5 Changes between 0.9.8za and 0.9.8zb [6 Aug 2014]
7 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8 to a denial of service attack. A malicious server can crash the client
9 with a null pointer dereference (read) by specifying an anonymous (EC)DH
10 ciphersuite and sending carefully crafted handshake messages.
12 Thanks to Felix Gröbert (Google) for discovering and researching this
17 *) By sending carefully crafted DTLS packets an attacker could cause openssl
18 to leak memory. This can be exploited through a Denial of Service attack.
19 Thanks to Adam Langley for discovering and researching this issue.
23 *) An attacker can force openssl to consume large amounts of memory whilst
24 processing DTLS handshake messages. This can be exploited through a
25 Denial of Service attack.
26 Thanks to Adam Langley for discovering and researching this issue.
30 *) An attacker can force an error condition which causes openssl to crash
31 whilst processing DTLS packets due to memory being freed twice. This
32 can be exploited through a Denial of Service attack.
33 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
38 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
39 X509_name_oneline, X509_name_print_ex et al. to leak some information
40 from the stack. Applications may be affected if they echo pretty printing
41 output to the attacker.
43 Thanks to Ivan Fratric (Google) for discovering this issue.
45 [Emilia Käsper, and Steve Henson]
47 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
48 for corner cases. (Certain input points at infinity could lead to
49 bogus results, with non-infinity inputs mapped to infinity too.)
52 Changes between 0.9.8y and 0.9.8za [5 Jun 2014]
54 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
55 handshake can force the use of weak keying material in OpenSSL
56 SSL/TLS clients and servers.
58 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
59 researching this issue. (CVE-2014-0224)
60 [KIKUCHI Masashi, Steve Henson]
62 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
63 OpenSSL DTLS client the code can be made to recurse eventually crashing
66 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
68 [Imre Rad, Steve Henson]
70 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
71 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
72 client or server. This is potentially exploitable to run arbitrary
73 code on a vulnerable client or server.
75 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
76 [Jüri Aedla, Steve Henson]
78 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
79 are subject to a denial of service attack.
81 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
82 this issue. (CVE-2014-3470)
83 [Felix Gröbert, Ivan Fratric, Steve Henson]
85 *) Fix for the attack described in the paper "Recovering OpenSSL
86 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
87 by Yuval Yarom and Naomi Benger. Details can be obtained from:
88 http://eprint.iacr.org/2014/140
90 Thanks to Yuval Yarom and Naomi Benger for discovering this
91 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
92 [Yuval Yarom and Naomi Benger]
94 Thanks to mancha for backporting the fix to the 0.9.8 branch.
96 *) Fix handling of warning-level alerts in SSL23 client mode so they
97 don't cause client-side termination (eg. on SNI unrecognized_name
98 warnings). Add client and server support for six additional alerts
99 per RFC 6066 and RFC 4279.
102 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
103 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
104 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
105 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
106 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
107 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
108 [Rob Stradling, Adam Langley]
110 Changes between 0.9.8x and 0.9.8y [5 Feb 2013]
112 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
114 This addresses the flaw in CBC record processing discovered by
115 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
116 at: http://www.isg.rhul.ac.uk/tls/
118 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
119 Security Group at Royal Holloway, University of London
120 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
121 Emilia Käsper for the initial patch.
123 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
125 *) Return an error when checking OCSP signatures when key is NULL.
126 This fixes a DoS attack. (CVE-2013-0166)
129 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
130 the right response is stapled. Also change SSL_get_certificate()
131 so it returns the certificate actually sent.
132 See http://rt.openssl.org/Ticket/Display.html?id=2836.
134 [Rob Stradling <rob.stradling@comodo.com>]
136 *) Fix possible deadlock when decoding public keys.
139 Changes between 0.9.8w and 0.9.8x [10 May 2012]
141 *) Sanity check record length before skipping explicit IV in DTLS
144 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
145 fuzzing as a service testing platform.
149 *) Initialise tkeylen properly when encrypting CMS messages.
150 Thanks to Solar Designer of Openwall for reporting this issue.
153 Changes between 0.9.8v and 0.9.8w [23 Apr 2012]
155 *) The fix for CVE-2012-2110 did not take into account that the
156 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an
157 int in OpenSSL 0.9.8, making it still vulnerable. Fix by
158 rejecting negative len parameter. (CVE-2012-2131)
159 [Tomas Hoger <thoger@redhat.com>]
161 Changes between 0.9.8u and 0.9.8v [19 Apr 2012]
163 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
164 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
165 in CRYPTO_realloc_clean.
167 Thanks to Tavis Ormandy, Google Security Team, for discovering this
168 issue and to Adam Langley <agl@chromium.org> for fixing it.
170 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
172 Changes between 0.9.8t and 0.9.8u [12 Mar 2012]
174 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
175 in CMS and PKCS7 code. When RSA decryption fails use a random key for
176 content decryption and always return the same error. Note: this attack
177 needs on average 2^20 messages so it only affects automated senders. The
178 old behaviour can be reenabled in the CMS code by setting the
179 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
180 an MMA defence is not necessary.
181 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
182 this issue. (CVE-2012-0884)
185 *) Fix CVE-2011-4619: make sure we really are receiving a
186 client hello before rejecting multiple SGC restarts. Thanks to
187 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
190 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
192 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
193 Thanks to Antonio Martin, Enterprise Secure Access Research and
194 Development, Cisco Systems, Inc. for discovering this bug and
195 preparing a fix. (CVE-2012-0050)
198 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
200 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
201 of the Vaudenay padding oracle attack on CBC mode encryption
202 which enables an efficient plaintext recovery attack against
203 the OpenSSL implementation of DTLS. Their attack exploits timing
204 differences arising during decryption processing. A research
205 paper describing this attack can be found at:
206 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
207 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
208 Security Group at Royal Holloway, University of London
209 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
210 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
211 for preparing the fix. (CVE-2011-4108)
212 [Robin Seggelmann, Michael Tuexen]
214 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
215 [Ben Laurie, Kasper <ekasper@google.com>]
217 *) Clear bytes used for block padding of SSL 3.0 records.
219 [Adam Langley (Google)]
221 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
222 Kadianakis <desnacked@gmail.com> for discovering this issue and
223 Adam Langley for preparing the fix. (CVE-2011-4619)
224 [Adam Langley (Google)]
226 *) Prevent malformed RFC3779 data triggering an assertion failure.
227 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
228 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
229 [Rob Austein <sra@hactrn.net>]
231 *) Fix ssl_ciph.c set-up race.
232 [Adam Langley (Google)]
234 *) Fix spurious failures in ecdsatest.c.
235 [Emilia Käsper (Google)]
237 *) Fix the BIO_f_buffer() implementation (which was mixing different
238 interpretations of the '..._len' fields).
239 [Adam Langley (Google)]
241 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
242 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
243 threads won't reuse the same blinding coefficients.
245 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
246 lock to call BN_BLINDING_invert_ex, and avoids one use of
247 BN_BLINDING_update for each BN_BLINDING structure (previously,
248 the last update always remained unused).
249 [Emilia Käsper (Google)]
251 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
252 for multi-threaded use of ECDH.
253 [Adam Langley (Google)]
255 *) Fix x509_name_ex_d2i memory leak on bad inputs.
258 *) Add protection against ECDSA timing attacks as mentioned in the paper
259 by Billy Bob Brumley and Nicola Tuveri, see:
261 http://eprint.iacr.org/2011/232.pdf
263 [Billy Bob Brumley and Nicola Tuveri]
265 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
267 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
268 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
270 *) Fix bug in string printing code: if *any* escaping is enabled we must
271 escape the escape character (backslash) or the resulting string is
275 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
277 *) Disable code workaround for ancient and obsolete Netscape browsers
278 and servers: an attacker can use it in a ciphersuite downgrade attack.
279 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
282 *) Fixed J-PAKE implementation error, originally discovered by
283 Sebastien Martini, further info and confirmation from Stefan
284 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
287 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
289 *) Fix extension code to avoid race conditions which can result in a buffer
290 overrun vulnerability: resumed sessions must not be modified as they can
291 be shared by multiple threads. CVE-2010-3864
294 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
297 *) Don't reencode certificate when calculating signature: cache and use
298 the original encoding instead. This makes signature verification of
299 some broken encodings work correctly.
302 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
303 is also one of the inputs.
304 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
306 *) Don't repeatedly append PBE algorithms to table if they already exist.
307 Sort table on each new add. This effectively makes the table read only
308 after all algorithms are added and subsequent calls to PKCS12_pbe_add
312 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
314 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
317 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
318 access or freeing data twice (CVE-2010-0742)
319 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
321 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
322 common in certificates and some applications which only call
323 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
327 Reduce copying into .apps and .test in makevms.com
328 Don't try to use blank CA certificate in CA.com
329 Allow use of C files from original directories in maketests.com
330 [Steven M. Schweda" <sms@antinode.info>]
332 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
334 *) When rejecting SSL/TLS records due to an incorrect version number, never
335 update s->server with a new major version number. As of
336 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
337 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
338 the previous behavior could result in a read attempt at NULL when
339 receiving specific incorrect SSL/TLS records once record payload
340 protection is active. (CVE-2010-0740)
341 [Bodo Moeller, Adam Langley <agl@chromium.org>]
343 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
344 could be crashed if the relevant tables were not present (e.g. chrooted).
345 [Tomas Hoger <thoger@redhat.com>]
347 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
349 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
350 [Martin Olsson, Neel Mehta]
352 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
353 accommodate for stack sorting, always a write lock!).
356 *) On some versions of WIN32 Heap32Next is very slow. This can cause
357 excessive delays in the RAND_poll(): over a minute. As a workaround
358 include a time check in the inner Heap32Next loop too.
361 *) The code that handled flushing of data in SSL/TLS originally used the
362 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
363 the problem outlined in PR#1949. The fix suggested there however can
364 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
365 of Apache). So instead simplify the code to flush unconditionally.
366 This should be fine since flushing with no data to flush is a no op.
369 *) Handle TLS versions 2.0 and later properly and correctly use the
370 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
371 off ancient servers have a habit of sticking around for a while...
374 *) Modify compression code so it frees up structures without using the
375 ex_data callbacks. This works around a problem where some applications
376 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
377 restarting) then use compression (e.g. SSL with compression) later.
378 This results in significant per-connection memory leaks and
379 has caused some security issues including CVE-2008-1678 and
383 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
384 change when encrypting or decrypting.
387 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
388 connect and renegotiate with servers which do not support RI.
389 Until RI is more widely deployed this option is enabled by default.
392 *) Add "missing" ssl ctrls to clear options and mode.
395 *) If client attempts to renegotiate and doesn't support RI respond with
396 a no_renegotiation alert as required by RFC5746. Some renegotiating
397 TLS clients will continue a connection gracefully when they receive
398 the alert. Unfortunately OpenSSL mishandled this alert and would hang
399 waiting for a server hello which it will never receive. Now we treat a
400 received no_renegotiation alert as a fatal error. This is because
401 applications requesting a renegotiation might well expect it to succeed
402 and would have no code in place to handle the server denying it so the
403 only safe thing to do is to terminate the connection.
406 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
407 peer supports secure renegotiation and 0 otherwise. Print out peer
408 renegotiation support in s_client/s_server.
411 *) Replace the highly broken and deprecated SPKAC certification method with
412 the updated NID creation version. This should correctly handle UTF8.
415 *) Implement RFC5746. Re-enable renegotiation but require the extension
416 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
417 turns out to be a bad idea. It has been replaced by
418 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
419 SSL_CTX_set_options(). This is really not recommended unless you
420 know what you are doing.
421 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
423 *) Fixes to stateless session resumption handling. Use initial_ctx when
424 issuing and attempting to decrypt tickets in case it has changed during
425 servername handling. Use a non-zero length session ID when attempting
426 stateless session resumption: this makes it possible to determine if
427 a resumption has occurred immediately after receiving server hello
428 (several places in OpenSSL subtly assume this) instead of later in
432 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
433 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
434 fixes for a few places where the return code is not checked
436 [Julia Lawall <julia@diku.dk>]
438 *) Add --strict-warnings option to Configure script to include devteam
439 warnings in other configurations.
442 *) Add support for --libdir option and LIBDIR variable in makefiles. This
443 makes it possible to install openssl libraries in locations which
444 have names other than "lib", for example "/usr/lib64" which some
446 [Steve Henson, based on patch from Jeremy Utley]
448 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
449 X690 8.9.12 and can produce some misleading textual output of OIDs.
450 [Steve Henson, reported by Dan Kaminsky]
452 *) Delete MD2 from algorithm tables. This follows the recommendation in
453 several standards that it is not used in new applications due to
454 several cryptographic weaknesses. For binary compatibility reasons
455 the MD2 API is still compiled in by default.
458 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
462 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
463 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
465 [Guenter <lists@gknw.net>]
467 *) Fix the server certificate chain building code to use X509_verify_cert(),
468 it used to have an ad-hoc builder which was unable to cope with anything
469 other than a simple chain.
470 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
472 *) Don't check self signed certificate signatures in X509_verify_cert()
473 by default (a flag can override this): it just wastes time without
474 adding any security. As a useful side effect self signed root CAs
475 with non-FIPS digests are now usable in FIPS mode.
478 *) In dtls1_process_out_of_seq_message() the check if the current message
479 is already buffered was missing. For every new message was memory
480 allocated, allowing an attacker to perform an denial of service attack
481 with sending out of seq handshake messages until there is no memory
482 left. Additionally every future messege was buffered, even if the
483 sequence number made no sense and would be part of another handshake.
484 So only messages with sequence numbers less than 10 in advance will be
485 buffered. (CVE-2009-1378)
486 [Robin Seggelmann, discovered by Daniel Mentz]
488 *) Records are buffered if they arrive with a future epoch to be
489 processed after finishing the corresponding handshake. There is
490 currently no limitation to this buffer allowing an attacker to perform
491 a DOS attack with sending records with future epochs until there is no
492 memory left. This patch adds the pqueue_size() function to detemine
493 the size of a buffer and limits the record buffer to 100 entries.
495 [Robin Seggelmann, discovered by Daniel Mentz]
497 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
498 parent structure is freed. (CVE-2009-1379)
501 *) Handle non-blocking I/O properly in SSL_shutdown() call.
