and s_server that use select() to determine when to use SSL_read;
5254 but for applications that know in advance when to expect data, it
5255 just makes things more complicated.)
5258 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
5262 *) Add a few more EBCDIC conditionals that make `req' and `x509'
5263 work better on such systems.
5264 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5266 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
5267 Update PKCS12_parse() so it copies the friendlyName and the
5268 keyid to the certificates aux info.
5271 *) Fix bug in PKCS7_verify() which caused an infinite loop
5272 if there was more than one signature.
5273 [Sven Uszpelkat <su@celocom.de>]
5275 *) Major change in util/mkdef.pl to include extra information
5276 about each symbol, as well as presentig variables as well
5277 as functions. This change means that there's n more need
5278 to rebuild the .num files when some algorithms are excluded.
5281 *) Allow the verify time to be set by an application,
5282 rather than always using the current time.
5285 *) Phase 2 verify code reorganisation. The certificate
5286 verify code now looks up an issuer certificate by a
5287 number of criteria: subject name, authority key id
5288 and key usage. It also verifies self signed certificates
5289 by the same criteria. The main comparison function is
5290 X509_check_issued() which performs these checks.
5292 Lot of changes were necessary in order to support this
5293 without completely rewriting the lookup code.
5295 Authority and subject key identifier are now cached.
5297 The LHASH 'certs' is X509_STORE has now been replaced
5298 by a STACK_OF(X509_OBJECT). This is mainly because an
5299 LHASH can't store or retrieve multiple objects with
5300 the same hash value.
5302 As a result various functions (which were all internal
5303 use only) have changed to handle the new X509_STORE
5304 structure. This will break anything that messed round
5305 with X509_STORE internally.
5307 The functions X509_STORE_add_cert() now checks for an
5308 exact match, rather than just subject name.
5310 The X509_STORE API doesn't directly support the retrieval
5311 of multiple certificates matching a given criteria, however
5312 this can be worked round by performing a lookup first
5313 (which will fill the cache with candidate certificates)
5314 and then examining the cache for matches. This is probably
5315 the best we can do without throwing out X509_LOOKUP
5316 entirely (maybe later...).
5318 The X509_VERIFY_CTX structure has been enhanced considerably.
5320 All certificate lookup operations now go via a get_issuer()
5321 callback. Although this currently uses an X509_STORE it
5322 can be replaced by custom lookups. This is a simple way
5323 to bypass the X509_STORE hackery necessary to make this
5324 work and makes it possible to use more efficient techniques
5325 in future. A very simple version which uses a simple
5326 STACK for its trusted certificate store is also provided
5327 using X509_STORE_CTX_trusted_stack().
5329 The verify_cb() and verify() callbacks now have equivalents
5330 in the X509_STORE_CTX structure.
5332 X509_STORE_CTX also has a 'flags' field which can be used
5333 to customise the verify behaviour.
5336 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
5337 excludes S/MIME capabilities.
5340 *) When a certificate request is read in keep a copy of the
5341 original encoding of the signed data and use it when outputing
5342 again. Signatures then use the original encoding rather than
5343 a decoded, encoded version which may cause problems if the
5344 request is improperly encoded.
5347 *) For consistency with other BIO_puts implementations, call
5348 buffer_write(b, ...) directly in buffer_puts instead of calling
5351 In BIO_puts, increment b->num_write as in BIO_write.
5352 [Peter.Sylvester@EdelWeb.fr]
5354 *) Fix BN_mul_word for the case where the word is 0. (We have to use
5355 BN_zero, we may not return a BIGNUM with an array consisting of
5359 *) Avoid calling abort() from within the library when problems are
5360 detected, except if preprocessor symbols have been defined
5361 (such as REF_CHECK, BN_DEBUG etc.).
5364 *) New openssl application 'rsautl'. This utility can be
5365 used for low level RSA operations. DER public key
5366 BIO/fp routines also added.
5369 *) New Configure entry and patches for compiling on QNX 4.
5370 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
5372 *) A demo state-machine implementation was sponsored by
5373 Nuron (http://www.nuron.com/) and is now available in
5374 demos/state_machine.
5377 *) New options added to the 'dgst' utility for signature
5378 generation and verification.
5381 *) Unrecognized PKCS#7 content types are now handled via a
5382 catch all ASN1_TYPE structure. This allows unsupported
5383 types to be stored as a "blob" and an application can
5384 encode and decode it manually.
5387 *) Fix various signed/unsigned issues to make a_strex.c
5389 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
5391 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
5392 length if passed a buffer. ASN1_INTEGER_to_BN failed
5393 if passed a NULL BN and its argument was negative.
5394 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
5396 *) Modification to PKCS#7 encoding routines to output definite
5397 length encoding. Since currently the whole structures are in
5398 memory there's not real point in using indefinite length
5399 constructed encoding. However if OpenSSL is compiled with
5400 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
5403 *) Added BIO_vprintf() and BIO_vsnprintf().
5406 *) Added more prefixes to parse for in the the strings written
5407 through a logging bio, to cover all the levels that are available
5408 through syslog. The prefixes are now:
5410 PANIC, EMERG, EMR => LOG_EMERG
5411 ALERT, ALR => LOG_ALERT
5412 CRIT, CRI => LOG_CRIT
5413 ERROR, ERR => LOG_ERR
5414 WARNING, WARN, WAR => LOG_WARNING
5415 NOTICE, NOTE, NOT => LOG_NOTICE
5416 INFO, INF => LOG_INFO
5417 DEBUG, DBG => LOG_DEBUG
5419 and as before, if none of those prefixes are present at the
5420 beginning of the string, LOG_ERR is chosen.
5422 On Win32, the LOG_* levels are mapped according to this:
5424 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
5425 LOG_WARNING => EVENTLOG_WARNING_TYPE
5426 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
5430 *) Made it possible to reconfigure with just the configuration
5431 argument "reconf" or "reconfigure". The command line arguments
5432 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
5433 and are retrieved from there when reconfiguring.
5437 [Assar Westerlund <assar@sics.se>, Richard Levitte]
5439 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
5442 *) The obj_dat.pl script was messing up the sorting of object
5443 names. The reason was that it compared the quoted version
5444 of strings as a result "OCSP" > "OCSP Signing" because
5445 " > SPACE. Changed script to store unquoted versions of
5446 names and add quotes on output. It was also omitting some
5447 names from the lookup table if they were given a default
5448 value (that is if SN is missing it is given the same
5449 value as LN and vice versa), these are now added on the
5450 grounds that if an object has a name we should be able to
5451 look it up. Finally added warning output when duplicate
5452 short or long names are found.
5455 *) Changes needed for Tandem NSK.
5456 [Scott Uroff <scott@xypro.com>]
5458 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
5459 RSA_padding_check_SSLv23(), special padding was never detected
5460 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
5461 version rollback attacks was not effective.
5463 In s23_clnt.c, don't use special rollback-attack detection padding
5464 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
5465 client; similarly, in s23_srvr.c, don't do the rollback check if
5466 SSL 2.0 is the only protocol enabled in the server.
5469 *) Make it possible to get hexdumps of unprintable data with 'openssl
5470 asn1parse'. By implication, the functions ASN1_parse_dump() and
5471 BIO_dump_indent() are added.
5474 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
5475 these print out strings and name structures based on various
5476 flags including RFC2253 support and proper handling of
5477 multibyte characters. Added options to the 'x509' utility
5478 to allow the various flags to be set.
5481 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
5482 Also change the functions X509_cmp_current_time() and
5483 X509_gmtime_adj() work with an ASN1_TIME structure,
5484 this will enable certificates using GeneralizedTime in validity
5485 dates to be checked.
5488 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
5489 negative public key encodings) on by default,
5490 NO_NEG_PUBKEY_BUG can be set to disable it.
5493 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
5494 content octets. An i2c_ASN1_OBJECT is unnecessary because
5495 the encoding can be trivially obtained from the structure.
5498 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
5499 not read locks (CRYPTO_r_[un]lock).
5502 *) A first attempt at creating official support for shared
5503 libraries through configuration. I've kept it so the
5504 default is static libraries only, and the OpenSSL programs
5505 are always statically linked for now, but there are
5506 preparations for dynamic linking in place.
5507 This has been tested on Linux and Tru64.
5510 *) Randomness polling function for Win9x, as described in:
5511 Peter Gutmann, Software Generation of Practically Strong
5515 *) Fix so PRNG is seeded in req if using an already existing
5519 *) New options to smime application. -inform and -outform
5520 allow alternative formats for the S/MIME message including
5521 PEM and DER. The -content option allows the content to be
5522 specified separately. This should allow things like Netscape
5523 form signing output easier to verify.
5526 *) Fix the ASN1 encoding of tags using the 'long form'.
5529 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
5530 STRING types. These convert content octets to and from the
5531 underlying type. The actual tag and length octets are
5532 already assumed to have been read in and checked. These
5533 are needed because all other string types have virtually
5534 identical handling apart from the tag. By having versions
5535 of the ASN1 functions that just operate on content octets
5536 IMPLICIT tagging can be handled properly. It also allows
5537 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
5538 and ASN1_INTEGER are identical apart from the tag.
5541 *) Change the handling of OID objects as follows:
5543 - New object identifiers are inserted in objects.txt, following
5544 the syntax given in objects.README.
5545 - objects.pl is used to process obj_mac.num and create a new
5547 - obj_dat.pl is used to create a new obj_dat.h, using the data in
5550 This is currently kind of a hack, and the perl code in objects.pl
5551 isn't very elegant, but it works as I intended. The simplest way
5552 to check that it worked correctly is to look in obj_dat.h and
5553 check the array nid_objs and make sure the objects haven't moved
5554 around (this is important!). Additions are OK, as well as
5555 consistent name changes.
5558 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
5561 *) Addition of the command line parameter '-rand file' to 'openssl req'.
5562 The given file adds to whatever has already been seeded into the
5563 random pool through the RANDFILE configuration file option or
5564 environment variable, or the default random state file.
5567 *) mkstack.pl now sorts each macro group into lexical order.
5568 Previously the output order depended on the order the files
5569 appeared in the directory, resulting in needless rewriting
5573 *) Patches to make OpenSSL compile under Win32 again. Mostly
5574 work arounds for the VC++ problem that it treats func() as
5575 func(void). Also stripped out the parts of mkdef.pl that
5576 added extra typesafe functions: these no longer exist.
5579 *) Reorganisation of the stack code. The macros are now all
5580 collected in safestack.h . Each macro is defined in terms of
5581 a "stack macro" of the form SKM_<name>(type, a, b). The
5582 DEBUG_SAFESTACK is now handled in terms of function casts,
5583 this has the advantage of retaining type safety without the
5584 use of additional functions. If DEBUG_SAFESTACK is not defined
5585 then the non typesafe macros are used instead. Also modified the
5586 mkstack.pl script to handle the new form. Needs testing to see
5587 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
5588 the default if no major problems. Similar behaviour for ASN1_SET_OF
5589 and PKCS12_STACK_OF.
5592 *) When some versions of IIS use the 'NET' form of private key the
5593 key derivation algorithm is different. Normally MD5(password) is
5594 used as a 128 bit RC4 key. In the modified case
5595 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
5596 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
5597 as the old Netscape_RSA functions except they have an additional
5598 'sgckey' parameter which uses the modified algorithm. Also added
5599 an -sgckey command line option to the rsa utility. Thanks to
5600 Adrian Peck <bertie@ncipher.com> for posting details of the modified
5601 algorithm to openssl-dev.
5604 *) The evp_local.h macros were using 'c.##kname' which resulted in
5605 invalid expansion on some systems (SCO 5.0.5 for example).
5606 Corrected to 'c.kname'.
5607 [Phillip Porch <root@theporch.com>]
5609 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
5610 a STACK of email addresses from a certificate or request, these look
5611 in the subject name and the subject alternative name extensions and
5612 omit any duplicate addresses.
5615 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
5616 This makes DSA verification about 2 % faster.
5619 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
5620 (meaning that now 2^5 values will be precomputed, which is only 4 KB
5621 plus overhead for 1024 bit moduli).
5622 This makes exponentiations about 0.5 % faster for 1024 bit
5623 exponents (as measured by "openssl speed rsa2048").
5626 *) Rename memory handling macros to avoid conflicts with other
5628 Malloc => OPENSSL_malloc
5629 Malloc_locked => OPENSSL_malloc_locked
5630 Realloc => OPENSSL_realloc
5631 Free => OPENSSL_free
5634 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
5635 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
5638 *) CygWin32 support.
5639 [John Jarvie <jjarvie@newsguy.com>]
5641 *) The type-safe stack code has been rejigged. It is now only compiled
5642 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
5643 by default all type-specific stack functions are "#define"d back to
5644 standard stack functions. This results in more streamlined output
5645 but retains the type-safety checking possibilities of the original
5649 *) The STACK code has been cleaned up, and certain type declarations
5650 that didn't make a lot of sense have been brought in line. This has
5651 also involved a cleanup of sorts in safestack.h to more correctly
5652 map type-safe stack functions onto their plain stack counterparts.