502 [Darryl Miles <darryl-mailinglists@netbauds.net>]
505 [Ilya O. <vrghost@gmail.com>]
507 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
509 *) Disable renegotiation completely - this fixes a severe security
510 problem (CVE-2009-3555) at the cost of breaking all
511 renegotiation. Renegotiation can be re-enabled by setting
512 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
513 run-time. This is really not recommended unless you know what
517 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
519 *) Don't set val to NULL when freeing up structures, it is freed up by
520 underlying code. If sizeof(void *) > sizeof(long) this can result in
521 zeroing past the valid field. (CVE-2009-0789)
522 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
524 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
525 checked correctly. This would allow some invalid signed attributes to
526 appear to verify correctly. (CVE-2009-0591)
527 [Ivan Nestlerode <inestlerode@us.ibm.com>]
529 *) Reject UniversalString and BMPString types with invalid lengths. This
530 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
531 a legal length. (CVE-2009-0590)
534 *) Set S/MIME signing as the default purpose rather than setting it
535 unconditionally. This allows applications to override it at the store
539 *) Permit restricted recursion of ASN1 strings. This is needed in practice
540 to handle some structures.
543 *) Improve efficiency of mem_gets: don't search whole buffer each time
545 [Jeremy Shapiro <jnshapir@us.ibm.com>]
547 *) New -hex option for openssl rand.
550 *) Print out UTF8String and NumericString when parsing ASN1.
553 *) Support NumericString type for name components.
556 *) Allow CC in the environment to override the automatically chosen
557 compiler. Note that nothing is done to ensure flags work with the
561 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
563 *) Properly check EVP_VerifyFinal() and similar return values
565 [Ben Laurie, Bodo Moeller, Google Security Team]
567 *) Enable TLS extensions by default.
570 *) Allow the CHIL engine to be loaded, whether the application is
571 multithreaded or not. (This does not release the developer from the
572 obligation to set up the dynamic locking callbacks.)
573 [Sander Temme <sander@temme.net>]
575 *) Use correct exit code if there is an error in dgst command.
576 [Steve Henson; problem pointed out by Roland Dirlewanger]
578 *) Tweak Configure so that you need to say "experimental-jpake" to enable
579 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
582 *) Add experimental JPAKE support, including demo authentication in
583 s_client and s_server.
586 *) Set the comparison function in v3_addr_canonize().
587 [Rob Austein <sra@hactrn.net>]
589 *) Add support for XMPP STARTTLS in s_client.
590 [Philip Paeps <philip@freebsd.org>]
592 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
593 to ensure that even with this option, only ciphersuites in the
594 server's preference list will be accepted. (Note that the option
595 applies only when resuming a session, so the earlier behavior was
596 just about the algorithm choice for symmetric cryptography.)
599 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
601 *) Fix NULL pointer dereference if a DTLS server received
602 ChangeCipherSpec as first record (CVE-2009-1386).
605 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
606 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
609 *) The fix in 0.9.8c that supposedly got rid of unsafe
610 double-checked locking was incomplete for RSA blinding,
611 addressing just one layer of what turns out to have been
612 doubly unsafe triple-checked locking.
614 So now fix this for real by retiring the MONT_HELPER macro
615 in crypto/rsa/rsa_eay.c.
617 [Bodo Moeller; problem pointed out by Marius Schilder]
619 *) Various precautionary measures:
621 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
623 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
624 (NB: This would require knowledge of the secret session ticket key
625 to exploit, in which case you'd be SOL either way.)
627 - Change bn_nist.c so that it will properly handle input BIGNUMs
628 outside the expected range.
630 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
633 [Neel Mehta, Bodo Moeller]
635 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
636 the load fails. Useful for distros.
637 [Ben Laurie and the FreeBSD team]
639 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
642 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
645 *) Expand ENGINE to support engine supplied SSL client certificate functions.
647 This work was sponsored by Logica.
650 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
651 keystores. Support for SSL/TLS client authentication too.
652 Not compiled unless enable-capieng specified to Configure.
654 This work was sponsored by Logica.
657 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
658 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
659 attribute creation routines such as certifcate requests and PKCS#12
663 Changes between 0.9.8g and 0.9.8h [28 May 2008]
665 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
666 handshake which could lead to a cilent crash as found using the
667 Codenomicon TLS test suite (CVE-2008-1672)
668 [Steve Henson, Mark Cox]
670 *) Fix double free in TLS server name extensions which could lead to
671 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
674 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
676 Clear the error queue to ensure that error entries left from
677 older function calls do not interfere with the correct operation.
678 [Lutz Jaenicke, Erik de Castro Lopo]
680 *) Remove root CA certificates of commercial CAs:
682 The OpenSSL project does not recommend any specific CA and does not
683 have any policy with respect to including or excluding any CA.
684 Therefore it does not make any sense to ship an arbitrary selection
685 of root CA certificates with the OpenSSL software.
688 *) RSA OAEP patches to fix two separate invalid memory reads.
689 The first one involves inputs when 'lzero' is greater than
690 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
691 before the beginning of from). The second one involves inputs where
692 the 'db' section contains nothing but zeroes (there is a one-byte
693 invalid read after the end of 'db').
694 [Ivan Nestlerode <inestlerode@us.ibm.com>]
696 *) Partial backport from 0.9.9-dev:
698 Introduce bn_mul_mont (dedicated Montgomery multiplication
699 procedure) as a candidate for BIGNUM assembler implementation.
700 While 0.9.9-dev uses assembler for various architectures, only
701 x86_64 is available by default here in the 0.9.8 branch, and
702 32-bit x86 is available through a compile-time setting.
704 To try the 32-bit x86 assembler implementation, use Configure
705 option "enable-montasm" (which exists only for this backport).
707 As "enable-montasm" for 32-bit x86 disclaims code stability
708 anyway, in this constellation we activate additional code
709 backported from 0.9.9-dev for further performance improvements,
710 namely BN_from_montgomery_word. (To enable this otherwise,
711 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
713 [Andy Polyakov (backport partially by Bodo Moeller)]
715 *) Add TLS session ticket callback. This allows an application to set
716 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
717 values. This is useful for key rollover for example where several key
718 sets may exist with different names.
721 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
722 This was broken until now in 0.9.8 releases, such that the only way
723 a registered ENGINE could be used (assuming it initialises
724 successfully on the host) was to explicitly set it as the default
725 for the relevant algorithms. This is in contradiction with 0.9.7
726 behaviour and the documentation. With this fix, when an ENGINE is
727 registered into a given algorithm's table of implementations, the
728 'uptodate' flag is reset so that auto-discovery will be used next
729 time a new context for that algorithm attempts to select an
731 [Ian Lister (tweaked by Geoff Thorpe)]
733 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
734 implemention in the following ways:
736 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
739 Lack of BER streaming support means one pass streaming processing is
740 only supported if data is detached: setting the streaming flag is
741 ignored for embedded content.
743 CMS support is disabled by default and must be explicitly enabled
744 with the enable-cms configuration option.
747 *) Update the GMP engine glue to do direct copies between BIGNUM and
748 mpz_t when openssl and GMP use the same limb size. Otherwise the
749 existing "conversion via a text string export" trick is still used.
750 [Paul Sheer <paulsheer@gmail.com>]
752 *) Zlib compression BIO. This is a filter BIO which compressed and
753 uncompresses any data passed through it.
756 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
757 RFC3394 compatible AES key wrapping.
760 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
761 sets string data without copying. X509_ALGOR_set0() and
762 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
763 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
764 from an X509_ATTRIBUTE structure optionally checking it occurs only
765 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
769 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
770 to get the expected BN_FLG_CONSTTIME behavior.
771 [Bodo Moeller (Google)]
775 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
776 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
777 - added some more tests to do_tests.pl
778 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
779 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
780 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
781 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
782 - various changes to netware.pl to enable gcc-cross builds on Win32
784 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
785 - various changes to fix missing prototype warnings
786 - fixed x86nasm.pl to create correct asm files for NASM COFF output
787 - added AES, WHIRLPOOL and CPUID assembler code to build files
788 - added missing AES assembler make rules to mk1mf.pl
789 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
790 [Guenter Knauf <eflash@gmx.net>]
792 *) Implement certificate status request TLS extension defined in RFC3546.
793 A client can set the appropriate parameters and receive the encoded
794 OCSP response via a callback. A server can query the supplied parameters
795 and set the encoded OCSP response in the callback. Add simplified examples
796 to s_client and s_server.
799 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
802 + Binary incompatibility of ssl_ctx_st structure
803 + DTLS interoperation with non-compliant servers
804 + Don't call get_session_cb() without proposed session
805 + Fix ia64 assembler code
806 [Andy Polyakov, Steve Henson]
808 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
810 *) DTLS Handshake overhaul. There were longstanding issues with
811 OpenSSL DTLS implementation, which were making it impossible for
812 RFC 4347 compliant client to communicate with OpenSSL server.
813 Unfortunately just fixing these incompatibilities would "cut off"
814 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
815 server keeps tolerating non RFC compliant syntax. The opposite is
816 not true, 0.9.8f client can not communicate with earlier server.
817 This update even addresses CVE-2007-4995.
820 *) Changes to avoid need for function casts in OpenSSL: some compilers
821 (gcc 4.2 and later) reject their use.
822 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
825 *) Add RFC4507 support to OpenSSL. This includes the corrections in
826 RFC4507bis. The encrypted ticket format is an encrypted encoded
827 SSL_SESSION structure, that way new session features are automatically
830 If a client application caches session in an SSL_SESSION structure
831 support is transparent because tickets are now stored in the encoded
834 The SSL_CTX structure automatically generates keys for ticket
835 protection in servers so again support should be possible
836 with no application modification.
838 If a client or server wishes to disable RFC4507 support then the option
839 SSL_OP_NO_TICKET can be set.
841 Add a TLS extension debugging callback to allow the contents of any client
842 or server extensions to be examined.
844 This work was sponsored by Google.
847 *) Add initial support for TLS extensions, specifically for the server_name
848 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
849 have new members for a host name. The SSL data structure has an
850 additional member SSL_CTX *initial_ctx so that new sessions can be
851 stored in that context to allow for session resumption, even after the
852 SSL has been switched to a new SSL_CTX in reaction to a client's
853 server_name extension.
855 New functions (subject to change):
858 SSL_get_servername_type()
861 New CTRL codes and macros (subject to change):
863 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
864 - SSL_CTX_set_tlsext_servername_callback()
865 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
866 - SSL_CTX_set_tlsext_servername_arg()
867 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
869 openssl s_client has a new '-servername ...' option.
871 openssl s_server has new options '-servername_host ...', '-cert2 ...',
872 '-key2 ...', '-servername_fatal' (subject to change). This allows
873 testing the HostName extension for a specific single host name ('-cert'
874 and '-key' remain fallbacks for handshakes without HostName
875 negotiation). If the unrecogninzed_name alert has to be sent, this by
876 default is a warning; it becomes fatal with the '-servername_fatal'
879 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
881 *) Add AES and SSE2 assembly language support to VC++ build.
884 *) Mitigate attack on final subtraction in Montgomery reduction.
887 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
888 (which previously caused an internal error).
891 *) Squeeze another 10% out of IGE mode when in != out.
894 *) AES IGE mode speedup.
895 [Dean Gaudet (Google)]
897 *) Add the Korean symmetric 128-bit cipher SEED (see
898 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
899 add SEED ciphersuites from RFC 4162:
901 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
902 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
903 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
904 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
906 To minimize changes between patchlevels in the OpenSSL 0.9.8
907 series, SEED remains excluded from compilation unless OpenSSL
908 is configured with 'enable-seed'.
911 *) Mitigate branch prediction attacks, which can be practical if a
912 single processor is shared, allowing a spy process to extract
913 information. For detailed background information, see
914 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
915 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
916 and Necessary Software Countermeasures"). The core of the change
917 are new versions BN_div_no_branch() and
918 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
919 respectively, which are slower, but avoid the security-relevant
920 conditional branches. These are automatically called by BN_div()
921 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
922 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
923 remove a conditional branch.
925 BN_FLG_CONSTTIME is the new name for the previous
926 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
927 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
928 in the exponent causes BN_mod_exp_mont() to use the alternative
929 implementation in BN_mod_exp_mont_consttime().) The old name
930 remains as a deprecated alias.
932 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
933 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
934 constant-time implementations for more than just exponentiation.
935 Here too the old name is kept as a deprecated alias.
937 BN_BLINDING_new() will now use BN_dup() for the modulus so that
938 the BN_BLINDING structure gets an independent copy of the
939 modulus. This means that the previous "BIGNUM *m" argument to
940 BN_BLINDING_new() and to BN_BLINDING_create_param() now
941 essentially becomes "const BIGNUM *m", although we can't actually
942 change this in the header file before 0.9.9. It allows
943 RSA_setup_blinding() to use BN_with_flags() on the modulus to
944 enable BN_FLG_CONSTTIME.
946 [Matthew D Wood (Intel Corp)]
948 *) In the SSL/TLS server implementation, be strict about session ID
949 context matching (which matters if an application uses a single
950 external cache for different purposes). Previously,
951 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
952 set. This did ensure strict client verification, but meant that,
953 with applications using a single external cache for quite
954 different requirements, clients could circumvent ciphersuite
955 restrictions for a given session ID context by starting a session
956 in a different context.
959 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
960 a ciphersuite string such as "DEFAULT:RSA" cannot enable
961 authentication-only ciphersuites.
964 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
965 not complete and could lead to a possible single byte overflow
966 (CVE-2007-5135) [Ben Laurie]
968 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
970 *) Since AES128 and AES256 (and similarly Camellia128 and
971 Camellia256) share a single mask bit in the logic of
972 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
973 kludge to work properly if AES128 is available and AES256 isn't
974 (or if Camellia128 is available and Camellia256 isn't).
977 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
978 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
979 When a point or a seed is encoded in a BIT STRING, we need to
980 prevent the removal of trailing zero bits to get the proper DER
981 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
982 of a NamedBitList, for which trailing 0 bits need to be removed.)
985 *) Have SSL/TLS server implementation tolerate "mismatched" record
986 protocol version while receiving ClientHello even if the
987 ClientHello is fragmented. (The server can't insist on the
988 particular protocol version it has chosen before the ServerHello
989 message has informed the client about his choice.)
992 *) Add RFC 3779 support.
993 [Rob Austein for ARIN, Ben Laurie]
995 *) Load error codes if they are not already present instead of using a
996 static variable. This allows them to be cleanly unloaded and reloaded.
997 Improve header file function name parsing.
1000 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
1001 or CAPABILITY handshake as required by RFCs.
1004 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
1006 *) Introduce limits to prevent malicious keys being able to
1007 cause a denial of service. (CVE-2006-2940)
1008 [Steve Henson, Bodo Moeller]
1010 *) Fix ASN.1 parsing of certain invalid structures that can result
1011 in a denial of service. (CVE-2006-2937) [Steve Henson]
1013 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1014 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1016 *) Fix SSL client code which could crash if connecting to a
1017 malicious SSLv2 server. (CVE-2006-4343)
1018 [Tavis Ormandy and Will Drewry, Google Security Team]
1020 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
1021 match only those. Before that, "AES256-SHA" would be interpreted
1022 as a pattern and match "AES128-SHA" too (since AES128-SHA got
1023 the same strength classification in 0.9.7h) as we currently only
1024 have a single AES bit in the ciphersuite description bitmap.
1025 That change, however, also applied to ciphersuite strings such as
1026 "RC4-MD5" that intentionally matched multiple ciphersuites --
1027 namely, SSL 2.0 ciphersuites in addition to the more common ones
1028 from SSL 3.0/TLS 1.0.
1030 So we change the selection algorithm again: Naming an explicit
1031 ciphersuite selects this one ciphersuite, and any other similar
1032 ciphersuite (same bitmap) from *other* protocol versions.
1033 Thus, "RC4-MD5" again will properly select both the SSL 2.0
1034 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
1036 Since SSL 2.0 does not have any ciphersuites for which the
1037 128/256 bit distinction would be relevant, this works for now.
1038 The proper fix will be to use different bits for AES128 and
1039 AES256, which would have avoided the problems from the beginning;
1040 however, bits are scarce, so we can only do this in a new release
1041 (not just a patchlevel) when we can change the SSL_CIPHER
1042 definition to split the single 'unsigned long mask' bitmap into
1043 multiple values to extend the available space.
1047 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
1049 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1050 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1052 *) Add AES IGE and biIGE modes.
1055 *) Change the Unix randomness entropy gathering to use poll() when
1056 possible instead of select(), since the latter has some
1057 undesirable limitations.
1058 [Darryl Miles via Richard Levitte and Bodo Moeller]
1060 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
1061 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
1062 cannot be implicitly activated as part of, e.g., the "AES" alias.
1063 However, please upgrade to OpenSSL 0.9.9[-dev] for
1064 non-experimental use of the ECC ciphersuites to get TLS extension
1065 support, which is required for curve and point format negotiation
1066 to avoid potential handshake problems.
1069 *) Disable rogue ciphersuites:
1071 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1072 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1073 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1075 The latter two were purportedly from
1076 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1079 Also deactivate the remaining ciphersuites from
1080 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1081 unofficial, and the ID has long expired.
1084 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1085 dual-core machines) and other potential thread-safety issues.
1088 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
1089 versions), which is now available for royalty-free use
1090 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
1091 Also, add Camellia TLS ciphersuites from RFC 4132.
1093 To minimize changes between patchlevels in the OpenSSL 0.9.8
1094 series, Camellia remains excluded from compilation unless OpenSSL
1095 is configured with 'enable-camellia'.
1098 *) Disable the padding bug check when compression is in use. The padding
1099 bug check assumes the first packet is of even length, this is not
1100 necessarily true if compresssion is enabled and can result in false
1101 positives causing handshake failure. The actual bug test is ancient
1102 code so it is hoped that implementations will either have fixed it by
1103 now or any which still have the bug do not support compression.
1106 Changes between 0.9.8a and 0.9.8b [04 May 2006]
1108 *) When applying a cipher rule check to see if string match is an explicit
1109 cipher suite and only match that one cipher suite if it is.
1112 *) Link in manifests for VC++ if needed.
1113 [Austin Ziegler <halostatue@gmail.com>]
1115 *) Update support for ECC-based TLS ciphersuites according to
1116 draft-ietf-tls-ecc-12.txt with proposed changes (but without
1117 TLS extensions, which are supported starting with the 0.9.9
1118 branch, not in the OpenSSL 0.9.8 branch).
1121 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
1122 opaque EVP_CIPHER_CTX handling.
1125 *) Fixes and enhancements to zlib compression code. We now only use
1126 "zlib1.dll" and use the default __cdecl calling convention on Win32
1127 to conform with the standards mentioned here:
1128 http://www.zlib.net/DLL_FAQ.txt
1129 Static zlib linking now works on Windows and the new --with-zlib-include
1130 --with-zlib-lib options to Configure can be used to supply the location
1131 of the headers and library. Gracefully handle case where zlib library
1135 *) Several fixes and enhancements to the OID generation code. The old code
1136 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
1137 handle numbers larger than ULONG_MAX, truncated printing and had a
1138 non standard OBJ_obj2txt() behaviour.
1141 *) Add support for building of engines under engine/ as shared libraries
1142 under VC++ build system.
1145 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
1146 Hopefully, we will not see any false combination of paths any more.
1149 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
1151 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1152 (part of SSL_OP_ALL). This option used to disable the
1153 countermeasure against man-in-the-middle protocol-version
1154 rollback in the SSL 2.0 server implementation, which is a bad
1155 idea. (CVE-2005-2969)
1157 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1158 for Information Security, National Institute of Advanced Industrial
1159 Science and Technology [AIST], Japan)]
1161 *) Add two function to clear and return the verify parameter flags.
1164 *) Keep cipherlists sorted in the source instead of sorting them at
1165 runtime, thus removing the need for a lock.
1168 *) Avoid some small subgroup attacks in Diffie-Hellman.
1169 [Nick Mathewson and Ben Laurie]
1171 *) Add functions for well-known primes.
1174 *) Extended Windows CE support.
1175 [Satoshi Nakamura and Andy Polyakov]
1177 *) Initialize SSL_METHOD structures at compile time instead of during
1178 runtime, thus removing the need for a lock.
1181 *) Make PKCS7_decrypt() work even if no certificate is supplied by
1182 attempting to decrypt each encrypted key in turn. Add support to
1186 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
1188 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
1191 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
1194 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
1195 key into the same file any more.
1198 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
1201 *) Add -utf8 command line and config file option to 'ca'.
1202 [Stefan <stf@udoma.org]
1204 *) Removed the macro des_crypt(), as it seems to conflict with some
1205 libraries. Use DES_crypt().
1208 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
1209 involves renaming the source and generated shared-libs for
1210 both. The engines will accept the corrected or legacy ids
1211 ('ncipher' and '4758_cca' respectively) when binding. NB,
1212 this only applies when building 'shared'.
1213 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
1215 *) Add attribute functions to EVP_PKEY structure. Modify
1216 PKCS12_create() to recognize a CSP name attribute and
1217 use it. Make -CSP option work again in pkcs12 utility.
1220 *) Add new functionality to the bn blinding code:
1221 - automatic re-creation of the BN_BLINDING parameters after
1222 a fixed number of uses (currently 32)
1223 - add new function for parameter creation
1224 - introduce flags to control the update behaviour of the
1225 BN_BLINDING parameters
1226 - hide BN_BLINDING structure
1227 Add a second BN_BLINDING slot to the RSA structure to improve
1228 performance when a single RSA object is shared among several
1232 *) Add support for DTLS.
1233 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
1235 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
1236 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
1239 *) Remove buggy and incompletet DH cert support from
1240 ssl/ssl_rsa.c and ssl/s3_both.c
1243 *) Use SHA-1 instead of MD5 as the default digest algorithm for
1244 the apps/openssl applications.
1247 *) Compile clean with "-Wall -Wmissing-prototypes
1248 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
1249 DEBUG_SAFESTACK must also be set.
1252 *) Change ./Configure so that certain algorithms can be disabled by default.
1253 The new counterpiece to "no-xxx" is "enable-xxx".
1255 The patented RC5 and MDC2 algorithms will now be disabled unless
1256 "enable-rc5" and "enable-mdc2", respectively, are specified.
1258 (IDEA remains enabled despite being patented. This is because IDEA
1259 is frequently required for interoperability, and there is no license
1260 fee for non-commercial use. As before, "no-idea" can be used to
1261 avoid this algorithm.)
1265 *) Add processing of proxy certificates (see RFC 3820). This work was
1266 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
1267 EGEE (Enabling Grids for E-science in Europe).
1270 *) RC4 performance overhaul on modern architectures/implementations, such
1271 as Intel P4, IA-64 and AMD64.
1274 *) New utility extract-section.pl. This can be used specify an alternative
1275 section number in a pod file instead of having to treat each file as
1276 a separate case in Makefile. This can be done by adding two lines to the
1279 =for comment openssl_section:XXX
1281 The blank line is mandatory.
1285 *) New arguments -certform, -keyform and -pass for s_client and s_server
1286 to allow alternative format key and certificate files and passphrase
1290 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
1291 update associated structures and add various utility functions.
1293 Add new policy related verify parameters, include policy checking in
1294 standard verify code. Enhance 'smime' application with extra parameters
1295 to support policy checking and print out.
1298 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
1299 Nehemiah processors. These extensions support AES encryption in hardware
1300 as well as RNG (though RNG support is currently disabled).
1301 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
1303 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
1306 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
1307 [Andy Polyakov and a number of other people]
1309 *) Improved PowerPC platform support. Most notably BIGNUM assembler
1310 implementation contributed by IBM.
1311 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
1313 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
1314 exponent rather than 'unsigned long'. There is a corresponding change to
1315 the new 'rsa_keygen' element of the RSA_METHOD structure.
1316 [Jelte Jansen, Geoff Thorpe]
1318 *) Functionality for creating the initial serial number file is now
1319 moved from CA.pl to the 'ca' utility with a new option -create_serial.
1321 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
1322 number file to 1, which is bound to cause problems. To avoid
1323 the problems while respecting compatibility between different 0.9.7
1324 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
1325 CA.pl for serial number initialization. With the new release 0.9.8,
1326 we can fix the problem directly in the 'ca' utility.)
1329 *) Reduced header interdepencies by declaring more opaque objects in
1330 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
1331 give fewer recursive includes, which could break lazy source code - so
1332 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
1333 developers should define this symbol when building and using openssl to
1334 ensure they track the recommended behaviour, interfaces, [etc], but
1335 backwards-compatible behaviour prevails when this isn't defined.
1338 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
1341 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
1342 This will generate a random key of the appropriate length based on the
1343 cipher context. The EVP_CIPHER can provide its own random key generation
1344 routine to support keys of a specific form. This is used in the des and
1345 3des routines to generate a key of the correct parity. Update S/MIME
1346 code to use new functions and hence generate correct parity DES keys.
1347 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
1348 valid (weak or incorrect parity).
1351 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
1352 as looking them up. This is useful when the verified structure may contain
1353 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
1354 present unless the new PKCS7_NO_CRL flag is asserted.
1357 *) Extend ASN1 oid configuration module. It now additionally accepts the
1360 shortName = some long name, 1.2.3.4
1363 *) Reimplemented the BN_CTX implementation. There is now no more static
1364 limitation on the number of variables it can handle nor the depth of the
1365 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
1366 information can now expand as required, and rather than having a single
1367 static array of bignums, BN_CTX now uses a linked-list of such arrays
1368 allowing it to expand on demand whilst maintaining the usefulness of
1369 BN_CTX's "bundling".
1372 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
1373 to allow all RSA operations to function using a single BN_CTX.
1376 *) Preliminary support for certificate policy evaluation and checking. This
1377 is initially intended to pass the tests outlined in "Conformance Testing
1378 of Relying Party Client Certificate Path Processing Logic" v1.07.
1381 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
1382 remained unused and not that useful. A variety of other little bignum
1383 tweaks and fixes have also been made continuing on from the audit (see
1387 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
1388 associated ASN1, EVP and SSL functions and old ASN1 macros.
1391 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
1392 and this should never fail. So the return value from the use of
1393 BN_set_word() (which can fail due to needless expansion) is now deprecated;
1394 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
1397 *) BN_CTX_get() should return zero-valued bignums, providing the same
1398 initialised value as BN_new().
1399 [Geoff Thorpe, suggested by Ulf Möller]
1401 *) Support for inhibitAnyPolicy certificate extension.
1404 *) An audit of the BIGNUM code is underway, for which debugging code is
1405 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
1406 is considered valid when processing BIGNUMs, and causes execution to
1407 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
1408 further steps are taken to deliberately pollute unused data in BIGNUM
1409 structures to try and expose faulty code further on. For now, openssl will
1410 (in its default mode of operation) continue to tolerate the inconsistent
1411 forms that it has tolerated in the past, but authors and packagers should
1412 consider trying openssl and their own applications when compiled with
1413 these debugging symbols defined. It will help highlight potential bugs in
1414 their own code, and will improve the test coverage for OpenSSL itself. At
1415 some point, these tighter rules will become openssl's default to improve
1416 maintainability, though the assert()s and other overheads will remain only
1417 in debugging configurations. See bn.h for more details.
1418 [Geoff Thorpe, Nils Larsch, Ulf Möller]
1420 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
1421 that can only be obtained through BN_CTX_new() (which implicitly
1422 initialises it). The presence of this function only made it possible
1423 to overwrite an existing structure (and cause memory leaks).
1426 *) Because of the callback-based approach for implementing LHASH as a
1427 template type, lh_insert() adds opaque objects to hash-tables and
1428 lh_doall() or lh_doall_arg() are typically used with a destructor callback
1429 to clean up those corresponding objects before destroying the hash table
1430 (and losing the object pointers). So some over-zealous constifications in
1431 LHASH have been relaxed so that lh_insert() does not take (nor store) the
1432 objects as "const" and the lh_doall[_arg] callback wrappers are not
1433 prototyped to have "const" restrictions on the object pointers they are
1434 given (and so aren't required to cast them away any more).
1437 *) The tmdiff.h API was so ugly and minimal that our own timing utility
1438 (speed) prefers to use its own implementation. The two implementations
1439 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
1440 its object type properly exposed (MS_TM) instead of casting to/from "char
1441 *". This may still change yet if someone realises MS_TM and "ms_time_***"
1442 aren't necessarily the greatest nomenclatures - but this is what was used
1443 internally to the implementation so I've used that for now.
1446 *) Ensure that deprecated functions do not get compiled when
1447 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
1448 the self-tests were still using deprecated key-generation functions so
1449 these have been updated also.
1452 *) Reorganise PKCS#7 code to separate the digest location functionality
1453 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
1454 New function PKCS7_set_digest() to set the digest type for PKCS#7
1455 digestedData type. Add additional code to correctly generate the
1456 digestedData type and add support for this type in PKCS7 initialization
1460 *) New function PKCS7_set0_type_other() this initializes a PKCS7
1461 structure of type "other".
1464 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
1465 sure the loop does correctly stop and breaking ("division by zero")
1466 modulus operations are not performed. The (pre-generated) prime
1467 table crypto/bn/bn_prime.h was already correct, but it could not be
1468 re-generated on some platforms because of the "division by zero"
1469 situation in the script.
1470 [Ralf S. Engelschall]
1472 *) Update support for ECC-based TLS ciphersuites according to
1473 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
1474 SHA-1 now is only used for "small" curves (where the
1475 representation of a field element takes up to 24 bytes); for
1476 larger curves, the field element resulting from ECDH is directly
1477 used as premaster secret.
1478 [Douglas Stebila (Sun Microsystems Laboratories)]
1480 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
1481 curve secp160r1 to the tests.