5653 This work has also resulted in a variety of "const"ifications of
5654 lots of the code, especially "_cmp" operations which should normally
5655 be prototyped with "const" parameters anyway.
5658 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
5659 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
5660 (The PRNG state consists of two parts, the large pool 'state' and 'md',
5661 where all of 'md' is used each time the PRNG is used, but 'state'
5662 is used only indexed by a cyclic counter. As entropy may not be
5663 well distributed from the beginning, 'md' is important as a
5664 chaining variable. However, the output function chains only half
5665 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
5666 all of 'md', and seeding with STATE_SIZE dummy bytes will result
5667 in all of 'state' being rewritten, with the new values depending
5668 on virtually all of 'md'. This overcomes the 80 bit limitation.)
5671 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
5672 the handshake is continued after ssl_verify_cert_chain();
5673 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
5674 can lead to 'unexplainable' connection aborts later.
5675 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
5677 *) Major EVP API cipher revision.
5678 Add hooks for extra EVP features. This allows various cipher
5679 parameters to be set in the EVP interface. Support added for variable
5680 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
5681 setting of RC2 and RC5 parameters.
5683 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
5686 Remove lots of duplicated code from the EVP library. For example *every*
5687 cipher init() function handles the 'iv' in the same way according to the
5688 cipher mode. They also all do nothing if the 'key' parameter is NULL and
5689 for CFB and OFB modes they zero ctx->num.
5691 New functionality allows removal of S/MIME code RC2 hack.
5693 Most of the routines have the same form and so can be declared in terms
5696 By shifting this to the top level EVP_CipherInit() it can be removed from
5697 all individual ciphers. If the cipher wants to handle IVs or keys
5698 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
5701 Change lots of functions like EVP_EncryptUpdate() to now return a
5702 value: although software versions of the algorithms cannot fail
5703 any installed hardware versions can.
5706 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
5707 this option is set, tolerate broken clients that send the negotiated
5708 protocol version number instead of the requested protocol version
5712 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
5713 i.e. non-zero for export ciphersuites, zero otherwise.
5714 Previous versions had this flag inverted, inconsistent with
5715 rsa_tmp_cb (..._TMP_RSA_CB).
5716 [Bodo Moeller; problem reported by Amit Chopra]
5718 *) Add missing DSA library text string. Work around for some IIS
5719 key files with invalid SEQUENCE encoding.
5722 *) Add a document (doc/standards.txt) that list all kinds of standards
5723 and so on that are implemented in OpenSSL.
5726 *) Enhance c_rehash script. Old version would mishandle certificates
5727 with the same subject name hash and wouldn't handle CRLs at all.
5728 Added -fingerprint option to crl utility, to support new c_rehash
5732 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
5735 *) Fix for SSL server purpose checking. Server checking was
5736 rejecting certificates which had extended key usage present
5737 but no ssl client purpose.
5738 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
5740 *) Make PKCS#12 code work with no password. The PKCS#12 spec
5741 is a little unclear about how a blank password is handled.
5742 Since the password in encoded as a BMPString with terminating
5743 double NULL a zero length password would end up as just the
5744 double NULL. However no password at all is different and is
5745 handled differently in the PKCS#12 key generation code. NS
5746 treats a blank password as zero length. MSIE treats it as no
5747 password on export: but it will try both on import. We now do
5748 the same: PKCS12_parse() tries zero length and no password if
5749 the password is set to "" or NULL (NULL is now a valid password:
5750 it wasn't before) as does the pkcs12 application.
5753 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
5754 perror when PEM_read_bio_X509_REQ fails, the error message must
5755 be obtained from the error queue.
5758 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
5759 it in ERR_remove_state if appropriate, and change ERR_get_state
5760 accordingly to avoid race conditions (this is necessary because
5761 thread_hash is no longer constant once set).
5764 *) Bugfix for linux-elf makefile.one.
5767 *) RSA_get_default_method() will now cause a default
5768 RSA_METHOD to be chosen if one doesn't exist already.
5769 Previously this was only set during a call to RSA_new()
5770 or RSA_new_method(NULL) meaning it was possible for
5771 RSA_get_default_method() to return NULL.
5774 *) Added native name translation to the existing DSO code
5775 that will convert (if the flag to do so is set) filenames
5776 that are sufficiently small and have no path information
5777 into a canonical native form. Eg. "blah" converted to
5778 "libblah.so" or "blah.dll" etc.
5781 *) New function ERR_error_string_n(e, buf, len) which is like
5782 ERR_error_string(e, buf), but writes at most 'len' bytes
5783 including the 0 terminator. For ERR_error_string_n, 'buf'
5785 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
5787 *) CONF library reworked to become more general. A new CONF
5788 configuration file reader "class" is implemented as well as a
5789 new functions (NCONF_*, for "New CONF") to handle it. The now
5790 old CONF_* functions are still there, but are reimplemented to
5791 work in terms of the new functions. Also, a set of functions
5792 to handle the internal storage of the configuration data is
5793 provided to make it easier to write new configuration file
5794 reader "classes" (I can definitely see something reading a
5795 configuration file in XML format, for example), called _CONF_*,
5796 or "the configuration storage API"...
5798 The new configuration file reading functions are:
5800 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
5801 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
5803 NCONF_default, NCONF_WIN32
5805 NCONF_dump_fp, NCONF_dump_bio
5807 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
5808 NCONF_new creates a new CONF object. This works in the same way
5809 as other interfaces in OpenSSL, like the BIO interface.
5810 NCONF_dump_* dump the internal storage of the configuration file,
5811 which is useful for debugging. All other functions take the same
5812 arguments as the old CONF_* functions wth the exception of the
5813 first that must be a `CONF *' instead of a `LHASH *'.
5815 To make it easer to use the new classes with the old CONF_* functions,
5816 the function CONF_set_default_method is provided.
5819 *) Add '-tls1' option to 'openssl ciphers', which was already
5820 mentioned in the documentation but had not been implemented.
5821 (This option is not yet really useful because even the additional
5822 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
5825 *) Initial DSO code added into libcrypto for letting OpenSSL (and
5826 OpenSSL-based applications) load shared libraries and bind to
5827 them in a portable way.
5828 [Geoff Thorpe, with contributions from Richard Levitte]
5830 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
5832 *) Make sure _lrotl and _lrotr are only used with MSVC.
5834 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
5835 (the default implementation of RAND_status).
5837 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
5838 to '-clrext' (= clear extensions), as intended and documented.
5839 [Bodo Moeller; inconsistency pointed out by Michael Attili
5842 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
5843 was larger than the MD block size.
5844 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
5846 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
5847 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
5848 using the passed key: if the passed key was a private key the result
5849 of X509_print(), for example, would be to print out all the private key
5853 *) des_quad_cksum() byte order bug fix.
5854 [Ulf Möller, using the problem description in krb4-0.9.7, where
5855 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
5857 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
5859 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
5861 *) For easily testing in shell scripts whether some command
5862 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
5863 returns with exit code 0 iff no command of the given name is available.
5864 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
5865 the output goes to stdout and nothing is printed to stderr.
5866 Additional arguments are always ignored.
5868 Since for each cipher there is a command of the same name,
5869 the 'no-cipher' compilation switches can be tested this way.
5871 ('openssl no-XXX' is not able to detect pseudo-commands such
5872 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
5875 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
5878 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
5879 is set; it will be thrown away anyway because each handshake creates
5881 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
5882 to parameters -- in previous versions (since OpenSSL 0.9.3) the
5883 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
5884 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
5887 *) New s_client option -ign_eof: EOF at stdin is ignored, and
5888 'Q' and 'R' lose their special meanings (quit/renegotiate).
5889 This is part of what -quiet does; unlike -quiet, -ign_eof
5890 does not suppress any output.
5893 *) Add compatibility options to the purpose and trust code. The
5894 purpose X509_PURPOSE_ANY is "any purpose" which automatically
5895 accepts a certificate or CA, this was the previous behaviour,
5896 with all the associated security issues.
5898 X509_TRUST_COMPAT is the old trust behaviour: only and
5899 automatically trust self signed roots in certificate store. A
5900 new trust setting X509_TRUST_DEFAULT is used to specify that
5901 a purpose has no associated trust setting and it should instead
5902 use the value in the default purpose.
5905 *) Fix the PKCS#8 DSA private key code so it decodes keys again
5906 and fix a memory leak.
5909 *) In util/mkerr.pl (which implements 'make errors'), preserve
5910 reason strings from the previous version of the .c file, as
5911 the default to have only downcase letters (and digits) in
5912 automatically generated reasons codes is not always appropriate.
5915 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
5916 using strerror. Previously, ERR_reason_error_string() returned
5917 library names as reason strings for SYSerr; but SYSerr is a special
5918 case where small numbers are errno values, not library numbers.
5921 *) Add '-dsaparam' option to 'openssl dhparam' application. This
5922 converts DSA parameters into DH parameters. (When creating parameters,
5923 DSA_generate_parameters is used.)
5926 *) Include 'length' (recommended exponent length) in C code generated
5927 by 'openssl dhparam -C'.
5930 *) The second argument to set_label in perlasm was already being used
5931 so couldn't be used as a "file scope" flag. Moved to third argument
5935 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
5936 instead of RAND_bytes for encryption IVs and salts.
5939 *) Include RAND_status() into RAND_METHOD instead of implementing
5940 it only for md_rand.c Otherwise replacing the PRNG by calling
5941 RAND_set_rand_method would be impossible.
5944 *) Don't let DSA_generate_key() enter an infinite loop if the random
5945 number generation fails.
5948 *) New 'rand' application for creating pseudo-random output.
5951 *) Added configuration support for Linux/IA64
5952 [Rolf Haberrecker <rolf@suse.de>]
5954 *) Assembler module support for Mingw32.
5957 *) Shared library support for HPUX (in shlib/).
5958 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
5960 *) Shared library support for Solaris gcc.
5961 [Lutz Behnke <behnke@trustcenter.de>]
5963 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
5965 *) PKCS7_encrypt() was adding text MIME headers twice because they
5966 were added manually and by SMIME_crlf_copy().
5969 *) In bntest.c don't call BN_rand with zero bits argument.
5970 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
5972 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
5973 case was implemented. This caused BN_div_recp() to fail occasionally.
5976 *) Add an optional second argument to the set_label() in the perl
5977 assembly language builder. If this argument exists and is set
5978 to 1 it signals that the assembler should use a symbol whose
5979 scope is the entire file, not just the current function. This
5980 is needed with MASM which uses the format label:: for this scope.
5981 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
5983 *) Change the ASN1 types so they are typedefs by default. Before
5984 almost all types were #define'd to ASN1_STRING which was causing
5985 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
5989 *) Change names of new functions to the new get1/get0 naming
5990 convention: After 'get1', the caller owns a reference count
5991 and has to call ..._free; 'get0' returns a pointer to some
5992 data structure without incrementing reference counters.
5993 (Some of the existing 'get' functions increment a reference
5994 counter, some don't.)
5995 Similarly, 'set1' and 'add1' functions increase reference
5996 counters or duplicate objects.
5999 *) Allow for the possibility of temp RSA key generation failure:
6000 the code used to assume it always worked and crashed on failure.
6003 *) Fix potential buffer overrun problem in BIO_printf().
6004 [Ulf Möller, using public domain code by Patrick Powell; problem
6005 pointed out by David Sacerdote <das33@cornell.edu>]
6007 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
6008 RAND_egd() and RAND_status(). In the command line application,
6009 the EGD socket can be specified like a seed file using RANDFILE
6013 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
6014 Some CAs (e.g. Verisign) distribute certificates in this form.
6017 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
6018 list to exclude them. This means that no special compilation option
6019 is needed to use anonymous DH: it just needs to be included in the
6023 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
6024 EVP_MD_type. The old functionality is available in a new macro called
6025 EVP_MD_md(). Change code that uses it and update docs.
6028 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
6029 where the 'void *' argument is replaced by a function pointer argument.
6030 Previously 'void *' was abused to point to functions, which works on
6031 many platforms, but is not correct. As these functions are usually
6032 called by macros defined in OpenSSL header files, most source code
6033 should work without changes.
6036 *) <openssl/opensslconf.h> (which is created by Configure) now contains
6037 sections with information on -D... compiler switches used for
6038 compiling the library so that applications can see them. To enable
6039 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
6040 must be defined. E.g.,
6041 #define OPENSSL_ALGORITHM_DEFINES
6042 #include <openssl/opensslconf.h>
6043 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
6044 [Richard Levitte, Ulf and Bodo Möller]
6046 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
6050 *) Change the 'other' type in certificate aux info to a STACK_OF
6051 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
6052 the required ASN1 format: arbitrary types determined by an OID.
6055 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
6056 argument to 'req'. This is not because the function is newer or
6057 better than others it just uses the work 'NEW' in the certificate
6058 request header lines. Some software needs this.