1482 [Douglas Stebila (Sun Microsystems Laboratories)]
1484 *) Add the possibility to load symbols globally with DSO.
1485 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
1487 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
1488 control of the error stack.
1491 *) Add support for STORE in ENGINE.
1494 *) Add the STORE type. The intention is to provide a common interface
1495 to certificate and key stores, be they simple file-based stores, or
1496 HSM-type store, or LDAP stores, or...
1497 NOTE: The code is currently UNTESTED and isn't really used anywhere.
1500 *) Add a generic structure called OPENSSL_ITEM. This can be used to
1501 pass a list of arguments to any function as well as provide a way
1502 for a function to pass data back to the caller.
1505 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
1506 works like BUF_strdup() but can be used to duplicate a portion of
1507 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
1511 *) Add the function sk_find_ex() which works like sk_find(), but will
1512 return an index to an element even if an exact match couldn't be
1513 found. The index is guaranteed to point at the element where the
1514 searched-for key would be inserted to preserve sorting order.
1517 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
1518 takes an extra flags argument for optional functionality. Currently,
1519 the following flags are defined:
1521 OBJ_BSEARCH_VALUE_ON_NOMATCH
1522 This one gets OBJ_bsearch_ex() to return a pointer to the first
1523 element where the comparing function returns a negative or zero
1526 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
1527 This one gets OBJ_bsearch_ex() to return a pointer to the first
1528 element where the comparing function returns zero. This is useful
1529 if there are more than one element where the comparing function
1533 *) Make it possible to create self-signed certificates with 'openssl ca'
1534 in such a way that the self-signed certificate becomes part of the
1535 CA database and uses the same mechanisms for serial number generation
1536 as all other certificate signing. The new flag '-selfsign' enables
1537 this functionality. Adapt CA.sh and CA.pl.in.
1540 *) Add functionality to check the public key of a certificate request
1541 against a given private. This is useful to check that a certificate
1542 request can be signed by that key (self-signing).
1545 *) Make it possible to have multiple active certificates with the same
1546 subject in the CA index file. This is done only if the keyword
1547 'unique_subject' is set to 'no' in the main CA section (default
1548 if 'CA_default') of the configuration file. The value is saved
1549 with the database itself in a separate index attribute file,
1550 named like the index file with '.attr' appended to the name.
1553 *) Generate muti valued AVAs using '+' notation in config files for
1557 *) Support for nameConstraints certificate extension.
1560 *) Support for policyConstraints certificate extension.
1563 *) Support for policyMappings certificate extension.
1566 *) Make sure the default DSA_METHOD implementation only uses its
1567 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
1568 and change its own handlers to be NULL so as to remove unnecessary
1569 indirection. This lets alternative implementations fallback to the
1570 default implementation more easily.
1573 *) Support for directoryName in GeneralName related extensions
1577 *) Make it possible to link applications using Makefile.shared.
1578 Make that possible even when linking against static libraries!
1581 *) Support for single pass processing for S/MIME signing. This now
1582 means that S/MIME signing can be done from a pipe, in addition
1583 cleartext signing (multipart/signed type) is effectively streaming
1584 and the signed data does not need to be all held in memory.
1586 This is done with a new flag PKCS7_STREAM. When this flag is set
1587 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
1588 is done after the data is output (and digests calculated) in
1589 SMIME_write_PKCS7().
1592 *) Add full support for -rpath/-R, both in shared libraries and
1593 applications, at least on the platforms where it's known how
1597 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
1598 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
1599 will now compute a table of multiples of the generator that
1600 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
1601 faster (notably in the case of a single point multiplication,
1602 scalar * generator).
1603 [Nils Larsch, Bodo Moeller]
1605 *) IPv6 support for certificate extensions. The various extensions
1606 which use the IP:a.b.c.d can now take IPv6 addresses using the
1607 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
1611 *) Added an ENGINE that implements RSA by performing private key
1612 exponentiations with the GMP library. The conversions to and from
1613 GMP's mpz_t format aren't optimised nor are any montgomery forms
1614 cached, and on x86 it appears OpenSSL's own performance has caught up.
1615 However there are likely to be other architectures where GMP could
1616 provide a boost. This ENGINE is not built in by default, but it can be
1617 specified at Configure time and should be accompanied by the necessary
1618 linker additions, eg;
1619 ./config -DOPENSSL_USE_GMP -lgmp
1622 *) "openssl engine" will not display ENGINE/DSO load failure errors when
1623 testing availability of engines with "-t" - the old behaviour is
1624 produced by increasing the feature's verbosity with "-tt".
1627 *) ECDSA routines: under certain error conditions uninitialized BN objects
1628 could be freed. Solution: make sure initialization is performed early
1629 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
1633 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
1634 and DH_METHOD (eg. by ENGINE implementations) to override the normal
1635 software implementations. For DSA and DH, parameter generation can
1636 also be overriden by providing the appropriate method callbacks.
1639 *) Change the "progress" mechanism used in key-generation and
1640 primality testing to functions that take a new BN_GENCB pointer in
1641 place of callback/argument pairs. The new API functions have "_ex"
1642 postfixes and the older functions are reimplemented as wrappers for
1643 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
1644 declarations of the old functions to help (graceful) attempts to
1645 migrate to the new functions. Also, the new key-generation API
1646 functions operate on a caller-supplied key-structure and return
1647 success/failure rather than returning a key or NULL - this is to
1648 help make "keygen" another member function of RSA_METHOD etc.
1650 Example for using the new callback interface:
1652 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
1656 BN_GENCB_set(&my_cb, my_callback, my_arg);
1658 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
1659 /* For the meaning of a, b in calls to my_callback(), see the
1660 * documentation of the function that calls the callback.
1661 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
1662 * my_callback should return 1 if it wants BN_is_prime_ex()
1663 * to continue, or 0 to stop.
1668 *) Change the ZLIB compression method to be stateful, and make it
1669 available to TLS with the number defined in
1670 draft-ietf-tls-compression-04.txt.
1673 *) Add the ASN.1 structures and functions for CertificatePair, which
1674 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
1676 CertificatePair ::= SEQUENCE {
1677 forward [0] Certificate OPTIONAL,
1678 reverse [1] Certificate OPTIONAL,
1679 -- at least one of the pair shall be present -- }
1681 Also implement the PEM functions to read and write certificate
1682 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
1684 This needed to be defined, mostly for the sake of the LDAP
1685 attribute crossCertificatePair, but may prove useful elsewhere as
1689 *) Make it possible to inhibit symlinking of shared libraries in
1690 Makefile.shared, for Cygwin's sake.
1693 *) Extend the BIGNUM API by creating a function
1694 void BN_set_negative(BIGNUM *a, int neg);
1695 and a macro that behave like
1696 int BN_is_negative(const BIGNUM *a);
1698 to avoid the need to access 'a->neg' directly in applications.
1701 *) Implement fast modular reduction for pseudo-Mersenne primes
1702 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
1703 EC_GROUP_new_curve_GFp() will now automatically use this
1705 [Nils Larsch <nla@trustcenter.de>]
1707 *) Add new lock type (CRYPTO_LOCK_BN).
1710 *) Change the ENGINE framework to automatically load engines
1711 dynamically from specific directories unless they could be
1712 found to already be built in or loaded. Move all the
1713 current engines except for the cryptodev one to a new
1715 The engines in engines/ are built as shared libraries if
1716 the "shared" options was given to ./Configure or ./config.
1717 Otherwise, they are inserted in libcrypto.a.
1718 /usr/local/ssl/engines is the default directory for dynamic
1719 engines, but that can be overriden at configure time through
1720 the usual use of --prefix and/or --openssldir, and at run
1721 time with the environment variable OPENSSL_ENGINES.
1722 [Geoff Thorpe and Richard Levitte]
1724 *) Add Makefile.shared, a helper makefile to build shared
1725 libraries. Addapt Makefile.org.
1728 *) Add version info to Win32 DLLs.
1729 [Peter 'Luna' Runestig" <peter@runestig.com>]
1731 *) Add new 'medium level' PKCS#12 API. Certificates and keys
1732 can be added using this API to created arbitrary PKCS#12
1733 files while avoiding the low level API.
1735 New options to PKCS12_create(), key or cert can be NULL and
1736 will then be omitted from the output file. The encryption
1737 algorithm NIDs can be set to -1 for no encryption, the mac
1738 iteration count can be set to 0 to omit the mac.
1740 Enhance pkcs12 utility by making the -nokeys and -nocerts
1741 options work when creating a PKCS#12 file. New option -nomac
1742 to omit the mac, NONE can be set for an encryption algorithm.
1743 New code is modified to use the enhanced PKCS12_create()
1744 instead of the low level API.
1747 *) Extend ASN1 encoder to support indefinite length constructed
1748 encoding. This can output sequences tags and octet strings in
1749 this form. Modify pk7_asn1.c to support indefinite length
1750 encoding. This is experimental and needs additional code to
1751 be useful, such as an ASN1 bio and some enhanced streaming
1754 Extend template encode functionality so that tagging is passed
1755 down to the template encoder.
1758 *) Let 'openssl req' fail if an argument to '-newkey' is not
1759 recognized instead of using RSA as a default.
1762 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
1763 As these are not official, they are not included in "ALL";
1764 the "ECCdraft" ciphersuite group alias can be used to select them.
1765 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
1767 *) Add ECDH engine support.
1768 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
1770 *) Add ECDH in new directory crypto/ecdh/.
1771 [Douglas Stebila (Sun Microsystems Laboratories)]
1773 *) Let BN_rand_range() abort with an error after 100 iterations
1774 without success (which indicates a broken PRNG).
1777 *) Change BN_mod_sqrt() so that it verifies that the input value
1778 is really the square of the return value. (Previously,
1779 BN_mod_sqrt would show GIGO behaviour.)
1782 *) Add named elliptic curves over binary fields from X9.62, SECG,
1783 and WAP/WTLS; add OIDs that were still missing.
1785 [Sheueling Chang Shantz and Douglas Stebila
1786 (Sun Microsystems Laboratories)]
1788 *) Extend the EC library for elliptic curves over binary fields
1789 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
1792 EC_GF2m_simple_method
1796 EC_GROUP_new_curve_GF2m
1797 EC_GROUP_set_curve_GF2m
1798 EC_GROUP_get_curve_GF2m
1799 EC_POINT_set_affine_coordinates_GF2m
1800 EC_POINT_get_affine_coordinates_GF2m
1801 EC_POINT_set_compressed_coordinates_GF2m
1803 Point compression for binary fields is disabled by default for
1804 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
1807 As binary polynomials are represented as BIGNUMs, various members
1808 of the EC_GROUP and EC_POINT data structures can be shared
1809 between the implementations for prime fields and binary fields;
1810 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
1811 are essentially identical to their ..._GFp counterparts.
1812 (For simplicity, the '..._GFp' prefix has been dropped from
1813 various internal method names.)
1815 An internal 'field_div' method (similar to 'field_mul' and
1816 'field_sqr') has been added; this is used only for binary fields.
1818 [Sheueling Chang Shantz and Douglas Stebila
1819 (Sun Microsystems Laboratories)]
1821 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
1822 through methods ('mul', 'precompute_mult').
1824 The generic implementations (now internally called 'ec_wNAF_mul'
1825 and 'ec_wNAF_precomputed_mult') remain the default if these
1826 methods are undefined.
1828 [Sheueling Chang Shantz and Douglas Stebila
1829 (Sun Microsystems Laboratories)]
1831 *) New function EC_GROUP_get_degree, which is defined through
1832 EC_METHOD. For curves over prime fields, this returns the bit
1833 length of the modulus.
1835 [Sheueling Chang Shantz and Douglas Stebila
1836 (Sun Microsystems Laboratories)]
1838 *) New functions EC_GROUP_dup, EC_POINT_dup.
1839 (These simply call ..._new and ..._copy).
1841 [Sheueling Chang Shantz and Douglas Stebila
1842 (Sun Microsystems Laboratories)]
1844 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
1845 Polynomials are represented as BIGNUMs (where the sign bit is not
1846 used) in the following functions [macros]:
1849 BN_GF2m_sub [= BN_GF2m_add]
1850 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
1851 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
1852 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
1854 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
1855 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
1856 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
1857 BN_GF2m_cmp [= BN_ucmp]
1859 (Note that only the 'mod' functions are actually for fields GF(2^m).
1860 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
1862 For some functions, an the irreducible polynomial defining a
1863 field can be given as an 'unsigned int[]' with strictly
1864 decreasing elements giving the indices of those bits that are set;
1865 i.e., p[] represents the polynomial
1866 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
1868 p[0] > p[1] > ... > p[k] = 0.
1869 This applies to the following functions:
1874 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
1875 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
1877 BN_GF2m_mod_sqrt_arr
1878 BN_GF2m_mod_solve_quad_arr
1882 Conversion can be performed by the following functions:
1887 bntest.c has additional tests for binary polynomial arithmetic.
1889 Two implementations for BN_GF2m_mod_div() are available.
1890 The default algorithm simply uses BN_GF2m_mod_inv() and
1891 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
1892 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
1893 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
1895 [Sheueling Chang Shantz and Douglas Stebila
1896 (Sun Microsystems Laboratories)]
1898 *) Add new error code 'ERR_R_DISABLED' that can be used when some
1899 functionality is disabled at compile-time.
1900 [Douglas Stebila <douglas.stebila@sun.com>]
1902 *) Change default behaviour of 'openssl asn1parse' so that more
1903 information is visible when viewing, e.g., a certificate:
1905 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
1906 mode the content of non-printable OCTET STRINGs is output in a
1907 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
1908 avoid the appearance of a printable string.
1909 [Nils Larsch <nla@trustcenter.de>]
1911 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
1913 EC_GROUP_set_asn1_flag()
1914 EC_GROUP_get_asn1_flag()
1915 EC_GROUP_set_point_conversion_form()
1916 EC_GROUP_get_point_conversion_form()
1917 These control ASN1 encoding details:
1918 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
1919 has been set to OPENSSL_EC_NAMED_CURVE.
1920 - Points are encoded in uncompressed form by default; options for
1921 asn1_for are as for point2oct, namely
1922 POINT_CONVERSION_COMPRESSED
1923 POINT_CONVERSION_UNCOMPRESSED
1924 POINT_CONVERSION_HYBRID
1926 Also add 'seed' and 'seed_len' members to EC_GROUP with access
1929 EC_GROUP_get0_seed()
1930 EC_GROUP_get_seed_len()
1931 This is used only for ASN1 purposes (so far).
1932 [Nils Larsch <nla@trustcenter.de>]
1934 *) Add 'field_type' member to EC_METHOD, which holds the NID
1935 of the appropriate field type OID. The new function
1936 EC_METHOD_get_field_type() returns this value.
1937 [Nils Larsch <nla@trustcenter.de>]
1942 EC_POINT_point2hex()
1943 EC_POINT_hex2point()
1944 providing useful interfaces to EC_POINT_point2oct() and
1945 EC_POINT_oct2point().
1946 [Nils Larsch <nla@trustcenter.de>]
1948 *) Change internals of the EC library so that the functions
1949 EC_GROUP_set_generator()
1950 EC_GROUP_get_generator()
1951 EC_GROUP_get_order()
1952 EC_GROUP_get_cofactor()
1953 are implemented directly in crypto/ec/ec_lib.c and not dispatched
1954 to methods, which would lead to unnecessary code duplication when
1955 adding different types of curves.
1956 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
1958 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
1959 arithmetic, and such that modified wNAFs are generated
1960 (which avoid length expansion in many cases).
1963 *) Add a function EC_GROUP_check_discriminant() (defined via
1964 EC_METHOD) that verifies that the curve discriminant is non-zero.
1966 Add a function EC_GROUP_check() that makes some sanity tests
1967 on a EC_GROUP, its generator and order. This includes
1968 EC_GROUP_check_discriminant().
1969 [Nils Larsch <nla@trustcenter.de>]
1971 *) Add ECDSA in new directory crypto/ecdsa/.
1973 Add applications 'openssl ecparam' and 'openssl ecdsa'
1974 (these are based on 'openssl dsaparam' and 'openssl dsa').
1976 ECDSA support is also included in various other files across the
1977 library. Most notably,
1978 - 'openssl req' now has a '-newkey ecdsa:file' option;
1979 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
1980 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
1981 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
1982 them suitable for ECDSA where domain parameters must be
1983 extracted before the specific public key;
1984 - ECDSA engine support has been added.
1985 [Nils Larsch <nla@trustcenter.de>]
1987 *) Include some named elliptic curves, and add OIDs from X9.62,
1988 SECG, and WAP/WTLS. Each curve can be obtained from the new
1990 EC_GROUP_new_by_curve_name(),
1991 and the list of available named curves can be obtained with
1992 EC_get_builtin_curves().
1993 Also add a 'curve_name' member to EC_GROUP objects, which can be
1995 EC_GROUP_set_curve_name()
1996 EC_GROUP_get_curve_name()
1997 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
1999 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2000 was actually never needed) and in BN_mul(). The removal in BN_mul()
2001 required a small change in bn_mul_part_recursive() and the addition
2002 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2003 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2004 bn_sub_words() and bn_add_words() except they take arrays with
2008 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
2010 *) Cleanse PEM buffers before freeing them since they may contain
2012 [Benjamin Bennett <ben@psc.edu>]
2014 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2015 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2016 authentication-only ciphersuites.
2019 *) Since AES128 and AES256 share a single mask bit in the logic of
2020 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2021 kludge to work properly if AES128 is available and AES256 isn't.
2024 *) Expand security boundary to match 1.1.1 module.
2027 *) Remove redundant features: hash file source, editing of test vectors
2028 modify fipsld to use external fips_premain.c signature.
2031 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
2032 run algorithm test programs.
2035 *) Make algorithm test programs more tolerant of whitespace.
2038 *) Have SSL/TLS server implementation tolerate "mismatched" record
2039 protocol version while receiving ClientHello even if the
2040 ClientHello is fragmented. (The server can't insist on the
2041 particular protocol version it has chosen before the ServerHello
2042 message has informed the client about his choice.)
2045 *) Load error codes if they are not already present instead of using a
2046 static variable. This allows them to be cleanly unloaded and reloaded.
2049 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
2051 *) Introduce limits to prevent malicious keys being able to
2052 cause a denial of service. (CVE-2006-2940)
2053 [Steve Henson, Bodo Moeller]
2055 *) Fix ASN.1 parsing of certain invalid structures that can result
2056 in a denial of service. (CVE-2006-2937) [Steve Henson]
2058 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2059 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2061 *) Fix SSL client code which could crash if connecting to a
2062 malicious SSLv2 server. (CVE-2006-4343)
2063 [Tavis Ormandy and Will Drewry, Google Security Team]
2065 *) Change ciphersuite string processing so that an explicit
2066 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
2067 will no longer include "AES128-SHA"), and any other similar
2068 ciphersuite (same bitmap) from *other* protocol versions (so that
2069 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
2070 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
2071 changes from 0.9.8b and 0.9.8d.
2074 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
2076 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2077 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2079 *) Change the Unix randomness entropy gathering to use poll() when
2080 possible instead of select(), since the latter has some
2081 undesirable limitations.
2082 [Darryl Miles via Richard Levitte and Bodo Moeller]
2084 *) Disable rogue ciphersuites:
2086 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2087 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2088 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2090 The latter two were purportedly from
2091 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2094 Also deactive the remaining ciphersuites from
2095 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2096 unofficial, and the ID has long expired.
2099 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2100 dual-core machines) and other potential thread-safety issues.
2103 Changes between 0.9.7i and 0.9.7j [04 May 2006]
2105 *) Adapt fipsld and the build system to link against the validated FIPS
2106 module in FIPS mode.
2109 *) Fixes for VC++ 2005 build under Windows.
2112 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
2113 from a Windows bash shell such as MSYS. It is autodetected from the
2114 "config" script when run from a VC++ environment. Modify standard VC++
2115 build to use fipscanister.o from the GNU make build.
2118 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
2120 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
2121 The value now differs depending on if you build for FIPS or not.
2122 BEWARE! A program linked with a shared FIPSed libcrypto can't be
2123 safely run with a non-FIPSed libcrypto, as it may crash because of
2124 the difference induced by this change.
2127 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
2129 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2130 (part of SSL_OP_ALL). This option used to disable the
2131 countermeasure against man-in-the-middle protocol-version
2132 rollback in the SSL 2.0 server implementation, which is a bad
2133 idea. (CVE-2005-2969)
2135 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2136 for Information Security, National Institute of Advanced Industrial
2137 Science and Technology [AIST], Japan)]
2139 *) Minimal support for X9.31 signatures and PSS padding modes. This is
2140 mainly for FIPS compliance and not fully integrated at this stage.
2143 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
2144 the exponentiation using a fixed-length exponent. (Otherwise,
2145 the information leaked through timing could expose the secret key
2146 after many signatures; cf. Bleichenbacher's attack on DSA with
2150 *) Make a new fixed-window mod_exp implementation the default for
2151 RSA, DSA, and DH private-key operations so that the sequence of
2152 squares and multiplies and the memory access pattern are
2153 independent of the particular secret key. This will mitigate
2154 cache-timing and potential related attacks.
2156 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
2157 and this is automatically used by BN_mod_exp_mont() if the new flag
2158 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
2159 will use this BN flag for private exponents unless the flag
2160 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
2161 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
2163 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
2165 *) Change the client implementation for SSLv23_method() and
2166 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
2167 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
2168 (Previously, the SSL 2.0 backwards compatible Client Hello
2169 message format would be used even with SSL_OP_NO_SSLv2.)
2172 *) Add support for smime-type MIME parameter in S/MIME messages which some
2176 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
2177 a threadsafe manner. Modify rsa code to use new function and add calls
2178 to dsa and dh code (which had race conditions before).
2181 *) Include the fixed error library code in the C error file definitions
2182 instead of fixing them up at runtime. This keeps the error code
2183 structures constant.
2186 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
2188 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
2191 *) Fixes for newer kerberos headers. NB: the casts are needed because
2192 the 'length' field is signed on one version and unsigned on another
2193 with no (?) obvious way to tell the difference, without these VC++
2194 complains. Also the "definition" of FAR (blank) is no longer included
2195 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
2196 some needed definitions.
2199 *) Undo Cygwin change.
2202 *) Added support for proxy certificates according to RFC 3820.
2203 Because they may be a security thread to unaware applications,
2204 they must be explicitely allowed in run-time. See
2205 docs/HOWTO/proxy_certificates.txt for further information.
2208 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
2210 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
2211 server and client random values. Previously
2212 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
2213 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
2215 This change has negligible security impact because:
2217 1. Server and client random values still have 24 bytes of pseudo random
2220 2. Server and client random values are sent in the clear in the initial
2223 3. The master secret is derived using the premaster secret (48 bytes in
2224 size for static RSA ciphersuites) as well as client server and random
2227 The OpenSSL team would like to thank the UK NISCC for bringing this issue
2230 [Stephen Henson, reported by UK NISCC]
2232 *) Use Windows randomness collection on Cygwin.
2235 *) Fix hang in EGD/PRNGD query when communication socket is closed
2236 prematurely by EGD/PRNGD.
2237 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
2239 *) Prompt for pass phrases when appropriate for PKCS12 input format.
2242 *) Back-port of selected performance improvements from development
2243 branch, as well as improved support for PowerPC platforms.
2246 *) Add lots of checks for memory allocation failure, error codes to indicate
2247 failure and freeing up memory if a failure occurs.
2248 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
2250 *) Add new -passin argument to dgst.
2253 *) Perform some character comparisons of different types in X509_NAME_cmp:
2254 this is needed for some certificates that reencode DNs into UTF8Strings
2255 (in violation of RFC3280) and can't or wont issue name rollover
2259 *) Make an explicit check during certificate validation to see that
2260 the CA setting in each certificate on the chain is correct. As a
2261 side effect always do the following basic checks on extensions,
2262 not just when there's an associated purpose to the check:
2264 - if there is an unhandled critical extension (unless the user
2265 has chosen to ignore this fault)
2266 - if the path length has been exceeded (if one is set at all)
2267 - that certain extensions fit the associated purpose (if one has
2271 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
2273 *) Avoid a race condition when CRLs are checked in a multi threaded
2274 environment. This would happen due to the reordering of the revoked
2275 entries during signature checking and serial number lookup. Now the
2276 encoding is cached and the serial number sort performed under a lock.
2277 Add new STACK function sk_is_sorted().
2280 *) Add Delta CRL to the extension code.
2283 *) Various fixes to s3_pkt.c so alerts are sent properly.
2284 [David Holmes <d.holmes@f5.com>]
2286 *) Reduce the chances of duplicate issuer name and serial numbers (in
2287 violation of RFC3280) using the OpenSSL certificate creation utilities.
2288 This is done by creating a random 64 bit value for the initial serial
2289 number when a serial number file is created or when a self signed
2290 certificate is created using 'openssl req -x509'. The initial serial
2291 number file is created using 'openssl x509 -next_serial' in CA.pl
2292 rather than being initialized to 1.
2295 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
2297 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
2298 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
2299 [Joe Orton, Steve Henson]
2301 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
2303 [Joe Orton, Steve Henson]
2305 *) Make it possible to have multiple active certificates with the same
2306 subject in the CA index file. This is done only if the keyword
2307 'unique_subject' is set to 'no' in the main CA section (default
2308 if 'CA_default') of the configuration file. The value is saved
2309 with the database itself in a separate index attribute file,
2310 named like the index file with '.attr' appended to the name.
2313 *) X509 verify fixes. Disable broken certificate workarounds when
2314 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
2315 keyUsage extension present. Don't accept CRLs with unhandled critical
2316 extensions: since verify currently doesn't process CRL extensions this
2317 rejects a CRL with *any* critical extensions. Add new verify error codes
2321 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
2322 A clarification of RFC2560 will require the use of OCTET STRINGs and
2323 some implementations cannot handle the current raw format. Since OpenSSL
2324 copies and compares OCSP nonces as opaque blobs without any attempt at
2325 parsing them this should not create any compatibility issues.
2328 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
2329 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
2330 this HMAC (and other) operations are several times slower than OpenSSL
2334 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
2335 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2337 *) Use the correct content when signing type "other".
2340 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
2342 *) Fix various bugs revealed by running the NISCC test suite:
2344 Stop out of bounds reads in the ASN1 code when presented with
2345 invalid tags (CVE-2003-0543 and CVE-2003-0544).
2347 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
2349 If verify callback ignores invalid public key errors don't try to check
2350 certificate signature with the NULL public key.
2354 *) New -ignore_err option in ocsp application to stop the server
2355 exiting on the first error in a request.
2358 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
2359 if the server requested one: as stated in TLS 1.0 and SSL 3.0
2363 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
2364 extra data after the compression methods not only for TLS 1.0
2365 but also for SSL 3.0 (as required by the specification).
2366 [Bodo Moeller; problem pointed out by Matthias Loepfe]
2368 *) Change X509_certificate_type() to mark the key as exported/exportable
2369 when it's 512 *bits* long, not 512 bytes.
2372 *) Change AES_cbc_encrypt() so it outputs exact multiple of
2373 blocks during encryption.
2376 *) Various fixes to base64 BIO and non blocking I/O. On write
2377 flushes were not handled properly if the BIO retried. On read
2378 data was not being buffered properly and had various logic bugs.
2379 This also affects blocking I/O when the data being decoded is a
2383 *) Various S/MIME bugfixes and compatibility changes:
2384 output correct application/pkcs7 MIME type if
2385 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
2386 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
2387 of files as .eml work). Correctly handle very long lines in MIME
2391 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
2393 *) Countermeasure against the Klima-Pokorny-Rosa extension of
2394 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
2395 a protocol version number mismatch like a decryption error
2396 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
2399 *) Turn on RSA blinding by default in the default implementation
2400 to avoid a timing attack. Applications that don't want it can call
2401 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
2402 They would be ill-advised to do so in most cases.
2403 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
2405 *) Change RSA blinding code so that it works when the PRNG is not
2406 seeded (in this case, the secret RSA exponent is abused as
2407 an unpredictable seed -- if it is not unpredictable, there
2408 is no point in blinding anyway). Make RSA blinding thread-safe
2409 by remembering the creator's thread ID in rsa->blinding and
2410 having all other threads use local one-time blinding factors
2411 (this requires more computation than sharing rsa->blinding, but
2412 avoids excessive locking; and if an RSA object is not shared
2413 between threads, blinding will still be very fast).
2416 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
2417 ENGINE as defaults for all supported algorithms irrespective of
2418 the 'flags' parameter. 'flags' is now honoured, so applications
2419 should make sure they are passing it correctly.
2422 *) Target "mingw" now allows native Windows code to be generated in
2423 the Cygwin environment as well as with the MinGW compiler.
2426 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
2428 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
2429 via timing by performing a MAC computation even if incorrrect
2430 block cipher padding has been found. This is a countermeasure
2431 against active attacks where the attacker has to distinguish
2432 between bad padding and a MAC verification error. (CVE-2003-0078)
2434 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
2435 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
2436 Martin Vuagnoux (EPFL, Ilion)]
2438 *) Make the no-err option work as intended. The intention with no-err
2439 is not to have the whole error stack handling routines removed from
2440 libcrypto, it's only intended to remove all the function name and
2441 reason texts, thereby removing some of the footprint that may not
2442 be interesting if those errors aren't displayed anyway.