6061 *) Reorganise password command line arguments: now passwords can be
6062 obtained from various sources. Delete the PEM_cb function and make
6063 it the default behaviour: i.e. if the callback is NULL and the
6064 usrdata argument is not NULL interpret it as a null terminated pass
6065 phrase. If usrdata and the callback are NULL then the pass phrase
6066 is prompted for as usual.
6069 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
6070 the support is automatically enabled. The resulting binaries will
6071 autodetect the card and use it if present.
6072 [Ben Laurie and Compaq Inc.]
6074 *) Work around for Netscape hang bug. This sends certificate request
6075 and server done in one record. Since this is perfectly legal in the
6076 SSL/TLS protocol it isn't a "bug" option and is on by default. See
6077 the bugs/SSLv3 entry for more info.
6080 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
6083 *) Add -rand argument to smime and pkcs12 applications and read/write
6087 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
6090 *) Add command line password options to the remaining applications.
6093 *) Bug fix for BN_div_recp() for numerators with an even number of
6097 *) More tests in bntest.c, and changed test_bn output.
6100 *) ./config recognizes MacOS X now.
6103 *) Bug fix for BN_div() when the first words of num and divsor are
6104 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
6107 *) Add support for various broken PKCS#8 formats, and command line
6108 options to produce them.
6111 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
6112 get temporary BIGNUMs from a BN_CTX.
6115 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
6119 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
6120 include a #define from the old name to the new. The original intent
6121 was that statically linked binaries could for example just call
6122 SSLeay_add_all_ciphers() to just add ciphers to the table and not
6123 link with digests. This never worked becayse SSLeay_add_all_digests()
6124 and SSLeay_add_all_ciphers() were in the same source file so calling
6125 one would link with the other. They are now in separate source files.
6128 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
6131 *) Use a less unusual form of the Miller-Rabin primality test (it used
6132 a binary algorithm for exponentiation integrated into the Miller-Rabin
6133 loop, our standard modexp algorithms are faster).
6136 *) Support for the EBCDIC character set completed.
6137 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
6139 *) Source code cleanups: use const where appropriate, eliminate casts,
6140 use void * instead of char * in lhash.
6143 *) Bugfix: ssl3_send_server_key_exchange was not restartable
6144 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
6145 this the server could overwrite ephemeral keys that the client
6149 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
6150 using 50 iterations of the Rabin-Miller test.
6152 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
6153 iterations of the Rabin-Miller test as required by the appendix
6154 to FIPS PUB 186[-1]) instead of DSA_is_prime.
6155 As BN_is_prime_fasttest includes trial division, DSA parameter
6156 generation becomes much faster.
6158 This implies a change for the callback functions in DSA_is_prime
6159 and DSA_generate_parameters: The callback function is called once
6160 for each positive witness in the Rabin-Miller test, not just
6161 occasionally in the inner loop; and the parameters to the
6162 callback function now provide an iteration count for the outer
6163 loop rather than for the current invocation of the inner loop.
6164 DSA_generate_parameters additionally can call the callback
6165 function with an 'iteration count' of -1, meaning that a
6166 candidate has passed the trial division test (when q is generated
6167 from an application-provided seed, trial division is skipped).
6170 *) New function BN_is_prime_fasttest that optionally does trial
6171 division before starting the Rabin-Miller test and has
6172 an additional BN_CTX * argument (whereas BN_is_prime always
6173 has to allocate at least one BN_CTX).
6174 'callback(1, -1, cb_arg)' is called when a number has passed the
6175 trial division stage.
6178 *) Fix for bug in CRL encoding. The validity dates weren't being handled
6182 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
6185 *) New function BN_pseudo_rand().
6188 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
6189 bignum version of BN_from_montgomery() with the working code from
6190 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
6194 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
6195 made it impossible to use the same SSL_SESSION data structure in
6196 SSL2 clients in multiple threads.
6199 *) The return value of RAND_load_file() no longer counts bytes obtained
6200 by stat(). RAND_load_file(..., -1) is new and uses the complete file
6201 to seed the PRNG (previously an explicit byte count was required).
6202 [Ulf Möller, Bodo Möller]
6204 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
6205 used (char *) instead of (void *) and had casts all over the place.
6208 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
6211 *) Retain source code compatibility for BN_prime_checks macro:
6212 BN_is_prime(..., BN_prime_checks, ...) now uses
6213 BN_prime_checks_for_size to determine the appropriate number of
6214 Rabin-Miller iterations.
6217 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
6218 DH_CHECK_P_NOT_SAFE_PRIME.
6219 (Check if this is true? OpenPGP calls them "strong".)
6222 *) Merge the functionality of "dh" and "gendh" programs into a new program
6223 "dhparam". The old programs are retained for now but will handle DH keys
6224 (instead of parameters) in future.
6227 *) Make the ciphers, s_server and s_client programs check the return values
6228 when a new cipher list is set.
6231 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
6232 ciphers. Before when the 56bit ciphers were enabled the sorting was
6235 The syntax for the cipher sorting has been extended to support sorting by
6236 cipher-strength (using the strength_bits hard coded in the tables).
6237 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
6239 Fix a bug in the cipher-command parser: when supplying a cipher command
6240 string with an "undefined" symbol (neither command nor alphanumeric
6241 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
6242 an error is flagged.
6244 Due to the strength-sorting extension, the code of the
6245 ssl_create_cipher_list() function was completely rearranged. I hope that
6246 the readability was also increased :-)
6247 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6249 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
6250 for the first serial number and places 2 in the serial number file. This
6251 avoids problems when the root CA is created with serial number zero and
6252 the first user certificate has the same issuer name and serial number
6256 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
6257 the new code. Add documentation for this stuff.
6260 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
6261 X509_*() to X509at_*() on the grounds that they don't handle X509
6262 structures and behave in an analagous way to the X509v3 functions:
6263 they shouldn't be called directly but wrapper functions should be used
6266 So we also now have some wrapper functions that call the X509at functions
6267 when passed certificate requests. (TO DO: similar things can be done with
6268 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
6269 things. Some of these need some d2i or i2d and print functionality
6270 because they handle more complex structures.)
6273 *) Add missing #ifndefs that caused missing symbols when building libssl
6274 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
6275 NO_RSA in ssl/s2*.c.
6276 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
6278 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
6279 has a return value which indicates the quality of the random data
6280 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
6281 error queue. New function RAND_pseudo_bytes() generates output that is
6282 guaranteed to be unique but not unpredictable. RAND_add is like
6283 RAND_seed, but takes an extra argument for an entropy estimate
6284 (RAND_seed always assumes full entropy).
6287 *) Do more iterations of Rabin-Miller probable prime test (specifically,
6288 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
6289 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
6290 in crypto/bn/bn_prime.c for the complete table). This guarantees a
6291 false-positive rate of at most 2^-80 for random input.
6294 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
6297 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
6298 in the 0.9.5 release), this returns the chain
6299 from an X509_CTX structure with a dup of the stack and all
6300 the X509 reference counts upped: so the stack will exist
6301 after X509_CTX_cleanup() has been called. Modify pkcs12.c
6304 Also make SSL_SESSION_print() print out the verify return
6308 *) Add manpage for the pkcs12 command. Also change the default
6309 behaviour so MAC iteration counts are used unless the new
6310 -nomaciter option is used. This improves file security and
6311 only older versions of MSIE (4.0 for example) need it.
6314 *) Honor the no-xxx Configure options when creating .DEF files.
6317 *) Add PKCS#10 attributes to field table: challengePassword,
6318 unstructuredName and unstructuredAddress. These are taken from
6319 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
6320 international characters are used.
6322 More changes to X509_ATTRIBUTE code: allow the setting of types
6323 based on strings. Remove the 'loc' parameter when adding
6324 attributes because these will be a SET OF encoding which is sorted
6328 *) Initial changes to the 'req' utility to allow request generation
6329 automation. This will allow an application to just generate a template
6330 file containing all the field values and have req construct the
6333 Initial support for X509_ATTRIBUTE handling. Stacks of these are
6334 used all over the place including certificate requests and PKCS#7
6335 structures. They are currently handled manually where necessary with
6336 some primitive wrappers for PKCS#7. The new functions behave in a
6337 manner analogous to the X509 extension functions: they allow
6338 attributes to be looked up by NID and added.
6340 Later something similar to the X509V3 code would be desirable to
6341 automatically handle the encoding, decoding and printing of the
6342 more complex types. The string types like challengePassword can
6343 be handled by the string table functions.
6345 Also modified the multi byte string table handling. Now there is
6346 a 'global mask' which masks out certain types. The table itself
6347 can use the flag STABLE_NO_MASK to ignore the mask setting: this
6348 is useful when for example there is only one permissible type
6349 (as in countryName) and using the mask might result in no valid
6353 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
6354 SSL_get_peer_finished to allow applications to obtain the latest
6355 Finished messages sent to the peer or expected from the peer,
6356 respectively. (SSL_get_peer_finished is usually the Finished message
6357 actually received from the peer, otherwise the protocol will be aborted.)
6359 As the Finished message are message digests of the complete handshake
6360 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
6361 be used for external authentication procedures when the authentication
6362 provided by SSL/TLS is not desired or is not enough.
6365 *) Enhanced support for Alpha Linux is added. Now ./config checks if
6366 the host supports BWX extension and if Compaq C is present on the
6367 $PATH. Just exploiting of the BWX extension results in 20-30%
6368 performance kick for some algorithms, e.g. DES and RC4 to mention
6369 a couple. Compaq C in turn generates ~20% faster code for MD5 and
6373 *) Add support for MS "fast SGC". This is arguably a violation of the
6374 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
6375 weak crypto and after checking the certificate is SGC a second one
6376 with strong crypto. MS SGC stops the first handshake after receiving
6377 the server certificate message and sends a second client hello. Since
6378 a server will typically do all the time consuming operations before
6379 expecting any further messages from the client (server key exchange
6380 is the most expensive) there is little difference between the two.
6382 To get OpenSSL to support MS SGC we have to permit a second client
6383 hello message after we have sent server done. In addition we have to
6384 reset the MAC if we do get this second client hello.
6387 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
6388 if a DER encoded private key is RSA or DSA traditional format. Changed
6389 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
6390 format DER encoded private key. Newer code should use PKCS#8 format which
6391 has the key type encoded in the ASN1 structure. Added DER private key
6392 support to pkcs8 application.
6395 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
6396 ciphersuites has been selected (as required by the SSL 3/TLS 1
6397 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
6398 is set, we interpret this as a request to violate the specification
6399 (the worst that can happen is a handshake failure, and 'correct'
6400 behaviour would result in a handshake failure anyway).
6403 *) In SSL_CTX_add_session, take into account that there might be multiple
6404 SSL_SESSION structures with the same session ID (e.g. when two threads
6405 concurrently obtain them from an external cache).
6406 The internal cache can handle only one SSL_SESSION with a given ID,
6407 so if there's a conflict, we now throw out the old one to achieve
6411 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
6412 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
6413 some routines that use cipher OIDs: some ciphers do not have OIDs
6414 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
6418 *) Simplify the trust setting structure and code. Now we just have
6419 two sequences of OIDs for trusted and rejected settings. These will
6420 typically have values the same as the extended key usage extension
6421 and any application specific purposes.
6423 The trust checking code now has a default behaviour: it will just
6424 check for an object with the same NID as the passed id. Functions can
6425 be provided to override either the default behaviour or the behaviour
6426 for a given id. SSL client, server and email already have functions
6427 in place for compatibility: they check the NID and also return "trusted"
6428 if the certificate is self signed.
6431 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
6432 traditional format into an EVP_PKEY structure.
6435 *) Add a password callback function PEM_cb() which either prompts for
6436 a password if usr_data is NULL or otherwise assumes it is a null
6437 terminated password. Allow passwords to be passed on command line
6438 environment or config files in a few more utilities.
6441 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
6442 keys. Add some short names for PKCS#8 PBE algorithms and allow them
6443 to be specified on the command line for the pkcs8 and pkcs12 utilities.
6444 Update documentation.
6447 *) Support for ASN1 "NULL" type. This could be handled before by using
6448 ASN1_TYPE but there wasn't any function that would try to read a NULL
6449 and produce an error if it couldn't. For compatibility we also have
6450 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
6451 don't allocate anything because they don't need to.
6454 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
6456 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
6458 *) Rebuild of the memory allocation routines used by OpenSSL code and
6459 possibly others as well. The purpose is to make an interface that
6460 provide hooks so anyone can build a separate set of allocation and
6461 deallocation routines to be used by OpenSSL, for example memory
6462 pool implementations, or something else, which was previously hard
6463 since Malloc(), Realloc() and Free() were defined as macros having
6464 the values malloc, realloc and free, respectively (except for Win32
6465 compilations). The same is provided for memory debugging code.
6466 OpenSSL already comes with functionality to find memory leaks, but
6467 this gives people a chance to debug other memory problems.