2444 NOTE: it's still possible for any application or module to have it's
2445 own set of error texts inserted. The routines are there, just not
2446 used by default when no-err is given.
2449 *) Add support for FreeBSD on IA64.
2450 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
2452 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
2453 Kerberos function mit_des_cbc_cksum(). Before this change,
2454 the value returned by DES_cbc_cksum() was like the one from
2455 mit_des_cbc_cksum(), except the bytes were swapped.
2456 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
2458 *) Allow an application to disable the automatic SSL chain building.
2459 Before this a rather primitive chain build was always performed in
2460 ssl3_output_cert_chain(): an application had no way to send the
2461 correct chain if the automatic operation produced an incorrect result.
2463 Now the chain builder is disabled if either:
2465 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
2467 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
2469 The reasoning behind this is that an application would not want the
2470 auto chain building to take place if extra chain certificates are
2471 present and it might also want a means of sending no additional
2472 certificates (for example the chain has two certificates and the
2476 *) Add the possibility to build without the ENGINE framework.
2477 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2479 *) Under Win32 gmtime() can return NULL: check return value in
2480 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
2483 *) DSA routines: under certain error conditions uninitialized BN objects
2484 could be freed. Solution: make sure initialization is performed early
2485 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
2486 Nils Larsch <nla@trustcenter.de> via PR#459)
2489 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
2490 checked on reconnect on the client side, therefore session resumption
2491 could still fail with a "ssl session id is different" error. This
2492 behaviour is masked when SSL_OP_ALL is used due to
2493 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
2494 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2495 followup to PR #377.
2498 *) IA-32 assembler support enhancements: unified ELF targets, support
2499 for SCO/Caldera platforms, fix for Cygwin shared build.
2502 *) Add support for FreeBSD on sparc64. As a consequence, support for
2503 FreeBSD on non-x86 processors is separate from x86 processors on
2504 the config script, much like the NetBSD support.
2505 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
2507 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
2509 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
2512 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
2513 code (06) was taken as the first octet of the session ID and the last
2514 octet was ignored consequently. As a result SSLv2 client side session
2515 caching could not have worked due to the session ID mismatch between
2517 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2521 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
2522 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
2526 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
2527 seems that in spite of existing for more than a year, many application
2528 author have done nothing to provide the necessary callbacks, which
2529 means that this particular engine will not work properly anywhere.
2530 This is a very unfortunate situation which forces us, in the name
2531 of usability, to give the hw_ncipher.c a static lock, which is part
2533 NOTE: This is for the 0.9.7 series ONLY. This hack will never
2534 appear in 0.9.8 or later. We EXPECT application authors to have
2535 dealt properly with this when 0.9.8 is released (unless we actually
2536 make such changes in the libcrypto locking code that changes will
2537 have to be made anyway).
2540 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
2541 octets have been read, EOF or an error occurs. Without this change
2542 some truncated ASN1 structures will not produce an error.
2545 *) Disable Heimdal support, since it hasn't been fully implemented.
2546 Still give the possibility to force the use of Heimdal, but with
2547 warnings and a request that patches get sent to openssl-dev.
2550 *) Add the VC-CE target, introduce the WINCE sysname, and add
2551 INSTALL.WCE and appropriate conditionals to make it build.
2552 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2554 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
2555 cygssl-x.y.z.dll, where x, y and z are the major, minor and
2556 edit numbers of the version.
2557 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2559 *) Introduce safe string copy and catenation functions
2560 (BUF_strlcpy() and BUF_strlcat()).
2561 [Ben Laurie (CHATS) and Richard Levitte]
2563 *) Avoid using fixed-size buffers for one-line DNs.
2564 [Ben Laurie (CHATS)]
2566 *) Add BUF_MEM_grow_clean() to avoid information leakage when
2567 resizing buffers containing secrets, and use where appropriate.
2568 [Ben Laurie (CHATS)]
2570 *) Avoid using fixed size buffers for configuration file location.
2571 [Ben Laurie (CHATS)]
2573 *) Avoid filename truncation for various CA files.
2574 [Ben Laurie (CHATS)]
2576 *) Use sizeof in preference to magic numbers.
2577 [Ben Laurie (CHATS)]
2579 *) Avoid filename truncation in cert requests.
2580 [Ben Laurie (CHATS)]
2582 *) Add assertions to check for (supposedly impossible) buffer
2584 [Ben Laurie (CHATS)]
2586 *) Don't cache truncated DNS entries in the local cache (this could
2587 potentially lead to a spoofing attack).
2588 [Ben Laurie (CHATS)]
2590 *) Fix various buffers to be large enough for hex/decimal
2591 representations in a platform independent manner.
2592 [Ben Laurie (CHATS)]
2594 *) Add CRYPTO_realloc_clean() to avoid information leakage when
2595 resizing buffers containing secrets, and use where appropriate.
2596 [Ben Laurie (CHATS)]
2598 *) Add BIO_indent() to avoid much slightly worrying code to do
2600 [Ben Laurie (CHATS)]
2602 *) Convert sprintf()/BIO_puts() to BIO_printf().
2603 [Ben Laurie (CHATS)]
2605 *) buffer_gets() could terminate with the buffer only half
2607 [Ben Laurie (CHATS)]
2609 *) Add assertions to prevent user-supplied crypto functions from
2610 overflowing internal buffers by having large block sizes, etc.
2611 [Ben Laurie (CHATS)]
2613 *) New OPENSSL_assert() macro (similar to assert(), but enabled
2615 [Ben Laurie (CHATS)]
2617 *) Eliminate unused copy of key in RC4.
2618 [Ben Laurie (CHATS)]
2620 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
2621 [Ben Laurie (CHATS)]
2623 *) Fix off-by-one error in EGD path.
2624 [Ben Laurie (CHATS)]
2626 *) If RANDFILE path is too long, ignore instead of truncating.
2627 [Ben Laurie (CHATS)]
2629 *) Eliminate unused and incorrectly sized X.509 structure
2631 [Ben Laurie (CHATS)]
2633 *) Eliminate unused and dangerous function knumber().
2634 [Ben Laurie (CHATS)]
2636 *) Eliminate unused and dangerous structure, KSSL_ERR.
2637 [Ben Laurie (CHATS)]
2639 *) Protect against overlong session ID context length in an encoded
2640 session object. Since these are local, this does not appear to be
2642 [Ben Laurie (CHATS)]
2644 *) Change from security patch (see 0.9.6e below) that did not affect
2645 the 0.9.6 release series:
2647 Remote buffer overflow in SSL3 protocol - an attacker could
2648 supply an oversized master key in Kerberos-enabled versions.
2650 [Ben Laurie (CHATS)]
2652 *) Change the SSL kerb5 codes to match RFC 2712.
2655 *) Make -nameopt work fully for req and add -reqopt switch.
2656 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
2658 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
2659 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
2661 *) Make sure tests can be performed even if the corresponding algorithms
2662 have been removed entirely. This was also the last step to make
2663 OpenSSL compilable with DJGPP under all reasonable conditions.
2664 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
2666 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
2667 to allow version independent disabling of normally unselected ciphers,
2668 which may be activated as a side-effect of selecting a single cipher.
2670 (E.g., cipher list string "RSA" enables ciphersuites that are left
2671 out of "ALL" because they do not provide symmetric encryption.
2672 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
2673 [Lutz Jaenicke, Bodo Moeller]
2675 *) Add appropriate support for separate platform-dependent build
2676 directories. The recommended way to make a platform-dependent
2677 build directory is the following (tested on Linux), maybe with
2680 # Place yourself outside of the OpenSSL source tree. In
2681 # this example, the environment variable OPENSSL_SOURCE
2682 # is assumed to contain the absolute OpenSSL source directory.
2683 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
2684 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
2685 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
2686 mkdir -p `dirname $F`
2687 ln -s $OPENSSL_SOURCE/$F $F
2690 To be absolutely sure not to disturb the source tree, a "make clean"
2691 is a good thing. If it isn't successfull, don't worry about it,
2692 it probably means the source directory is very clean.
2695 *) Make sure any ENGINE control commands make local copies of string
2696 pointers passed to them whenever necessary. Otherwise it is possible
2697 the caller may have overwritten (or deallocated) the original string
2698 data when a later ENGINE operation tries to use the stored values.
2699 [Götz Babin-Ebell <babinebell@trustcenter.de>]
2701 *) Improve diagnostics in file reading and command-line digests.
2702 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
2704 *) Add AES modes CFB and OFB to the object database. Correct an
2705 error in AES-CFB decryption.
2708 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
2709 allows existing EVP_CIPHER_CTX structures to be reused after
2710 calling EVP_*Final(). This behaviour is used by encryption
2711 BIOs and some applications. This has the side effect that
2712 applications must explicitly clean up cipher contexts with
2713 EVP_CIPHER_CTX_cleanup() or they will leak memory.
2716 *) Check the values of dna and dnb in bn_mul_recursive before calling
2717 bn_mul_comba (a non zero value means the a or b arrays do not contain
2718 n2 elements) and fallback to bn_mul_normal if either is not zero.
2721 *) Fix escaping of non-ASCII characters when using the -subj option
2722 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
2725 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
2726 form for "surname", serialNumber has no short form.
2727 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
2728 therefore remove "mail" short name for "internet 7".
2729 The OID for unique identifiers in X509 certificates is
2730 x500UniqueIdentifier, not uniqueIdentifier.
2731 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
2734 *) Add an "init" command to the ENGINE config module and auto initialize
2735 ENGINEs. Without any "init" command the ENGINE will be initialized
2736 after all ctrl commands have been executed on it. If init=1 the
2737 ENGINE is initailized at that point (ctrls before that point are run
2738 on the uninitialized ENGINE and after on the initialized one). If
2739 init=0 then the ENGINE will not be iniatialized at all.
2742 *) Fix the 'app_verify_callback' interface so that the user-defined
2743 argument is actually passed to the callback: In the
2744 SSL_CTX_set_cert_verify_callback() prototype, the callback
2745 declaration has been changed from
2748 int (*cb)(X509_STORE_CTX *,void *);
2749 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
2750 i=s->ctx->app_verify_callback(&ctx)
2751 has been changed into
2752 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
2754 To update applications using SSL_CTX_set_cert_verify_callback(),
2755 a dummy argument can be added to their callback functions.
2756 [D. K. Smetters <smetters@parc.xerox.com>]
2758 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
2759 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
2761 *) Add and OPENSSL_LOAD_CONF define which will cause
2762 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
2763 This allows older applications to transparently support certain
2764 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
2765 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
2766 load the config file and OPENSSL_add_all_algorithms_conf() which will
2767 always load it have also been added.
2770 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
2771 Adjust NIDs and EVP layer.
2772 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2774 *) Config modules support in openssl utility.
2776 Most commands now load modules from the config file,
2777 though in a few (such as version) this isn't done
2778 because it couldn't be used for anything.
2780 In the case of ca and req the config file used is
2781 the same as the utility itself: that is the -config
2782 command line option can be used to specify an
2786 *) Move default behaviour from OPENSSL_config(). If appname is NULL
2787 use "openssl_conf" if filename is NULL use default openssl config file.
2790 *) Add an argument to OPENSSL_config() to allow the use of an alternative
2791 config section name. Add a new flag to tolerate a missing config file
2792 and move code to CONF_modules_load_file().
2795 *) Support for crypto accelerator cards from Accelerated Encryption
2796 Processing, www.aep.ie. (Use engine 'aep')
2797 The support was copied from 0.9.6c [engine] and adapted/corrected
2798 to work with the new engine framework.
2799 [AEP Inc. and Richard Levitte]
2801 *) Support for SureWare crypto accelerator cards from Baltimore
2802 Technologies. (Use engine 'sureware')
2803 The support was copied from 0.9.6c [engine] and adapted
2804 to work with the new engine framework.
2807 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
2808 make the newer ENGINE framework commands for the CHIL engine work.
2809 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
2811 *) Make it possible to produce shared libraries on ReliantUNIX.
2812 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
2814 *) Add the configuration target debug-linux-ppro.
2815 Make 'openssl rsa' use the general key loading routines
2816 implemented in apps.c, and make those routines able to
2817 handle the key format FORMAT_NETSCAPE and the variant
2819 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2821 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2822 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2824 *) Add -keyform to rsautl, and document -engine.
2825 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
2827 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
2828 BIO_R_NO_SUCH_FILE error code rather than the generic
2829 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
2832 *) Add new functions
2834 ERR_peek_last_error_line
2835 ERR_peek_last_error_line_data.
2836 These are similar to
2839 ERR_peek_error_line_data,
2840 but report on the latest error recorded rather than the first one
2841 still in the error queue.
2842 [Ben Laurie, Bodo Moeller]
2844 *) default_algorithms option in ENGINE config module. This allows things
2846 default_algorithms = ALL
2847 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
2850 *) Prelminary ENGINE config module.
2853 *) New experimental application configuration code.
2856 *) Change the AES code to follow the same name structure as all other
2857 symmetric ciphers, and behave the same way. Move everything to
2858 the directory crypto/aes, thereby obsoleting crypto/rijndael.
2859 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2861 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
2862 [Ben Laurie and Theo de Raadt]
2864 *) Add option to output public keys in req command.
2865 [Massimiliano Pala madwolf@openca.org]
2867 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
2868 (up to about 10% better than before for P-192 and P-224).
2871 *) New functions/macros
2873 SSL_CTX_set_msg_callback(ctx, cb)
2874 SSL_CTX_set_msg_callback_arg(ctx, arg)
2875 SSL_set_msg_callback(ssl, cb)
2876 SSL_set_msg_callback_arg(ssl, arg)
2878 to request calling a callback function
2880 void cb(int write_p, int version, int content_type,
2881 const void *buf, size_t len, SSL *ssl, void *arg)
2883 whenever a protocol message has been completely received
2884 (write_p == 0) or sent (write_p == 1). Here 'version' is the
2885 protocol version according to which the SSL library interprets
2886 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
2887 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
2888 the content type as defined in the SSL 3.0/TLS 1.0 protocol
2889 specification (change_cipher_spec(20), alert(21), handshake(22)).
2890 'buf' and 'len' point to the actual message, 'ssl' to the
2891 SSL object, and 'arg' is the application-defined value set by
2892 SSL[_CTX]_set_msg_callback_arg().
2894 'openssl s_client' and 'openssl s_server' have new '-msg' options
2895 to enable a callback that displays all protocol messages.
2898 *) Change the shared library support so shared libraries are built as
2899 soon as the corresponding static library is finished, and thereby get
2900 openssl and the test programs linked against the shared library.