6469 With these changes, a new set of functions and macros have appeared:
6471 CRYPTO_set_mem_debug_functions() [F]
6472 CRYPTO_get_mem_debug_functions() [F]
6473 CRYPTO_dbg_set_options() [F]
6474 CRYPTO_dbg_get_options() [F]
6475 CRYPTO_malloc_debug_init() [M]
6477 The memory debug functions are NULL by default, unless the library
6478 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
6479 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
6480 gives the standard debugging functions that come with OpenSSL) or
6481 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
6482 provided by the library user) must be used. When the standard
6483 debugging functions are used, CRYPTO_dbg_set_options can be used to
6484 request additional information:
6485 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
6486 the CRYPTO_MDEBUG_xxx macro when compiling the library.
6488 Also, things like CRYPTO_set_mem_functions will always give the
6489 expected result (the new set of functions is used for allocation
6490 and deallocation) at all times, regardless of platform and compiler
6493 To finish it up, some functions that were never use in any other
6494 way than through macros have a new API and new semantic:
6497 CRYPTO_dbg_realloc()
6500 All macros of value have retained their old syntax.
6501 [Richard Levitte and Bodo Moeller]
6503 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
6504 ordering of SMIMECapabilities wasn't in "strength order" and there
6505 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
6509 *) Some ASN1 types with illegal zero length encoding (INTEGER,
6510 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
6511 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
6513 *) Merge in my S/MIME library for OpenSSL. This provides a simple
6514 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
6515 functionality to handle multipart/signed properly) and a utility
6516 called 'smime' to call all this stuff. This is based on code I
6517 originally wrote for Celo who have kindly allowed it to be
6518 included in OpenSSL.
6521 *) Add variants des_set_key_checked and des_set_key_unchecked of
6522 des_set_key (aka des_key_sched). Global variable des_check_key
6523 decides which of these is called by des_set_key; this way
6524 des_check_key behaves as it always did, but applications and
6525 the library itself, which was buggy for des_check_key == 1,
6526 have a cleaner way to pick the version they need.
6529 *) New function PKCS12_newpass() which changes the password of a
6533 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
6534 dynamic mix. In both cases the ids can be used as an index into the
6535 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
6536 functions so they accept a list of the field values and the
6537 application doesn't need to directly manipulate the X509_TRUST
6541 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
6545 *) Modify the way the V3 extension code looks up extensions. This now
6546 works in a similar way to the object code: we have some "standard"
6547 extensions in a static table which is searched with OBJ_bsearch()
6548 and the application can add dynamic ones if needed. The file
6549 crypto/x509v3/ext_dat.h now has the info: this file needs to be
6550 updated whenever a new extension is added to the core code and kept
6551 in ext_nid order. There is a simple program 'tabtest.c' which checks
6552 this. New extensions are not added too often so this file can readily
6553 be maintained manually.
6555 There are two big advantages in doing things this way. The extensions
6556 can be looked up immediately and no longer need to be "added" using
6557 X509V3_add_standard_extensions(): this function now does nothing.
6558 [Side note: I get *lots* of email saying the extension code doesn't
6559 work because people forget to call this function]
6560 Also no dynamic allocation is done unless new extensions are added:
6561 so if we don't add custom extensions there is no need to call
6562 X509V3_EXT_cleanup().
6565 *) Modify enc utility's salting as follows: make salting the default. Add a
6566 magic header, so unsalted files fail gracefully instead of just decrypting
6567 to garbage. This is because not salting is a big security hole, so people
6568 should be discouraged from doing it.
6571 *) Fixes and enhancements to the 'x509' utility. It allowed a message
6572 digest to be passed on the command line but it only used this
6573 parameter when signing a certificate. Modified so all relevant
6574 operations are affected by the digest parameter including the
6575 -fingerprint and -x509toreq options. Also -x509toreq choked if a
6576 DSA key was used because it didn't fix the digest.
6579 *) Initial certificate chain verify code. Currently tests the untrusted
6580 certificates for consistency with the verify purpose (which is set
6581 when the X509_STORE_CTX structure is set up) and checks the pathlength.
6583 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
6584 this is because it will reject chains with invalid extensions whereas
6585 every previous version of OpenSSL and SSLeay made no checks at all.
6587 Trust code: checks the root CA for the relevant trust settings. Trust
6588 settings have an initial value consistent with the verify purpose: e.g.
6589 if the verify purpose is for SSL client use it expects the CA to be
6590 trusted for SSL client use. However the default value can be changed to
6591 permit custom trust settings: one example of this would be to only trust
6592 certificates from a specific "secure" set of CAs.
6594 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
6595 which should be used for version portability: especially since the
6596 verify structure is likely to change more often now.
6598 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
6599 to set them. If not set then assume SSL clients will verify SSL servers
6602 Two new options to the verify program: -untrusted allows a set of
6603 untrusted certificates to be passed in and -purpose which sets the
6604 intended purpose of the certificate. If a purpose is set then the
6605 new chain verify code is used to check extension consistency.
6608 *) Support for the authority information access extension.
6611 *) Modify RSA and DSA PEM read routines to transparently handle
6612 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
6613 public keys in a format compatible with certificate
6614 SubjectPublicKeyInfo structures. Unfortunately there were already
6615 functions called *_PublicKey_* which used various odd formats so
6616 these are retained for compatibility: however the DSA variants were
6617 never in a public release so they have been deleted. Changed dsa/rsa
6618 utilities to handle the new format: note no releases ever handled public
6619 keys so we should be OK.
6621 The primary motivation for this change is to avoid the same fiasco
6622 that dogs private keys: there are several incompatible private key
6623 formats some of which are standard and some OpenSSL specific and
6624 require various evil hacks to allow partial transparent handling and
6625 even then it doesn't work with DER formats. Given the option anything
6626 other than PKCS#8 should be dumped: but the other formats have to
6627 stay in the name of compatibility.
6629 With public keys and the benefit of hindsight one standard format
6630 is used which works with EVP_PKEY, RSA or DSA structures: though
6631 it clearly returns an error if you try to read the wrong kind of key.
6633 Added a -pubkey option to the 'x509' utility to output the public key.
6634 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
6635 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
6636 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
6637 that do the same as the EVP_PKEY_assign_*() except they up the
6638 reference count of the added key (they don't "swallow" the
6642 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
6643 CRLs would fail if the file contained no certificates or no CRLs:
6644 added a new function to read in both types and return the number
6645 read: this means that if none are read it will be an error. The
6646 DER versions of the certificate and CRL reader would always fail
6647 because it isn't possible to mix certificates and CRLs in DER format
6648 without choking one or the other routine. Changed this to just read
6649 a certificate: this is the best we can do. Also modified the code
6650 in apps/verify.c to take notice of return codes: it was previously
6651 attempting to read in certificates from NULL pointers and ignoring
6652 any errors: this is one reason why the cert and CRL reader seemed
6653 to work. It doesn't check return codes from the default certificate
6654 routines: these may well fail if the certificates aren't installed.
6657 *) Code to support otherName option in GeneralName.
6660 *) First update to verify code. Change the verify utility
6661 so it warns if it is passed a self signed certificate:
6662 for consistency with the normal behaviour. X509_verify
6663 has been modified to it will now verify a self signed
6664 certificate if *exactly* the same certificate appears
6665 in the store: it was previously impossible to trust a
6666 single self signed certificate. This means that:
6667 openssl verify ss.pem
6668 now gives a warning about a self signed certificate but
6669 openssl verify -CAfile ss.pem ss.pem
6673 *) For servers, store verify_result in SSL_SESSION data structure
6674 (and add it to external session representation).
6675 This is needed when client certificate verifications fails,
6676 but an application-provided verification callback (set by
6677 SSL_CTX_set_cert_verify_callback) allows accepting the session
6678 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
6679 but returns 1): When the session is reused, we have to set
6680 ssl->verify_result to the appropriate error code to avoid
6682 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
6684 *) Fix a bug in the new PKCS#7 code: it didn't consider the
6685 case in PKCS7_dataInit() where the signed PKCS7 structure
6686 didn't contain any existing data because it was being created.
6687 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
6689 *) Add a salt to the key derivation routines in enc.c. This
6690 forms the first 8 bytes of the encrypted file. Also add a
6691 -S option to allow a salt to be input on the command line.
6694 *) New function X509_cmp(). Oddly enough there wasn't a function
6695 to compare two certificates. We do this by working out the SHA1
6696 hash and comparing that. X509_cmp() will be needed by the trust
6700 *) SSL_get1_session() is like SSL_get_session(), but increments
6701 the reference count in the SSL_SESSION returned.
6702 [Geoff Thorpe <geoff@eu.c2.net>]
6704 *) Fix for 'req': it was adding a null to request attributes.
6705 Also change the X509_LOOKUP and X509_INFO code to handle
6706 certificate auxiliary information.
6709 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
6713 *) Add the possibility to add extra information to the memory leak
6714 detecting output, to form tracebacks, showing from where each
6715 allocation was originated: CRYPTO_push_info("constant string") adds
6716 the string plus current file name and line number to a per-thread
6717 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
6718 is like calling CYRPTO_pop_info() until the stack is empty.
6719 Also updated memory leak detection code to be multi-thread-safe.
6722 *) Add options -text and -noout to pkcs7 utility and delete the
6723 encryption options which never did anything. Update docs.
6726 *) Add options to some of the utilities to allow the pass phrase
6727 to be included on either the command line (not recommended on
6728 OSes like Unix) or read from the environment. Update the
6729 manpages and fix a few bugs.
6732 *) Add a few manpages for some of the openssl commands.
6735 *) Fix the -revoke option in ca. It was freeing up memory twice,
6736 leaking and not finding already revoked certificates.
6739 *) Extensive changes to support certificate auxiliary information.
6740 This involves the use of X509_CERT_AUX structure and X509_AUX
6741 functions. An X509_AUX function such as PEM_read_X509_AUX()
6742 can still read in a certificate file in the usual way but it
6743 will also read in any additional "auxiliary information". By
6744 doing things this way a fair degree of compatibility can be
6745 retained: existing certificates can have this information added
6746 using the new 'x509' options.
6748 Current auxiliary information includes an "alias" and some trust
6749 settings. The trust settings will ultimately be used in enhanced
6750 certificate chain verification routines: currently a certificate
6751 can only be trusted if it is self signed and then it is trusted
6755 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
6756 The problem was that one of the replacement routines had not been working
6757 since SSLeay releases. For now the offending routine has been replaced
6758 with non-optimised assembler. Even so, this now gives around 95%
6759 performance improvement for 1024 bit RSA signs.
6762 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
6763 handling. Most clients have the effective key size in bits equal to
6764 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
6765 A few however don't do this and instead use the size of the decrypted key
6766 to determine the RC2 key length and the AlgorithmIdentifier to determine
6767 the effective key length. In this case the effective key length can still
6768 be 40 bits but the key length can be 168 bits for example. This is fixed
6769 by manually forcing an RC2 key into the EVP_PKEY structure because the
6770 EVP code can't currently handle unusual RC2 key sizes: it always assumes
6771 the key length and effective key length are equal.
6774 *) Add a bunch of functions that should simplify the creation of
6775 X509_NAME structures. Now you should be able to do:
6776 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
6777 and have it automatically work out the correct field type and fill in
6778 the structures. The more adventurous can try:
6779 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
6780 and it will (hopefully) work out the correct multibyte encoding.
6783 *) Change the 'req' utility to use the new field handling and multibyte
6784 copy routines. Before the DN field creation was handled in an ad hoc
6785 way in req, ca, and x509 which was rather broken and didn't support
6786 BMPStrings or UTF8Strings. Since some software doesn't implement
6787 BMPStrings or UTF8Strings yet, they can be enabled using the config file
6788 using the dirstring_type option. See the new comment in the default
6789 openssl.cnf for more info.
6792 *) Make crypto/rand/md_rand.c more robust:
6793 - Assure unique random numbers after fork().
6794 - Make sure that concurrent threads access the global counter and
6795 md serializably so that we never lose entropy in them
6796 or use exactly the same state in multiple threads.
6797 Access to the large state is not always serializable because
6798 the additional locking could be a performance killer, and
6799 md should be large enough anyway.
6802 *) New file apps/app_rand.c with commonly needed functionality
6803 for handling the random seed file.
6805 Use the random seed file in some applications that previously did not:
6807 dsaparam -genkey (which also ignored its '-rand' option),
6810 x509 (when signing).
6811 Except on systems with /dev/urandom, it is crucial to have a random
6812 seed file at least for key creation, DSA signing, and for DH exchanges;
6813 for RSA signatures we could do without one.
6815 gendh and gendsa (unlike genrsa) used to read only the first byte
6816 of each file listed in the '-rand' option. The function as previously
6817 found in genrsa is now in app_rand.c and is used by all programs
6818 that support '-rand'.
6821 *) In RAND_write_file, use mode 0600 for creating files;
6822 don't just chmod when it may be too late.