2901 This still only happens when the keyword "shard" has been given to
2902 the configuration scripts.
2904 NOTE: shared library support is still an experimental thing, and
2905 backward binary compatibility is still not guaranteed.
2906 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
2908 *) Add support for Subject Information Access extension.
2909 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2911 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
2912 additional bytes when new memory had to be allocated, not just
2913 when reusing an existing buffer.
2916 *) New command line and configuration option 'utf8' for the req command.
2917 This allows field values to be specified as UTF8 strings.
2920 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
2921 runs for the former and machine-readable output for the latter.
2924 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
2925 of the e-mail address in the DN (i.e., it will go into a certificate
2926 extension only). The new configuration file option 'email_in_dn = no'
2927 has the same effect.
2928 [Massimiliano Pala madwolf@openca.org]
2930 *) Change all functions with names starting with des_ to be starting
2931 with DES_ instead. Add wrappers that are compatible with libdes,
2932 but are named _ossl_old_des_*. Finally, add macros that map the
2933 des_* symbols to the corresponding _ossl_old_des_* if libdes
2934 compatibility is desired. If OpenSSL 0.9.6c compatibility is
2935 desired, the des_* symbols will be mapped to DES_*, with one
2938 Since we provide two compatibility mappings, the user needs to
2939 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
2940 compatibility is desired. The default (i.e., when that macro
2941 isn't defined) is OpenSSL 0.9.6c compatibility.
2943 There are also macros that enable and disable the support of old
2944 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
2945 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
2946 are defined, the default will apply: to support the old des routines.
2948 In either case, one must include openssl/des.h to get the correct
2949 definitions. Do not try to just include openssl/des_old.h, that
2952 NOTE: This is a major break of an old API into a new one. Software
2953 authors are encouraged to switch to the DES_ style functions. Some
2954 time in the future, des_old.h and the libdes compatibility functions
2955 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
2956 default), and then completely removed.
2959 *) Test for certificates which contain unsupported critical extensions.
2960 If such a certificate is found during a verify operation it is
2961 rejected by default: this behaviour can be overridden by either
2962 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
2963 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
2964 X509_supported_extension() has also been added which returns 1 if a
2965 particular extension is supported.
2968 *) Modify the behaviour of EVP cipher functions in similar way to digests
2969 to retain compatibility with existing code.
2972 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
2973 compatibility with existing code. In particular the 'ctx' parameter does
2974 not have to be to be initialized before the call to EVP_DigestInit() and
2975 it is tidied up after a call to EVP_DigestFinal(). New function
2976 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
2977 EVP_MD_CTX_copy() changed to not require the destination to be
2978 initialized valid and new function EVP_MD_CTX_copy_ex() added which
2979 requires the destination to be valid.
2981 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
2982 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
2985 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
2986 so that complete 'Handshake' protocol structures are kept in memory
2987 instead of overwriting 'msg_type' and 'length' with 'body' data.
2990 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
2991 [Massimo Santin via Richard Levitte]
2993 *) Major restructuring to the underlying ENGINE code. This includes
2994 reduction of linker bloat, separation of pure "ENGINE" manipulation
2995 (initialisation, etc) from functionality dealing with implementations
2996 of specific crypto iterfaces. This change also introduces integrated
2997 support for symmetric ciphers and digest implementations - so ENGINEs
2998 can now accelerate these by providing EVP_CIPHER and EVP_MD
2999 implementations of their own. This is detailed in crypto/engine/README
3000 as it couldn't be adequately described here. However, there are a few
3001 API changes worth noting - some RSA, DSA, DH, and RAND functions that
3002 were changed in the original introduction of ENGINE code have now
3003 reverted back - the hooking from this code to ENGINE is now a good
3004 deal more passive and at run-time, operations deal directly with
3005 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
3006 dereferencing through an ENGINE pointer any more. Also, the ENGINE
3007 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
3008 they were not being used by the framework as there is no concept of a
3009 BIGNUM_METHOD and they could not be generalised to the new
3010 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
3011 ENGINE_cpy() has been removed as it cannot be consistently defined in
3015 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
3018 *) Change mkdef.pl to sort symbols that get the same entry number,
3019 and make sure the automatically generated functions ERR_load_*
3020 become part of libeay.num as well.
3023 *) New function SSL_renegotiate_pending(). This returns true once
3024 renegotiation has been requested (either SSL_renegotiate() call
3025 or HelloRequest/ClientHello receveived from the peer) and becomes
3026 false once a handshake has been completed.
3027 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
3028 sends a HelloRequest, but does not ensure that a handshake takes
3029 place. SSL_renegotiate_pending() is useful for checking if the
3030 client has followed the request.)
3033 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
3034 By default, clients may request session resumption even during
3035 renegotiation (if session ID contexts permit); with this option,
3036 session resumption is possible only in the first handshake.
3038 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
3039 more bits available for options that should not be part of
3040 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
3043 *) Add some demos for certificate and certificate request creation.
3046 *) Make maximum certificate chain size accepted from the peer application
3047 settable (SSL*_get/set_max_cert_list()), as proposed by
3048 "Douglas E. Engert" <deengert@anl.gov>.
3051 *) Add support for shared libraries for Unixware-7
3052 (Boyd Lynn Gerber <gerberb@zenez.com>).
3055 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
3056 be done prior to destruction. Use this to unload error strings from
3057 ENGINEs that load their own error strings. NB: This adds two new API
3058 functions to "get" and "set" this destroy handler in an ENGINE.
3061 *) Alter all existing ENGINE implementations (except "openssl" and
3062 "openbsd") to dynamically instantiate their own error strings. This
3063 makes them more flexible to be built both as statically-linked ENGINEs
3064 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
3065 Also, add stub code to each that makes building them as self-contained
3066 shared-libraries easier (see README.ENGINE).
3069 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
3070 implementations into applications that are completely implemented in
3071 self-contained shared-libraries. The "dynamic" ENGINE exposes control
3072 commands that can be used to configure what shared-library to load and
3073 to control aspects of the way it is handled. Also, made an update to
3074 the README.ENGINE file that brings its information up-to-date and
3075 provides some information and instructions on the "dynamic" ENGINE
3076 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
3079 *) Make it possible to unload ranges of ERR strings with a new
3080 "ERR_unload_strings" function.
3083 *) Add a copy() function to EVP_MD.
3086 *) Make EVP_MD routines take a context pointer instead of just the
3087 md_data void pointer.
3090 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
3091 that the digest can only process a single chunk of data
3092 (typically because it is provided by a piece of
3093 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
3094 is only going to provide a single chunk of data, and hence the
3095 framework needn't accumulate the data for oneshot drivers.
3098 *) As with "ERR", make it possible to replace the underlying "ex_data"
3099 functions. This change also alters the storage and management of global
3100 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
3101 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
3102 index counters. The API functions that use this state have been changed
3103 to take a "class_index" rather than pointers to the class's local STACK
3104 and counter, and there is now an API function to dynamically create new
3105 classes. This centralisation allows us to (a) plug a lot of the
3106 thread-safety problems that existed, and (b) makes it possible to clean
3107 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
3108 such data would previously have always leaked in application code and
3109 workarounds were in place to make the memory debugging turn a blind eye
3110 to it. Application code that doesn't use this new function will still
3111 leak as before, but their memory debugging output will announce it now
3112 rather than letting it slide.
3114 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
3115 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
3116 has a return value to indicate success or failure.
3119 *) Make it possible to replace the underlying "ERR" functions such that the
3120 global state (2 LHASH tables and 2 locks) is only used by the "default"
3121 implementation. This change also adds two functions to "get" and "set"
3122 the implementation prior to it being automatically set the first time
3123 any other ERR function takes place. Ie. an application can call "get",
3124 pass the return value to a module it has just loaded, and that module
3125 can call its own "set" function using that value. This means the
3126 module's "ERR" operations will use (and modify) the error state in the
3127 application and not in its own statically linked copy of OpenSSL code.
3130 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
3131 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
3132 the operation, and provides a more encapsulated way for external code
3133 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
3134 to use these functions rather than manually incrementing the counts.
3136 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
3139 *) Add EVP test program.
3142 *) Add symmetric cipher support to ENGINE. Expect the API to change!
3145 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
3146 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
3147 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
3148 These allow a CRL to be built without having to access X509_CRL fields
3149 directly. Modify 'ca' application to use new functions.
3152 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
3153 bug workarounds. Rollback attack detection is a security feature.
3154 The problem will only arise on OpenSSL servers when TLSv1 is not
3155 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
3156 Software authors not wanting to support TLSv1 will have special reasons
3157 for their choice and can explicitly enable this option.
3158 [Bodo Moeller, Lutz Jaenicke]
3160 *) Rationalise EVP so it can be extended: don't include a union of
3161 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
3162 (similar to those existing for EVP_CIPHER_CTX).
3167 EVP_MD_CTX_init(&md); /* new function call */
3168 EVP_DigestInit(&md, EVP_sha1());
3169 EVP_DigestUpdate(&md, in, len);
3170 EVP_DigestFinal(&md, out, NULL);
3171 EVP_MD_CTX_cleanup(&md); /* new function call */
3175 *) Make DES key schedule conform to the usual scheme, as well as
3176 correcting its structure. This means that calls to DES functions
3177 now have to pass a pointer to a des_key_schedule instead of a
3178 plain des_key_schedule (which was actually always a pointer
3181 des_key_schedule ks;
3183 des_set_key_checked(..., &ks);
3184 des_ncbc_encrypt(..., &ks, ...);
3186 (Note that a later change renames 'des_...' into 'DES_...'.)
3189 *) Initial reduction of linker bloat: the use of some functions, such as
3190 PEM causes large amounts of unused functions to be linked in due to
3191 poor organisation. For example pem_all.c contains every PEM function
3192 which has a knock on effect of linking in large amounts of (unused)
3193 ASN1 code. Grouping together similar functions and splitting unrelated
3194 functions prevents this.
3197 *) Cleanup of EVP macros.
3200 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
3201 correct _ecb suffix.
3204 *) Add initial OCSP responder support to ocsp application. The
3205 revocation information is handled using the text based index
3206 use by the ca application. The responder can either handle
3207 requests generated internally, supplied in files (for example
3208 via a CGI script) or using an internal minimal server.
3211 *) Add configuration choices to get zlib compression for TLS.
3214 *) Changes to Kerberos SSL for RFC 2712 compliance:
3215 1. Implemented real KerberosWrapper, instead of just using
3216 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
3217 2. Implemented optional authenticator field of KerberosWrapper.
3219 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
3220 and authenticator structs; see crypto/krb5/.
3222 Generalized Kerberos calls to support multiple Kerberos libraries.
3223 [Vern Staats <staatsvr@asc.hpc.mil>,
3224 Jeffrey Altman <jaltman@columbia.edu>
3225 via Richard Levitte]
3227 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
3228 already does with RSA. testdsa.h now has 'priv_key/pub_key'
3229 values for each of the key sizes rather than having just
3230 parameters (and 'speed' generating keys each time).
3233 *) Speed up EVP routines.
3236 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
3237 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
3238 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
3239 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
3241 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
3242 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
3243 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
3246 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
3248 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
3251 *) Added the OS2-EMX target.
3252 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
3254 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
3255 to support NCONF routines in extension code. New function CONF_set_nconf()
3256 to allow functions which take an NCONF to also handle the old LHASH
3257 structure: this means that the old CONF compatible routines can be
3258 retained (in particular wrt extensions) without having to duplicate the
3259 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
3262 *) Enhance the general user interface with mechanisms for inner control
3263 and with possibilities to have yes/no kind of prompts.
3266 *) Change all calls to low level digest routines in the library and
3267 applications to use EVP. Add missing calls to HMAC_cleanup() and
3268 don't assume HMAC_CTX can be copied using memcpy().
3269 [Verdon Walker <VWalker@novell.com>, Steve Henson]
3271 *) Add the possibility to control engines through control names but with
3272 arbitrary arguments instead of just a string.
3273 Change the key loaders to take a UI_METHOD instead of a callback
3274 function pointer. NOTE: this breaks binary compatibility with earlier
3275 versions of OpenSSL [engine].
3276 Adapt the nCipher code for these new conditions and add a card insertion
3280 *) Enhance the general user interface with mechanisms to better support
3281 dialog box interfaces, application-defined prompts, the possibility
3282 to use defaults (for example default passwords from somewhere else)
3283 and interrupts/cancellations.
3286 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
3287 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
3290 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
3291 tidy up some unnecessarily weird code in 'sk_new()').
3292 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
3294 *) Change the key loading routines for ENGINEs to use the same kind
3295 callback (pem_password_cb) as all other routines that need this
3299 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
3300 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
3301 than this minimum value is recommended.
3304 *) New random seeder for OpenVMS, using the system process statistics
3305 that are easily reachable.
3308 *) Windows apparently can't transparently handle global
3309 variables defined in DLLs. Initialisations such as:
3311 const ASN1_ITEM *it = &ASN1_INTEGER_it;
3313 wont compile. This is used by the any applications that need to
3314 declare their own ASN1 modules. This was fixed by adding the option
3315 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
3316 needed for static libraries under Win32.
3319 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
3320 setting of purpose and trust fields. New X509_STORE trust and
3321 purpose functions and tidy up setting in other SSL functions.
3324 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
3325 structure. These are inherited by X509_STORE_CTX when it is
3326 initialised. This allows various defaults to be set in the
3327 X509_STORE structure (such as flags for CRL checking and custom
3328 purpose or trust settings) for functions which only use X509_STORE_CTX
3329 internally such as S/MIME.
3331 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
3332 trust settings if they are not set in X509_STORE. This allows X509_STORE
3333 purposes and trust (in S/MIME for example) to override any set by default.
3335 Add command line options for CRL checking to smime, s_client and s_server
3339 *) Initial CRL based revocation checking. If the CRL checking flag(s)
3340 are set then the CRL is looked up in the X509_STORE structure and
3341 its validity and signature checked, then if the certificate is found
3342 in the CRL the verify fails with a revoked error.
3344 Various new CRL related callbacks added to X509_STORE_CTX structure.
3346 Command line options added to 'verify' application to support this.
3348 This needs some additional work, such as being able to handle multiple
3349 CRLs with different times, extension based lookup (rather than just
3350 by subject name) and ultimately more complete V2 CRL extension
3354 *) Add a general user interface API (crypto/ui/). This is designed
3355 to replace things like des_read_password and friends (backward
3356 compatibility functions using this new API are provided).
3357 The purpose is to remove prompting functions from the DES code
3358 section as well as provide for prompting through dialog boxes in
3359 a window system and the like.
3362 *) Add "ex_data" support to ENGINE so implementations can add state at a
3363 per-structure level rather than having to store it globally.