6825 *) Report an error from X509_STORE_load_locations
6826 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
6829 *) New function ASN1_mbstring_copy() this copies a string in either
6830 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
6831 into an ASN1_STRING type. A mask of permissible types is passed
6832 and it chooses the "minimal" type to use or an error if not type
6836 *) Add function equivalents to the various macros in asn1.h. The old
6837 macros are retained with an M_ prefix. Code inside the library can
6838 use the M_ macros. External code (including the openssl utility)
6839 should *NOT* in order to be "shared library friendly".
6842 *) Add various functions that can check a certificate's extensions
6843 to see if it usable for various purposes such as SSL client,
6844 server or S/MIME and CAs of these types. This is currently
6845 VERY EXPERIMENTAL but will ultimately be used for certificate chain
6846 verification. Also added a -purpose flag to x509 utility to
6847 print out all the purposes.
6850 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
6854 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
6855 for, obtain and decode and extension and obtain its critical flag.
6856 This allows all the necessary extension code to be handled in a
6857 single function call.
6860 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
6861 platforms. See crypto/rc4/rc4_enc.c for further details.
6864 *) New -noout option to asn1parse. This causes no output to be produced
6865 its main use is when combined with -strparse and -out to extract data
6866 from a file (which may not be in ASN.1 format).
6869 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
6870 when producing the local key id.
6871 [Richard Levitte <levitte@stacken.kth.se>]
6873 *) New option -dhparam in s_server. This allows a DH parameter file to be
6874 stated explicitly. If it is not stated then it tries the first server
6875 certificate file. The previous behaviour hard coded the filename
6879 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
6880 a public key to be input or output. For example:
6881 openssl rsa -in key.pem -pubout -out pubkey.pem
6882 Also added necessary DSA public key functions to handle this.
6885 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
6886 in the message. This was handled by allowing
6887 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
6888 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
6890 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
6891 to the end of the strings whereas this didn't. This would cause problems
6892 if strings read with d2i_ASN1_bytes() were later modified.
6893 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
6895 *) Fix for base64 decode bug. When a base64 bio reads only one line of
6896 data and it contains EOF it will end up returning an error. This is
6897 caused by input 46 bytes long. The cause is due to the way base64
6898 BIOs find the start of base64 encoded data. They do this by trying a
6899 trial decode on each line until they find one that works. When they
6900 do a flag is set and it starts again knowing it can pass all the
6901 data directly through the decoder. Unfortunately it doesn't reset
6902 the context it uses. This means that if EOF is reached an attempt
6903 is made to pass two EOFs through the context and this causes the
6904 resulting error. This can also cause other problems as well. As is
6905 usual with these problems it takes *ages* to find and the fix is
6906 trivial: move one line.
6907 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
6909 *) Ugly workaround to get s_client and s_server working under Windows. The
6910 old code wouldn't work because it needed to select() on sockets and the
6911 tty (for keypresses and to see if data could be written). Win32 only
6912 supports select() on sockets so we select() with a 1s timeout on the
6913 sockets and then see if any characters are waiting to be read, if none
6914 are present then we retry, we also assume we can always write data to
6915 the tty. This isn't nice because the code then blocks until we've
6916 received a complete line of data and it is effectively polling the
6917 keyboard at 1s intervals: however it's quite a bit better than not
6918 working at all :-) A dedicated Windows application might handle this
6919 with an event loop for example.
6922 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
6923 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
6924 will be called when RSA_sign() and RSA_verify() are used. This is useful
6925 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
6926 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
6927 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
6928 This necessitated the support of an extra signature type NID_md5_sha1
6929 for SSL signatures and modifications to the SSL library to use it instead
6930 of calling RSA_public_decrypt() and RSA_private_encrypt().
6933 *) Add new -verify -CAfile and -CApath options to the crl program, these
6934 will lookup a CRL issuers certificate and verify the signature in a
6935 similar way to the verify program. Tidy up the crl program so it
6936 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
6937 less strict. It will now permit CRL extensions even if it is not
6938 a V2 CRL: this will allow it to tolerate some broken CRLs.
6941 *) Initialize all non-automatic variables each time one of the openssl
6942 sub-programs is started (this is necessary as they may be started
6943 multiple times from the "OpenSSL>" prompt).
6944 [Lennart Bang, Bodo Moeller]
6946 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
6947 removing all other RSA functionality (this is what NO_RSA does). This
6948 is so (for example) those in the US can disable those operations covered
6949 by the RSA patent while allowing storage and parsing of RSA keys and RSA
6953 *) Non-copying interface to BIO pairs.
6954 (still largely untested)
6957 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
6958 ASCII string. This was handled independently in various places before.
6961 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
6962 UTF8 strings a character at a time.
6965 *) Use client_version from client hello to select the protocol
6966 (s23_srvr.c) and for RSA client key exchange verification
6967 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
6970 *) Add various utility functions to handle SPKACs, these were previously
6971 handled by poking round in the structure internals. Added new function
6972 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
6973 print, verify and generate SPKACs. Based on an original idea from
6974 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
6977 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
6980 *) Allow the config file extension section to be overwritten on the
6981 command line. Based on an original idea from Massimiliano Pala
6982 <madwolf@comune.modena.it>. The new option is called -extensions
6983 and can be applied to ca, req and x509. Also -reqexts to override
6984 the request extensions in req and -crlexts to override the crl extensions
6988 *) Add new feature to the SPKAC handling in ca. Now you can include
6989 the same field multiple times by preceding it by "XXXX." for example:
6992 this is the same syntax as used in the req config file.
6995 *) Allow certificate extensions to be added to certificate requests. These
6996 are specified in a 'req_extensions' option of the req section of the
6997 config file. They can be printed out with the -text option to req but
6998 are otherwise ignored at present.
7001 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
7002 data read consists of only the final block it would not decrypted because
7003 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
7004 A misplaced 'break' also meant the decrypted final block might not be
7005 copied until the next read.
7008 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
7009 a few extra parameters to the DH structure: these will be useful if
7010 for example we want the value of 'q' or implement X9.42 DH.
7013 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
7014 provides hooks that allow the default DSA functions or functions on a
7015 "per key" basis to be replaced. This allows hardware acceleration and
7016 hardware key storage to be handled without major modification to the
7017 library. Also added low level modexp hooks and CRYPTO_EX structure and
7018 associated functions.
7021 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
7022 as "read only": it can't be written to and the buffer it points to will
7023 not be freed. Reading from a read only BIO is much more efficient than
7024 a normal memory BIO. This was added because there are several times when
7025 an area of memory needs to be read from a BIO. The previous method was
7026 to create a memory BIO and write the data to it, this results in two
7027 copies of the data and an O(n^2) reading algorithm. There is a new
7028 function BIO_new_mem_buf() which creates a read only memory BIO from
7029 an area of memory. Also modified the PKCS#7 routines to use read only
7033 *) Bugfix: ssl23_get_client_hello did not work properly when called in
7034 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
7035 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
7036 but a retry condition occured while trying to read the rest.
7039 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
7040 NID_pkcs7_encrypted by default: this was wrong since this should almost
7041 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
7042 the encrypted data type: this is a more sensible place to put it and it
7043 allows the PKCS#12 code to be tidied up that duplicated this
7047 *) Changed obj_dat.pl script so it takes its input and output files on
7048 the command line. This should avoid shell escape redirection problems
7052 *) Initial support for certificate extension requests, these are included
7053 in things like Xenroll certificate requests. Included functions to allow
7054 extensions to be obtained and added.
7057 *) -crlf option to s_client and s_server for sending newlines as
7058 CRLF (as required by many protocols).
7061 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
7063 *) Install libRSAglue.a when OpenSSL is built with RSAref.
7064 [Ralf S. Engelschall]
7066 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
7067 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
7069 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
7073 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
7074 DH parameters/keys (q is lost during that conversion, but the resulting
7075 DH parameters contain its length).
7077 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
7078 much faster than DH_generate_parameters (which creates parameters
7079 where p = 2*q + 1), and also the smaller q makes DH computations
7080 much more efficient (160-bit exponentiation instead of 1024-bit
7081 exponentiation); so this provides a convenient way to support DHE
7082 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
7083 utter importance to use
7084 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7086 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7087 when such DH parameters are used, because otherwise small subgroup
7088 attacks may become possible!
7091 *) Avoid memory leak in i2d_DHparams.
7094 *) Allow the -k option to be used more than once in the enc program:
7095 this allows the same encrypted message to be read by multiple recipients.
7098 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
7099 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
7100 it will always use the numerical form of the OID, even if it has a short
7104 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
7105 method only got called if p,q,dmp1,dmq1,iqmp components were present,
7106 otherwise bn_mod_exp was called. In the case of hardware keys for example
7107 no private key components need be present and it might store extra data
7108 in the RSA structure, which cannot be accessed from bn_mod_exp.
7109 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
7110 private key operations.
7113 *) Added support for SPARC Linux.
7116 *) pem_password_cb function type incompatibly changed from
7117 typedef int pem_password_cb(char *buf, int size, int rwflag);
7119 ....(char *buf, int size, int rwflag, void *userdata);
7120 so that applications can pass data to their callbacks:
7121 The PEM[_ASN1]_{read,write}... functions and macros now take an
7122 additional void * argument, which is just handed through whenever
7123 the password callback is called.
7124 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
7126 New function SSL_CTX_set_default_passwd_cb_userdata.
7128 Compatibility note: As many C implementations push function arguments
7129 onto the stack in reverse order, the new library version is likely to
7130 interoperate with programs that have been compiled with the old
7131 pem_password_cb definition (PEM_whatever takes some data that
7132 happens to be on the stack as its last argument, and the callback
7133 just ignores this garbage); but there is no guarantee whatsoever that
7136 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
7137 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
7138 problems not only on Windows, but also on some Unix platforms.
7139 To avoid problematic command lines, these definitions are now in an
7140 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
7141 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
7144 *) MIPS III/IV assembler module is reimplemented.
7147 *) More DES library cleanups: remove references to srand/rand and
7148 delete an unused file.
7151 *) Add support for the the free Netwide assembler (NASM) under Win32,
7152 since not many people have MASM (ml) and it can be hard to obtain.
7153 This is currently experimental but it seems to work OK and pass all
7154 the tests. Check out INSTALL.W32 for info.
7157 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
7158 without temporary keys kept an extra copy of the server key,
7159 and connections with temporary keys did not free everything in case
7163 *) New function RSA_check_key and new openssl rsa option -check
7164 for verifying the consistency of RSA keys.
7165 [Ulf Moeller, Bodo Moeller]
7167 *) Various changes to make Win32 compile work:
7168 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
7169 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
7170 comparison" warnings.
7171 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
7174 *) Add a debugging option to PKCS#5 v2 key generation function: when
7175 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
7176 derived keys are printed to stderr.
7179 *) Copy the flags in ASN1_STRING_dup().
7180 [Roman E. Pavlov <pre@mo.msk.ru>]
7182 *) The x509 application mishandled signing requests containing DSA
7183 keys when the signing key was also DSA and the parameters didn't match.
7185 It was supposed to omit the parameters when they matched the signing key:
7186 the verifying software was then supposed to automatically use the CA's
7187 parameters if they were absent from the end user certificate.
7189 Omitting parameters is no longer recommended. The test was also
7190 the wrong way round! This was probably due to unusual behaviour in
7191 EVP_cmp_parameters() which returns 1 if the parameters match.
7192 This meant that parameters were omitted when they *didn't* match and
7193 the certificate was useless. Certificates signed with 'ca' didn't have
7195 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
7197 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
7198 The interface is as follows:
7199 Applications can use
7200 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
7201 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
7202 "off" is now the default.
7203 The library internally uses
7204 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
7205 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
7206 to disable memory-checking temporarily.
7208 Some inconsistent states that previously were possible (and were
7209 even the default) are now avoided.
7211 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
7212 with each memory chunk allocated; this is occasionally more helpful
7213 than just having a counter.
7215 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
7217 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
7221 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
7222 which largely parallels "options", but is for changing API behaviour,
7223 whereas "options" are about protocol behaviour.
7224 Initial "mode" flags are:
7226 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
7227 a single record has been written.
7228 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
7229 retries use the same buffer location.
7230 (But all of the contents must be
7234 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
7237 *) Fix problems with no-hmac etc.
7238 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
7240 *) New functions RSA_get_default_method(), RSA_set_method() and
7241 RSA_get_method(). These allows replacement of RSA_METHODs without having
7242 to mess around with the internals of an RSA structure.
7245 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
7246 Also really enable memory leak checks in openssl.c and in some
7248 [Chad C. Mulligan, Bodo Moeller]
7250 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
7251 up the length of negative integers. This has now been simplified to just
7252 store the length when it is first determined and use it later, rather
7253 than trying to keep track of where data is copied and updating it to
7255 [Steve Henson, reported by Brien Wheeler
7256 <bwheeler@authentica-security.com>]
7258 *) Add a new function PKCS7_signatureVerify. This allows the verification
7259 of a PKCS#7 signature but with the signing certificate passed to the
7260 function itself. This contrasts with PKCS7_dataVerify which assumes the
7261 certificate is present in the PKCS#7 structure. This isn't always the
7262 case: certificates can be omitted from a PKCS#7 structure and be
7263 distributed by "out of band" means (such as a certificate database).
7266 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
7267 function prototypes in pem.h, also change util/mkdef.pl to add the
7268 necessary function names.
7271 *) mk1mf.pl (used by Windows builds) did not properly read the
7272 options set by Configure in the top level Makefile, and Configure
7273 was not even able to write more than one option correctly.
7274 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
7277 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
7278 file to be loaded from a BIO or FILE pointer. The BIO version will
7279 for example allow memory BIOs to contain config info.
7282 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
7283 Whoever hopes to achieve shared-library compatibility across versions
7284 must use this, not the compile-time macro.
7285 (Exercise 0.9.4: Which is the minimum library version required by
7287 Note: All this applies only to multi-threaded programs, others don't
7291 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
7292 through a BIO pair triggered the default case, i.e.
7293 SSLerr(...,SSL_R_UNKNOWN_STATE).
7296 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
7297 can use the SSL library even if none of the specific BIOs is
7301 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
7302 for the encoded length.
7303 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
7305 *) Add initial documentation of the X509V3 functions.
7308 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
7309 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
7310 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
7311 secure PKCS#8 private key format with a high iteration count.
7314 *) Fix determination of Perl interpreter: A perl or perl5
7315 _directory_ in $PATH was also accepted as the interpreter.
7316 [Ralf S. Engelschall]
7318 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
7319 wrong with it but it was very old and did things like calling
7320 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
7324 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
7325 to use the new extension code.
7328 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
7329 with macros. This should make it easier to change their form, add extra
7330 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
7334 *) Add to configuration table a new entry that can specify an alternative
7335 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
7336 according to Mark Crispin <MRC@Panda.COM>.
7340 *) DES CBC did not update the IV. Weird.
7343 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
7344 Changing the behaviour of the former might break existing programs --
7345 where IV updating is needed, des_ncbc_encrypt can be used.
7348 *) When bntest is run from "make test" it drives bc to check its
7349 calculations, as well as internally checking them. If an internal check
7350 fails, it needs to cause bc to give a non-zero result or make test carries
7351 on without noticing the failure. Fixed.
7354 *) DES library cleanups.
7357 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
7358 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
7359 ciphers. NOTE: although the key derivation function has been verified
7360 against some published test vectors it has not been extensively tested
7361 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
7365 *) Instead of "mkdir -p", which is not fully portable, use new
7366 Perl script "util/mkdir-p.pl".
7369 *) Rewrite the way password based encryption (PBE) is handled. It used to
7370 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
7371 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
7372 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
7373 the 'parameter' field of the AlgorithmIdentifier is passed to the
7374 underlying key generation function so it must do its own ASN1 parsing.
7375 This has also changed the EVP_PBE_CipherInit() function which now has a
7376 'parameter' argument instead of literal salt and iteration count values
7377 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
7380 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
7381 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
7382 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
7383 KEY" because this clashed with PKCS#8 unencrypted string. Since this
7384 value was just used as a "magic string" and not used directly its
7385 value doesn't matter.
7388 *) Introduce some semblance of const correctness to BN. Shame C doesn't
7392 *) "linux-sparc64" configuration (ultrapenguin).
7393 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
7394 "linux-sparc" configuration.
7395 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
7397 *) config now generates no-xxx options for missing ciphers.
7400 *) Support the EBCDIC character set (work in progress).
7401 File ebcdic.c not yet included because it has a different license.
7402 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7404 *) Support BS2000/OSD-POSIX.
7405 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7407 *) Make callbacks for key generation use void * instead of char *.
7410 *) Make S/MIME samples compile (not yet tested).
7413 *) Additional typesafe stacks.
7416 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
7420 Changes between 0.9.3 and 0.9.3a [29 May 1999]
7422 *) New configuration variant "sco5-gcc".
7424 *) Updated some demos.
7425 [Sean O Riordain, Wade Scholine]
7427 *) Add missing BIO_free at exit of pkcs12 application.
7430 *) Fix memory leak in conf.c.
7433 *) Updates for Win32 to assembler version of MD5.
7436 *) Set #! path to perl in apps/der_chop to where we found it
7437 instead of using a fixed path.
7440 *) SHA library changes for irix64-mips4-cc.
7443 *) Improvements for VMS support.
7447 Changes between 0.9.2b and 0.9.3 [24 May 1999]
7449 *) Bignum library bug fix. IRIX 6 passes "make test" now!
7450 This also avoids the problems with SC4.2 and unpatched SC5.
7451 [Andy Polyakov <appro@fy.chalmers.se>]
7453 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
7454 These are required because of the typesafe stack would otherwise break
7455 existing code. If old code used a structure member which used to be STACK
7456 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
7457 sk_num or sk_value it would produce an error because the num, data members
7458 are not present in STACK_OF. Now it just produces a warning. sk_set
7459 replaces the old method of assigning a value to sk_value
7460 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
7461 that does this will no longer work (and should use sk_set instead) but
7462 this could be regarded as a "questionable" behaviour anyway.
7465 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
7466 correctly handle encrypted S/MIME data.
7469 *) Change type of various DES function arguments from des_cblock
7470 (which means, in function argument declarations, pointer to char)
7471 to des_cblock * (meaning pointer to array with 8 char elements),
7472 which allows the compiler to do more typechecking; it was like
7473 that back in SSLeay, but with lots of ugly casts.
7475 Introduce new type const_des_cblock.
7478 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
7479 problems: find RecipientInfo structure that matches recipient certificate
7480 and initialise the ASN1 structures properly based on passed cipher.
7483 *) Belatedly make the BN tests actually check the results.
7486 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
7487 to and from BNs: it was completely broken. New compilation option
7488 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
7489 key elements as negative integers.
7492 *) Reorganize and speed up MD5.
7493 [Andy Polyakov <appro@fy.chalmers.se>]
7496 [Richard Levitte <richard@levitte.org>]
7498 *) New option -out to asn1parse to allow the parsed structure to be
7499 output to a file. This is most useful when combined with the -strparse
7500 option to examine the output of things like OCTET STRINGS.
7503 *) Make SSL library a little more fool-proof by not requiring any longer
7504 that SSL_set_{accept,connect}_state be called before
7505 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
7506 in many applications because usually everything *appeared* to work as
7507 intended anyway -- now it really works as intended).
7510 *) Move openssl.cnf out of lib/.
7513 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
7514 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
7515 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
7516 [Ralf S. Engelschall]
7518 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
7519 handle PKCS#7 enveloped data properly.
7520 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
7522 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
7523 copying pointers. The cert_st handling is changed by this in
7524 various ways (and thus what used to be known as ctx->default_cert
7525 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
7526 any longer when s->cert does not give us what we need).
7527 ssl_cert_instantiate becomes obsolete by this change.
7528 As soon as we've got the new code right (possibly it already is?),
7529 we have solved a couple of bugs of the earlier code where s->cert
7530 was used as if it could not have been shared with other SSL structures.
7532 Note that using the SSL API in certain dirty ways now will result
7533 in different behaviour than observed with earlier library versions:
7534 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
7535 does not influence s as it used to.
7537 In order to clean up things more thoroughly, inside SSL_SESSION
7538 we don't use CERT any longer, but a new structure SESS_CERT
7539 that holds per-session data (if available); currently, this is
7540 the peer's certificate chain and, for clients, the server's certificate
7541 and temporary key. CERT holds only those values that can have
7542 meaningful defaults in an SSL_CTX.
7545 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
7546 from the internal representation. Various PKCS#7 fixes: remove some
7547 evil casts and set the enc_dig_alg field properly based on the signing
7551 *) Allow PKCS#12 password to be set from the command line or the
7552 environment. Let 'ca' get its config file name from the environment
7553 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
7557 *) Allow certificate policies extension to use an IA5STRING for the
7558 organization field. This is contrary to the PKIX definition but
7559 VeriSign uses it and IE5 only recognises this form. Document 'x509'
7563 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
7564 without disallowing inline assembler and the like for non-pedantic builds.
7567 *) Support Borland C++ builder.
7568 [Janez Jere <jj@void.si>, modified by Ulf Möller]
7573 *) SHA-1 cleanups and performance enhancements.
7574 [Andy Polyakov <appro@fy.chalmers.se>]
7576 *) Sparc v8plus assembler for the bignum library.
7577 [Andy Polyakov <appro@fy.chalmers.se>]
7579 *) Accept any -xxx and +xxx compiler options in Configure.
7582 *) Update HPUX configuration.
7585 *) Add missing sk_<type>_unshift() function to safestack.h
7586 [Ralf S. Engelschall]
7588 *) New function SSL_CTX_use_certificate_chain_file that sets the
7589 "extra_cert"s in addition to the certificate. (This makes sense
7590 only for "PEM" format files, as chains as a whole are not
7594 *) Support verify_depth from the SSL API.
7595 x509_vfy.c had what can be considered an off-by-one-error:
7596 Its depth (which was not part of the external interface)
7597 was actually counting the number of certificates in a chain;
7598 now it really counts the depth.
7601 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
7602 instead of X509err, which often resulted in confusing error
7603 messages since the error codes are not globally unique
7604 (e.g. an alleged error in ssl3_accept when a certificate
7605 didn't match the private key).
7607 *) New function SSL_CTX_set_session_id_context that allows to set a default
7608 value (so that you don't need SSL_set_session_id_context for each
7609 connection using the SSL_CTX).
7612 *) OAEP decoding bug fix.
7615 *) Support INSTALL_PREFIX for package builders, as proposed by
7619 *) New Configure options "threads" and "no-threads". For systems
7620 where the proper compiler options are known (currently Solaris
7621 and Linux), "threads" is the default.
7624 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
7627 *) Install various scripts to $(OPENSSLDIR)/misc, not to
7628 $(INSTALLTOP)/bin -- they shouldn't clutter directories
7629 such as /usr/local/bin.
7632 *) "make linux-shared" to build shared libraries.
7633 [Niels Poppe <niels@netbox.org>]
7635 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
7638 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
7639 extension adding in x509 utility.
7642 *) Remove NOPROTO sections and error code comments.
7645 *) Partial rewrite of the DEF file generator to now parse the ANSI
7649 *) New Configure options --prefix=DIR and --openssldir=DIR.
7652 *) Complete rewrite of the error code script(s). It is all now handled
7653 by one script at the top level which handles error code gathering,
7654 header rewriting and C source file generation. It should be much better
7655 than the old method: it now uses a modified version of Ulf's parser to
7656 read the ANSI prototypes in all header files (thus the old K&R definitions
7657 aren't needed for error creation any more) and do a better job of
7658 translating function codes into names. The old 'ASN1 error code imbedded
7659 in a comment' is no longer necessary and it doesn't use .err files which
7660 have now been deleted. Also the error code call doesn't have to appear all
7661 on one line (which resulted in some large lines...).
7664 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
7667 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
7668 0 (which usually indicates a closed connection), but continue reading.
7671 *) Fix some race conditions.
7674 *) Add support for CRL distribution points extension. Add Certificate
7675 Policies and CRL distribution points documentation.
7678 *) Move the autogenerated header file parts to crypto/opensslconf.h.
7681 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
7682 8 of keying material. Merlin has also confirmed interop with this fix
7683 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
7684 [Merlin Hughes <merlin@baltimore.ie>]
7686 *) Fix lots of warnings.
7687 [Richard Levitte <levitte@stacken.kth.se>]
7689 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
7690 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
7691 [Richard Levitte <levitte@stacken.kth.se>]
7693 *) Fix problems with sizeof(long) == 8.
7694 [Andy Polyakov <appro@fy.chalmers.se>]
7696 *) Change functions to ANSI C.
7699 *) Fix typos in error codes.
7700 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
7702 *) Remove defunct assembler files from Configure.
7705 *) SPARC v8 assembler BIGNUM implementation.
7706 [Andy Polyakov <appro@fy.chalmers.se>]
7708 *) Support for Certificate Policies extension: both print and set.
7709 Various additions to support the r2i method this uses.
7712 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
7713 return a const string when you are expecting an allocated buffer.
7716 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
7717 types DirectoryString and DisplayText.
7720 *) Add code to allow r2i extensions to access the configuration database,
7721 add an LHASH database driver and add several ctx helper functions.
7724 *) Fix an evil bug in bn_expand2() which caused various BN functions to
7725 fail when they extended the size of a BIGNUM.
7728 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
7729 support typesafe stack.
7732 *) Fix typo in SSL_[gs]et_options().
7733 [Nils Frostberg <nils@medcom.se>]
7735 *) Delete various functions and files that belonged to the (now obsolete)
7736 old X509V3 handling code.
7739 *) New Configure option "rsaref".
7742 *) Don't auto-generate pem.h.
7745 *) Introduce type-safe ASN.1 SETs.
7748 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
7749 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
7751 *) Introduce type-safe STACKs. This will almost certainly break lots of code
7752 that links with OpenSSL (well at least cause lots of warnings), but fear
7753 not: the conversion is trivial, and it eliminates loads of evil casts. A
7754 few STACKed things have been converted already. Feel free to convert more.
7755 In the fullness of time, I'll do away with the STACK type altogether.
7758 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
7759 specified in <certfile> by updating the entry in the index.txt file.
7760 This way one no longer has to edit the index.txt file manually for
7761 revoking a certificate. The -revoke option does the gory details now.
7762 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
7764 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
7765 `-text' option at all and this way the `-noout -text' combination was
7766 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
7767 [Ralf S. Engelschall]
7769 *) Make sure a corresponding plain text error message exists for the
7770 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
7771 verify callback function determined that a certificate was revoked.
7772 [Ralf S. Engelschall]
7774 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
7775 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
7776 all available cipers including rc5, which was forgotten until now.
7777 In order to let the testing shell script know which algorithms
7778 are available, a new (up to now undocumented) command
7779 "openssl list-cipher-commands" is used.
7782 *) Bugfix: s_client occasionally would sleep in select() when
7783 it should have checked SSL_pending() first.
7786 *) New functions DSA_do_sign and DSA_do_verify to provide access to
7787 the raw DSA values prior to ASN.1 encoding.
7790 *) Tweaks to Configure
7791 [Niels Poppe <niels@netbox.org>]
7793 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
7797 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
7800 *) New config option to avoid instructions that are illegal on the 80386.
7801 The default code is faster, but requires at least a 486.
7804 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
7805 SSL2_SERVER_VERSION (not used at all) macros, which are now the
7806 same as SSL2_VERSION anyway.
7809 *) New "-showcerts" option for s_client.
7812 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
7813 application. Various cleanups and fixes.
7816 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
7817 modify error routines to work internally. Add error codes and PBE init
7818 to library startup routines.
7821 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
7822 packing functions to asn1 and evp. Changed function names and error
7823 codes along the way.
7826 *) PKCS12 integration: and so it begins... First of several patches to
7827 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
7828 objects to objects.h
7831 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
7832 and display support for Thawte strong extranet extension.
7835 *) Add LinuxPPC support.
7836 [Jeff Dubrule <igor@pobox.org>]
7838 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
7839 bn_div_words in alpha.s.
7840 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
7842 *) Make sure the RSA OAEP test is skipped under -DRSAref because
7843 OAEP isn't supported when OpenSSL is built with RSAref.
7844 [Ulf Moeller <ulf@fitug.de>]
7846 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
7847 so they no longer are missing under -DNOPROTO.
7848 [Soren S. Jorvang <soren@t.dk>]
7851 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
7853 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
7854 doesn't work when the session is reused. Coming soon!
7857 *) Fix a security hole, that allows sessions to be reused in the wrong
7858 context thus bypassing client cert protection! All software that uses
7859 client certs and session caches in multiple contexts NEEDS PATCHING to
7860 allow session reuse! A fuller solution is in the works.
7861 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
7863 *) Some more source tree cleanups (removed obsolete files
7864 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
7865 permission on "config" script to be executable) and a fix for the INSTALL
7867 [Ulf Moeller <ulf@fitug.de>]
7869 *) Remove some legacy and erroneous uses of malloc, free instead of
7871 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
7873 *) Make rsa_oaep_test return non-zero on error.
7874 [Ulf Moeller <ulf@fitug.de>]
7876 *) Add support for native Solaris shared libraries. Configure
7877 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
7878 if someone would make that last step automatic.
7879 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
7881 *) ctx_size was not built with the right compiler during "make links". Fixed.
7884 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
7885 except NULL ciphers". This means the default cipher list will no longer
7886 enable NULL ciphers. They need to be specifically enabled e.g. with
7887 the string "DEFAULT:eNULL".
7890 *) Fix to RSA private encryption routines: if p < q then it would
7891 occasionally produce an invalid result. This will only happen with
7892 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
7895 *) Be less restrictive and allow also `perl util/perlpath.pl
7896 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
7897 because this way one can also use an interpreter named `perl5' (which is
7898 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
7899 installed as `perl').
7900 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7902 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
7903 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7905 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
7906 advapi32.lib to Win32 build and change the pem test comparision
7907 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
7908 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
7909 and crypto/des/ede_cbcm_enc.c.
7912 *) DES quad checksum was broken on big-endian architectures. Fixed.
7915 *) Comment out two functions in bio.h that aren't implemented. Fix up the
7916 Win32 test batch file so it (might) work again. The Win32 test batch file
7917 is horrible: I feel ill....
7920 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
7921 in e_os.h. Audit of header files to check ANSI and non ANSI
7922 sections: 10 functions were absent from non ANSI section and not exported
7923 from Windows DLLs. Fixed up libeay.num for new functions.
7926 *) Make `openssl version' output lines consistent.
7927 [Ralf S. Engelschall]
7929 *) Fix Win32 symbol export lists for BIO functions: Added
7930 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
7931 to ms/libeay{16,32}.def.
7932 [Ralf S. Engelschall]
7934 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
7935 fine under Unix and passes some trivial tests I've now added. But the
7936 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
7937 added to make sure no one expects that this stuff really works in the
7938 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
7939 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
7941 [Ralf S. Engelschall]
7943 *) Fix the generation of two part addresses in perl.
7944 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
7946 *) Add config entry for Linux on MIPS.
7947 [John Tobey <jtobey@channel1.com>]
7949 *) Make links whenever Configure is run, unless we are on Windoze.
7952 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
7953 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
7957 *) Add a useful kludge to allow package maintainers to specify compiler and
7958 other platforms details on the command line without having to patch the
7959 Configure script everytime: One now can use ``perl Configure
7960 <id>:<details>'', i.e. platform ids are allowed to have details appended
7961 to them (seperated by colons). This is treated as there would be a static
7962 pre-configured entry in Configure's %table under key <id> with value
7963 <details> and ``perl Configure <id>'' is called. So, when you want to
7964 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
7965 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
7966 now, which overrides the FreeBSD-elf entry on-the-fly.
7967 [Ralf S. Engelschall]
7969 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
7972 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
7973 on the `perl Configure ...' command line. This way one can compile
7974 OpenSSL libraries with Position Independent Code (PIC) which is needed
7975 for linking it into DSOs.
7976 [Ralf S. Engelschall]
7978 *) Remarkably, export ciphers were totally broken and no-one had noticed!
7982 *) Cleaned up the LICENSE document: The official contact for any license
7983 questions now is the OpenSSL core team under openssl-core@openssl.org.
7984 And add a paragraph about the dual-license situation to make sure people
7985 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
7986 to the OpenSSL toolkit.
7987 [Ralf S. Engelschall]
7989 *) General source tree makefile cleanups: Made `making xxx in yyy...'
7990 display consistent in the source tree and replaced `/bin/rm' by `rm'.
7991 Additonally cleaned up the `make links' target: Remove unnecessary
7992 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
7993 to speed processing and no longer clutter the display with confusing
7994 stuff. Instead only the actually done links are displayed.
7995 [Ralf S. Engelschall]
7997 *) Permit null encryption ciphersuites, used for authentication only. It used
7998 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
7999 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
8003 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
8004 signed attributes when verifying signatures (this would break them),
8005 the detached data encoding was wrong and public keys obtained using
8006 X509_get_pubkey() weren't freed.
8009 *) Add text documentation for the BUFFER functions. Also added a work around
8010 to a Win95 console bug. This was triggered by the password read stuff: the
8011 last character typed gets carried over to the next fread(). If you were
8012 generating a new cert request using 'req' for example then the last
8013 character of the passphrase would be CR which would then enter the first
8017 *) Added the new `Includes OpenSSL Cryptography Software' button as
8018 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
8019 button and can be used by applications based on OpenSSL to show the
8020 relationship to the OpenSSL project.
8021 [Ralf S. Engelschall]
8023 *) Remove confusing variables in function signatures in files
8024 ssl/ssl_lib.c and ssl/ssl.h.
8025 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8027 *) Don't install bss_file.c under PREFIX/include/
8028 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8030 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
8031 functions that return function pointers and has support for NT specific
8032 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
8033 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
8034 unsigned to signed types: this was killing the Win32 compile.
8037 *) Add new certificate file to stack functions,
8038 SSL_add_dir_cert_subjects_to_stack() and
8039 SSL_add_file_cert_subjects_to_stack(). These largely supplant
8040 SSL_load_client_CA_file(), and can be used to add multiple certs easily
8041 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
8042 This means that Apache-SSL and similar packages don't have to mess around
8043 to add as many CAs as they want to the preferred list.
8046 *) Experiment with doxygen documentation. Currently only partially applied to
8048 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
8049 openssl.doxy as the configuration file.
8052 *) Get rid of remaining C++-style comments which strict C compilers hate.
8053 [Ralf S. Engelschall, pointed out by Carlos Amengual]
8055 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
8056 compiled in by default: it has problems with large keys.
8059 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
8060 DH private keys and/or callback functions which directly correspond to
8061 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
8062 is needed for applications which have to configure certificates on a
8063 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
8065 For the RSA certificate situation is makes no difference, but
8066 for the DSA certificate situation this fixes the "no shared cipher"
8067 problem where the OpenSSL cipher selection procedure failed because the
8068 temporary keys were not overtaken from the context and the API provided
8069 no way to reconfigure them.
8070 The new functions now let applications reconfigure the stuff and they
8071 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
8072 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
8073 non-public-API function ssl_cert_instantiate() is used as a helper
8074 function and also to reduce code redundancy inside ssl_rsa.c.
8075 [Ralf S. Engelschall]
8077 *) Move s_server -dcert and -dkey options out of the undocumented feature
8078 area because they are useful for the DSA situation and should be
8079 recognized by the users.
8080 [Ralf S. Engelschall]
8082 *) Fix the cipher decision scheme for export ciphers: the export bits are
8083 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
8084 SSL_EXP_MASK. So, the original variable has to be used instead of the
8085 already masked variable.
8086 [Richard Levitte <levitte@stacken.kth.se>]
8088 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
8089 [Richard Levitte <levitte@stacken.kth.se>]
8091 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
8092 from `int' to `unsigned int' because it's a length and initialized by
8093 EVP_DigestFinal() which expects an `unsigned int *'.
8094 [Richard Levitte <levitte@stacken.kth.se>]
8096 *) Don't hard-code path to Perl interpreter on shebang line of Configure
8097 script. Instead use the usual Shell->Perl transition trick.
8098 [Ralf S. Engelschall]
8100 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
8101 (in addition to RSA certificates) to match the behaviour of `openssl dsa
8102 -noout -modulus' as it's already the case for `openssl rsa -noout
8103 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
8104 currently the public key is printed (a decision which was already done by
8105 `openssl dsa -modulus' in the past) which serves a similar purpose.
8106 Additionally the NO_RSA no longer completely removes the whole -modulus
8107 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
8109 [Ralf S. Engelschall]
8111 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
8112 BIO. See the source (crypto/evp/bio_ok.c) for more info.
8113 [Arne Ansper <arne@ats.cyber.ee>]
8115 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
8116 to be added. Now both 'req' and 'ca' can use new objects defined in the
8120 *) Add cool BIO that does syslog (or event log on NT).
8121 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
8123 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
8124 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
8125 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
8126 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
8129 *) Add preliminary config info for new extension code.
8132 *) Make RSA_NO_PADDING really use no padding.
8133 [Ulf Moeller <ulf@fitug.de>]
8135 *) Generate errors when private/public key check is done.
8138 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
8139 for some CRL extensions and new objects added.
8142 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
8143 key usage extension and fuller support for authority key id.
8146 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
8147 padding method for RSA, which is recommended for new applications in PKCS
8148 #1 v2.0 (RFC 2437, October 1998).
8149 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
8150 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
8151 against Bleichbacher's attack on RSA.
8152 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
8155 *) Updates to the new SSL compression code
8156 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8158 *) Fix so that the version number in the master secret, when passed
8159 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
8160 (because the server will not accept higher), that the version number
8161 is 0x03,0x01, not 0x03,0x00
8162 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8164 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
8165 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
8166 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
8169 *) Support for RAW extensions where an arbitrary extension can be
8170 created by including its DER encoding. See apps/openssl.cnf for
8174 *) Make sure latest Perl versions don't interpret some generated C array
8175 code as Perl array code in the crypto/err/err_genc.pl script.
8176 [Lars Weber <3weber@informatik.uni-hamburg.de>]
8178 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
8179 not many people have the assembler. Various Win32 compilation fixes and
8180 update to the INSTALL.W32 file with (hopefully) more accurate Win32
8184 *) Modify configure script 'Configure' to automatically create crypto/date.h
8185 file under Win32 and also build pem.h from pem.org. New script
8186 util/mkfiles.pl to create the MINFO file on environments that can't do a
8187 'make files': perl util/mkfiles.pl >MINFO should work.
8190 *) Major rework of DES function declarations, in the pursuit of correctness
8191 and purity. As a result, many evil casts evaporated, and some weirdness,
8192 too. You may find this causes warnings in your code. Zapping your evil
8193 casts will probably fix them. Mostly.
8196 *) Fix for a typo in asn1.h. Bug fix to object creation script
8197 obj_dat.pl. It considered a zero in an object definition to mean
8198 "end of object": none of the objects in objects.h have any zeros
8199 so it wasn't spotted.
8200 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
8202 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
8203 Masking (CBCM). In the absence of test vectors, the best I have been able
8204 to do is check that the decrypt undoes the encrypt, so far. Send me test
8205 vectors if you have them.
8208 *) Correct calculation of key length for export ciphers (too much space was
8209 allocated for null ciphers). This has not been tested!
8212 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
8213 message is now correct (it understands "crypto" and "ssl" on its
8214 command line). There is also now an "update" option. This will update
8215 the util/ssleay.num and util/libeay.num files with any new functions.
8217 perl util/mkdef.pl crypto ssl update
8218 it will update them.
8221 *) Overhauled the Perl interface (perl/*):
8222 - ported BN stuff to OpenSSL's different BN library
8223 - made the perl/ source tree CVS-aware
8224 - renamed the package from SSLeay to OpenSSL (the files still contain
8225 their history because I've copied them in the repository)
8226 - removed obsolete files (the test scripts will be replaced
8227 by better Test::Harness variants in the future)
8228 [Ralf S. Engelschall]
8230 *) First cut for a very conservative source tree cleanup:
8231 1. merge various obsolete readme texts into doc/ssleay.txt
8232 where we collect the old documents and readme texts.
8233 2. remove the first part of files where I'm already sure that we no
8234 longer need them because of three reasons: either they are just temporary
8235 files which were left by Eric or they are preserved original files where
8236 I've verified that the diff is also available in the CVS via "cvs diff
8237 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
8238 the crypto/md/ stuff).
8239 [Ralf S. Engelschall]
8241 *) More extension code. Incomplete support for subject and issuer alt
8242 name, issuer and authority key id. Change the i2v function parameters
8243 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
8244 what that's for :-) Fix to ASN1 macro which messed up
8245 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
8248 *) Preliminary support for ENUMERATED type. This is largely copied from the
8252 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
8253 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8255 *) Make sure `make rehash' target really finds the `openssl' program.
8256 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8258 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
8259 like to hear about it if this slows down other processors.
8262 *) Add CygWin32 platform information to Configure script.
8263 [Alan Batie <batie@aahz.jf.intel.com>]
8265 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
8266 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
8268 *) New program nseq to manipulate netscape certificate sequences
8271 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
8275 *) Fixes to BN code. Previously the default was to define BN_RECURSION
8276 but the BN code had some problems that would cause failures when
8277 doing certificate verification and some other functions.
8278 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8280 *) Add ASN1 and PEM code to support netscape certificate sequences.
8283 *) Add ASN1 and PEM code to support netscape certificate sequences.
8286 *) Add several PKIX and private extended key usage OIDs.
8289 *) Modify the 'ca' program to handle the new extension code. Modify
8290 openssl.cnf for new extension format, add comments.
8293 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
8294 and add a sample to openssl.cnf so req -x509 now adds appropriate
8298 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
8299 error code, add initial support to X509_print() and x509 application.
8302 *) Takes a deep breath and start addding X509 V3 extension support code. Add
8303 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
8304 stuff is currently isolated and isn't even compiled yet.
8307 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
8308 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
8309 Removed the versions check from X509 routines when loading extensions:
8310 this allows certain broken certificates that don't set the version
8311 properly to be processed.
8314 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
8315 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
8316 can still be regenerated with "make depend".
8319 *) Spelling mistake in C version of CAST-128.
8320 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
8322 *) Changes to the error generation code. The perl script err-code.pl
8323 now reads in the old error codes and retains the old numbers, only
8324 adding new ones if necessary. It also only changes the .err files if new
8325 codes are added. The makefiles have been modified to only insert errors
8326 when needed (to avoid needlessly modifying header files). This is done
8327 by only inserting errors if the .err file is newer than the auto generated
8328 C file. To rebuild all the error codes from scratch (the old behaviour)
8329 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
8330 or delete all the .err files.
8333 *) CAST-128 was incorrectly implemented for short keys. The C version has
8334 been fixed, but is untested. The assembler versions are also fixed, but
8335 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
8336 to regenerate it if needed.
8337 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
8338 Hagino <itojun@kame.net>]
8340 *) File was opened incorrectly in randfile.c.
8341 [Ulf Möller <ulf@fitug.de>]
8343 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
8344 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
8345 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
8346 al: it's just almost always a UTCTime. Note this patch adds new error
8347 codes so do a "make errors" if there are problems.
8350 *) Correct Linux 1 recognition in config.
8351 [Ulf Möller <ulf@fitug.de>]
8353 *) Remove pointless MD5 hash when using DSA keys in ca.
8354 [Anonymous <nobody@replay.com>]
8356 *) Generate an error if given an empty string as a cert directory. Also
8357 generate an error if handed NULL (previously returned 0 to indicate an
8358 error, but didn't set one).
8359 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
8361 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
8364 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
8365 parameters. This was causing a warning which killed off the Win32 compile.
8368 *) Remove C++ style comments from crypto/bn/bn_local.h.
8369 [Neil Costigan <neil.costigan@celocom.com>]
8371 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
8372 based on a text string, looking up short and long names and finally
8373 "dot" format. The "dot" format stuff didn't work. Added new function
8374 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
8375 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
8376 OID is not part of the table.
8379 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
8380 X509_LOOKUP_by_alias().
8383 *) Sort openssl functions by name.
8386 *) Get the gendsa program working (hopefully) and add it to app list. Remove
8387 encryption from sample DSA keys (in case anyone is interested the password
8391 *) Make _all_ *_free functions accept a NULL pointer.
8392 [Frans Heymans <fheymans@isaserver.be>]
8394 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
8396 [Anonymous <nobody@replay.com>]
8398 *) s_server should send the CAfile as acceptable CAs, not its own cert.
8399 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8401 *) Don't blow it for numeric -newkey arguments to apps/req.
8402 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8404 *) Temp key "for export" tests were wrong in s3_srvr.c.
8405 [Anonymous <nobody@replay.com>]
8407 *) Add prototype for temp key callback functions
8408 SSL_CTX_set_tmp_{rsa,dh}_callback().
8411 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
8412 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
8415 *) X509_name_add_entry() freed the wrong thing after an error.
8416 [Arne Ansper <arne@ats.cyber.ee>]
8418 *) rsa_eay.c would attempt to free a NULL context.
8419 [Arne Ansper <arne@ats.cyber.ee>]
8421 *) BIO_s_socket() had a broken should_retry() on Windoze.
8422 [Arne Ansper <arne@ats.cyber.ee>]
8424 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
8425 [Arne Ansper <arne@ats.cyber.ee>]
8427 *) Make sure the already existing X509_STORE->depth variable is initialized
8428 in X509_STORE_new(), but document the fact that this variable is still
8429 unused in the certificate verification process.
8430 [Ralf S. Engelschall]
8432 *) Fix the various library and apps files to free up pkeys obtained from
8433 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
8436 *) Fix reference counting in X509_PUBKEY_get(). This makes
8437 demos/maurice/example2.c work, amongst others, probably.
8438 [Steve Henson and Ben Laurie]
8440 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
8441 `openssl' and second, the shortcut symlinks for the `openssl <command>'
8442 are no longer created. This way we have a single and consistent command
8443 line interface `openssl <command>', similar to `cvs <command>'.
8444 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
8446 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
8447 BIT STRING wrapper always have zero unused bits.
8450 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
8453 *) Make the top-level INSTALL documentation easier to understand.
8456 *) Makefiles updated to exit if an error occurs in a sub-directory
8457 make (including if user presses ^C) [Paul Sutton]
8459 *) Make Montgomery context stuff explicit in RSA data structure.
8462 *) Fix build order of pem and err to allow for generated pem.h.
8465 *) Fix renumbering bug in X509_NAME_delete_entry().
8468 *) Enhanced the err-ins.pl script so it makes the error library number
8469 global and can add a library name. This is needed for external ASN1 and
8470 other error libraries.
8473 *) Fixed sk_insert which never worked properly.
8476 *) Fix ASN1 macros so they can handle indefinite length construted
8477 EXPLICIT tags. Some non standard certificates use these: they can now
8481 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
8482 into a single doc/ssleay.txt bundle. This way the information is still
8483 preserved but no longer messes up this directory. Now it's new room for
8484 the new set of documenation files.
8485 [Ralf S. Engelschall]
8487 *) SETs were incorrectly DER encoded. This was a major pain, because they
8488 shared code with SEQUENCEs, which aren't coded the same. This means that
8489 almost everything to do with SETs or SEQUENCEs has either changed name or
8490 number of arguments.
8491 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
8493 *) Fix test data to work with the above.
8496 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
8497 was already fixed by Eric for 0.9.1 it seems.
8498 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
8500 *) Autodetect FreeBSD3.
8503 *) Fix various bugs in Configure. This affects the following platforms:
8507 unixware-2.0-pentium
8511 *) Eliminate generated files from CVS. Reorder tests to regenerate files
8512 before they are needed.
8515 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
8519 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
8521 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
8522 changed SSLeay to OpenSSL in version strings.
8523 [Ralf S. Engelschall]
8525 *) Some fixups to the top-level documents.
8528 *) Fixed the nasty bug where rsaref.h was not found under compile-time
8529 because the symlink to include/ was missing.
8530 [Ralf S. Engelschall]
8532 *) Incorporated the popular no-RSA/DSA-only patches
8533 which allow to compile a RSA-free SSLeay.
8534 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
8536 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
8537 when "ssleay" is still not found.
8538 [Ralf S. Engelschall]
8540 *) Added more platforms to Configure: Cray T3E, HPUX 11,
8541 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
8543 *) Updated the README file.
8544 [Ralf S. Engelschall]
8546 *) Added various .cvsignore files in the CVS repository subdirs
8547 to make a "cvs update" really silent.
8548 [Ralf S. Engelschall]
8550 *) Recompiled the error-definition header files and added
8551 missing symbols to the Win32 linker tables.
8552 [Ralf S. Engelschall]
8554 *) Cleaned up the top-level documents;
8555 o new files: CHANGES and LICENSE
8556 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
8557 o merged COPYRIGHT into LICENSE
8558 o removed obsolete TODO file
8559 o renamed MICROSOFT to INSTALL.W32
8560 [Ralf S. Engelschall]
8562 *) Removed dummy files from the 0.9.1b source tree:
8563 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
8564 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
8565 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
8566 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
8567 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
8568 [Ralf S. Engelschall]
8570 *) Added various platform portability fixes.
8573 *) The Genesis of the OpenSSL rpject:
8574 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
8575 Young and Tim J. Hudson created while they were working for C2Net until
8577 [The OpenSSL Project]
8580 Changes between 0.9.0b and 0.9.1b [not released]
8582 *) Updated a few CA certificates under certs/
8585 *) Changed some BIGNUM api stuff.
8588 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
8589 DGUX x86, Linux Alpha, etc.
8592 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
8593 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
8597 *) Add -strparse option to asn1pars program which parses nested
8599 [Dr Stephen Henson <shenson@bigfoot.com>]
8601 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
8604 *) DSA fix for "ca" program.
8607 *) Added "-genkey" option to "dsaparam" program.
8610 *) Added RIPE MD160 (rmd160) message digest.
8613 *) Added -a (all) option to "ssleay version" command.
8616 *) Added PLATFORM define which is the id given to Configure.
8619 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
8622 *) Extended the ASN.1 parser routines.
8625 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
8628 *) Added a BN_CTX to the BN library.
8631 *) Fixed the weak key values in DES library
8634 *) Changed API in EVP library for cipher aliases.
8637 *) Added support for RC2/64bit cipher.
8640 *) Converted the lhash library to the crypto/mem.c functions.
8643 *) Added more recognized ASN.1 object ids.
8646 *) Added more RSA padding checks for SSL/TLS.
8649 *) Added BIO proxy/filter functionality.
8652 *) Added extra_certs to SSL_CTX which can be used
8653 send extra CA certificates to the client in the CA cert chain sending
8654 process. It can be configured with SSL_CTX_add_extra_chain_cert().
8657 *) Now Fortezza is denied in the authentication phase because
8658 this is key exchange mechanism is not supported by SSLeay at all.
8661 *) Additional PKCS1 checks.
8664 *) Support the string "TLSv1" for all TLS v1 ciphers.
8667 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
8668 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
8671 *) Fixed a few memory leaks.
8674 *) Fixed various code and comment typos.
8677 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
8678 bytes sent in the client random.
8679 [Edward Bishop <ebishop@spyglass.com>]