3366 *) Make it possible for ENGINE structures to be copied when retrieved by
3367 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
3368 This causes the "original" ENGINE structure to act like a template,
3369 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
3370 operational state can be localised to each ENGINE structure, despite the
3371 fact they all share the same "methods". New ENGINE structures returned in
3372 this case have no functional references and the return value is the single
3373 structural reference. This matches the single structural reference returned
3374 by ENGINE_by_id() normally, when it is incremented on the pre-existing
3378 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
3379 needs to match any other type at all we need to manually clear the
3383 *) Changes to the "openssl engine" utility to include;
3384 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
3385 about an ENGINE's available control commands.
3386 - executing control commands from command line arguments using the
3387 '-pre' and '-post' switches. '-post' is only used if '-t' is
3388 specified and the ENGINE is successfully initialised. The syntax for
3389 the individual commands are colon-separated, for example;
3390 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
3393 *) New dynamic control command support for ENGINEs. ENGINEs can now
3394 declare their own commands (numbers), names (strings), descriptions,
3395 and input types for run-time discovery by calling applications. A
3396 subset of these commands are implicitly classed as "executable"
3397 depending on their input type, and only these can be invoked through
3398 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
3399 can be based on user input, config files, etc). The distinction is
3400 that "executable" commands cannot return anything other than a boolean
3401 result and can only support numeric or string input, whereas some
3402 discoverable commands may only be for direct use through
3403 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
3404 pointers, or other custom uses. The "executable" commands are to
3405 support parameterisations of ENGINE behaviour that can be
3406 unambiguously defined by ENGINEs and used consistently across any
3407 OpenSSL-based application. Commands have been added to all the
3408 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
3409 control over shared-library paths without source code alterations.
3412 *) Changed all ENGINE implementations to dynamically allocate their
3413 ENGINEs rather than declaring them statically. Apart from this being
3414 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
3415 this also allows the implementations to compile without using the
3416 internal engine_int.h header.
3419 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
3420 'const' value. Any code that should be able to modify a RAND_METHOD
3421 should already have non-const pointers to it (ie. they should only
3422 modify their own ones).
3425 *) Made a variety of little tweaks to the ENGINE code.
3426 - "atalla" and "ubsec" string definitions were moved from header files
3427 to C code. "nuron" string definitions were placed in variables
3428 rather than hard-coded - allowing parameterisation of these values
3429 later on via ctrl() commands.
3430 - Removed unused "#if 0"'d code.
3431 - Fixed engine list iteration code so it uses ENGINE_free() to release
3432 structural references.
3433 - Constified the RAND_METHOD element of ENGINE structures.
3434 - Constified various get/set functions as appropriate and added
3435 missing functions (including a catch-all ENGINE_cpy that duplicates
3436 all ENGINE values onto a new ENGINE except reference counts/state).
3437 - Removed NULL parameter checks in get/set functions. Setting a method
3438 or function to NULL is a way of cancelling out a previously set
3439 value. Passing a NULL ENGINE parameter is just plain stupid anyway
3440 and doesn't justify the extra error symbols and code.
3441 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
3442 flags from engine_int.h to engine.h.
3443 - Changed prototypes for ENGINE handler functions (init(), finish(),
3444 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
3447 *) Implement binary inversion algorithm for BN_mod_inverse in addition
3448 to the algorithm using long division. The binary algorithm can be
3449 used only if the modulus is odd. On 32-bit systems, it is faster
3450 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
3451 roughly 5-15% for 256-bit moduli), so we use it only for moduli
3452 up to 450 bits. In 64-bit environments, the binary algorithm
3453 appears to be advantageous for much longer moduli; here we use it
3454 for moduli up to 2048 bits.
3457 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
3458 could not support the combine flag in choice fields.
3461 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
3462 extensions from a certificate request to the certificate.
3465 *) Allow multiple 'certopt' and 'nameopt' options to be separated
3466 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
3467 file: this allows the display of the certificate about to be
3468 signed to be customised, to allow certain fields to be included
3469 or excluded and extension details. The old system didn't display
3470 multicharacter strings properly, omitted fields not in the policy
3471 and couldn't display additional details such as extensions.
3474 *) Function EC_POINTs_mul for multiple scalar multiplication
3475 of an arbitrary number of elliptic curve points
3476 \sum scalars[i]*points[i],
3477 optionally including the generator defined for the EC_GROUP:
3478 scalar*generator + \sum scalars[i]*points[i].
3480 EC_POINT_mul is a simple wrapper function for the typical case
3481 that the point list has just one item (besides the optional
3485 *) First EC_METHODs for curves over GF(p):
3487 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
3488 operations and provides various method functions that can also
3489 operate with faster implementations of modular arithmetic.
3491 EC_GFp_mont_method() reuses most functions that are part of
3492 EC_GFp_simple_method, but uses Montgomery arithmetic.
3494 [Bodo Moeller; point addition and point doubling
3495 implementation directly derived from source code provided by
3496 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
3498 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
3499 crypto/ec/ec_lib.c):
3501 Curves are EC_GROUP objects (with an optional group generator)
3502 based on EC_METHODs that are built into the library.
3504 Points are EC_POINT objects based on EC_GROUP objects.
3506 Most of the framework would be able to handle curves over arbitrary
3507 finite fields, but as there are no obvious types for fields other
3508 than GF(p), some functions are limited to that for now.
3511 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
3512 that the file contains a complete HTTP response.
3515 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
3516 change the def and num file printf format specifier from "%-40sXXX"
3517 to "%-39s XXX". The latter will always guarantee a space after the
3518 field while the former will cause them to run together if the field
3519 is 40 of more characters long.
3522 *) Constify the cipher and digest 'method' functions and structures
3523 and modify related functions to take constant EVP_MD and EVP_CIPHER
3527 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
3528 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
3531 *) Modify EVP_Digest*() routines so they now return values. Although the
3532 internal software routines can never fail additional hardware versions
3536 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
3538 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
3539 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
3542 ERR_R_NESTED_ASN1_ERROR
3544 ERR_R_MISSING_ASN1_EOS
3545 were 4 .. 9, conflicting with
3546 ERR_LIB_RSA (= ERR_R_RSA_LIB)
3548 ERR_LIB_PEM (= ERR_R_PEM_LIB).
3549 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
3551 Add new error code 'ERR_R_INTERNAL_ERROR'.
3554 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
3558 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
3559 sets the subject name for a new request or supersedes the
3560 subject name in a given request. Formats that can be parsed are
3561 'CN=Some Name, OU=myOU, C=IT'
3563 'CN=Some Name/OU=myOU/C=IT'.
3565 Add options '-batch' and '-verbose' to 'openssl req'.
3566 [Massimiliano Pala <madwolf@hackmasters.net>]
3568 *) Introduce the possibility to access global variables through
3569 functions on platform were that's the best way to handle exporting
3570 global variables in shared libraries. To enable this functionality,
3571 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
3572 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
3573 is normally done by Configure or something similar).
3575 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
3576 in the source file (foo.c) like this:
3578 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
3579 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
3581 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
3582 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
3584 OPENSSL_DECLARE_GLOBAL(int,foo);
3585 #define foo OPENSSL_GLOBAL_REF(foo)
3586 OPENSSL_DECLARE_GLOBAL(double,bar);
3587 #define bar OPENSSL_GLOBAL_REF(bar)
3589 The #defines are very important, and therefore so is including the
3590 header file everywhere where the defined globals are used.
3592 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
3593 of ASN.1 items, but that structure is a bit different.
3595 The largest change is in util/mkdef.pl which has been enhanced with
3596 better and easier to understand logic to choose which symbols should
3597 go into the Windows .def files as well as a number of fixes and code
3598 cleanup (among others, algorithm keywords are now sorted
3599 lexicographically to avoid constant rewrites).
3602 *) In BN_div() keep a copy of the sign of 'num' before writing the
3603 result to 'rm' because if rm==num the value will be overwritten
3604 and produce the wrong result if 'num' is negative: this caused
3605 problems with BN_mod() and BN_nnmod().
3608 *) Function OCSP_request_verify(). This checks the signature on an
3609 OCSP request and verifies the signer certificate. The signer
3610 certificate is just checked for a generic purpose and OCSP request
3614 *) Add OCSP_check_validity() function to check the validity of OCSP
3615 responses. OCSP responses are prepared in real time and may only
3616 be a few seconds old. Simply checking that the current time lies
3617 between thisUpdate and nextUpdate max reject otherwise valid responses
3618 caused by either OCSP responder or client clock inaccuracy. Instead
3619 we allow thisUpdate and nextUpdate to fall within a certain period of
3620 the current time. The age of the response can also optionally be
3621 checked. Two new options -validity_period and -status_age added to
3625 *) If signature or public key algorithm is unrecognized print out its
3626 OID rather that just UNKNOWN.
3629 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
3630 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
3631 ID to be generated from the issuer certificate alone which can then be
3632 passed to OCSP_id_issuer_cmp().
3635 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
3636 ASN1 modules to export functions returning ASN1_ITEM pointers
3637 instead of the ASN1_ITEM structures themselves. This adds several
3638 new macros which allow the underlying ASN1 function/structure to
3639 be accessed transparently. As a result code should not use ASN1_ITEM
3640 references directly (such as &X509_it) but instead use the relevant
3641 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
3642 use of the new ASN1 code on platforms where exporting structures
3643 is problematical (for example in shared libraries) but exporting
3644 functions returning pointers to structures is not.
3647 *) Add support for overriding the generation of SSL/TLS session IDs.
3648 These callbacks can be registered either in an SSL_CTX or per SSL.
3649 The purpose of this is to allow applications to control, if they wish,
3650 the arbitrary values chosen for use as session IDs, particularly as it
3651 can be useful for session caching in multiple-server environments. A
3652 command-line switch for testing this (and any client code that wishes
3653 to use such a feature) has been added to "s_server".
3654 [Geoff Thorpe, Lutz Jaenicke]
3656 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
3657 of the form '#if defined(...) || defined(...) || ...' and
3658 '#if !defined(...) && !defined(...) && ...'. This also avoids
3659 the growing number of special cases it was previously handling.
3662 *) Make all configuration macros available for application by making
3663 sure they are available in opensslconf.h, by giving them names starting
3664 with "OPENSSL_" to avoid conflicts with other packages and by making
3665 sure e_os2.h will cover all platform-specific cases together with
3667 Additionally, it is now possible to define configuration/platform-
3668 specific names (called "system identities"). In the C code, these
3669 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
3670 macro with the name beginning with "OPENSSL_SYS_", which is determined
3671 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
3675 *) New option -set_serial to 'req' and 'x509' this allows the serial
3676 number to use to be specified on the command line. Previously self
3677 signed certificates were hard coded with serial number 0 and the
3678 CA options of 'x509' had to use a serial number in a file which was
3682 *) New options to 'ca' utility to support V2 CRL entry extensions.
3683 Currently CRL reason, invalidity date and hold instruction are
3684 supported. Add new CRL extensions to V3 code and some new objects.
3687 *) New function EVP_CIPHER_CTX_set_padding() this is used to
3688 disable standard block padding (aka PKCS#5 padding) in the EVP
3689 API, which was previously mandatory. This means that the data is
3690 not padded in any way and so the total length much be a multiple
3691 of the block size, otherwise an error occurs.
3694 *) Initial (incomplete) OCSP SSL support.
3697 *) New function OCSP_parse_url(). This splits up a URL into its host,
3698 port and path components: primarily to parse OCSP URLs. New -url
3699 option to ocsp utility.
3702 *) New nonce behavior. The return value of OCSP_check_nonce() now
3703 reflects the various checks performed. Applications can decide
3704 whether to tolerate certain situations such as an absent nonce
3705 in a response when one was present in a request: the ocsp application
3706 just prints out a warning. New function OCSP_add1_basic_nonce()
3707 this is to allow responders to include a nonce in a response even if
3708 the request is nonce-less.
3711 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
3712 skipped when using openssl x509 multiple times on a single input file,
3713 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
3716 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
3717 set string type: to handle setting ASN1_TIME structures. Fix ca
3718 utility to correctly initialize revocation date of CRLs.
3721 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
3722 the clients preferred ciphersuites and rather use its own preferences.
3723 Should help to work around M$ SGC (Server Gated Cryptography) bug in
3724 Internet Explorer by ensuring unchanged hash method during stepup.
3725 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
3728 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
3729 to aes and add a new 'exist' option to print out symbols that don't
3733 *) Additional options to ocsp utility to allow flags to be set and
3734 additional certificates supplied.
3737 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
3738 OCSP client a number of certificate to only verify the response
3742 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
3743 handle the new API. Currently only ECB, CBC modes supported. Add new
3746 Add TLS AES ciphersuites as described in RFC3268, "Advanced
3747 Encryption Standard (AES) Ciphersuites for Transport Layer
3748 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
3749 not enabled by default and were not part of the "ALL" ciphersuite
3750 alias because they were not yet official; they could be
3751 explicitly requested by specifying the "AESdraft" ciphersuite
3752 group alias. In the final release of OpenSSL 0.9.7, the group
3753 alias is called "AES" and is part of "ALL".)
3754 [Ben Laurie, Steve Henson, Bodo Moeller]
3756 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
3757 request to response.
3760 *) Functions for OCSP responders. OCSP_request_onereq_count(),
3761 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
3762 extract information from a certificate request. OCSP_response_create()
3763 creates a response and optionally adds a basic response structure.
3764 OCSP_basic_add1_status() adds a complete single response to a basic
3765 response and returns the OCSP_SINGLERESP structure just added (to allow
3766 extensions to be included for example). OCSP_basic_add1_cert() adds a
3767 certificate to a basic response and OCSP_basic_sign() signs a basic
3768 response with various flags. New helper functions ASN1_TIME_check()
3769 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
3770 (converts ASN1_TIME to GeneralizedTime).
3773 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
3774 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
3775 structure from a certificate. X509_pubkey_digest() digests the public_key
3776 contents: this is used in various key identifiers.
3779 *) Make sk_sort() tolerate a NULL argument.
3780 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
3782 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
3783 passed by the function are trusted implicitly. If any of them signed the
3784 response then it is assumed to be valid and is not verified.
3787 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
3788 to data. This was previously part of the PKCS7 ASN1 code. This
3789 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
3790 [Steve Henson, reported by Kenneth R. Robinette
3791 <support@securenetterm.com>]
3793 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
3794 routines: without these tracing memory leaks is very painful.
3795 Fix leaks in PKCS12 and PKCS7 routines.
3798 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
3799 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
3800 effectively meant GeneralizedTime would never be used. Now it
3801 is initialised to -1 but X509_time_adj() now has to check the value
3802 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
3803 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
3804 [Steve Henson, reported by Kenneth R. Robinette
3805 <support@securenetterm.com>]
3807 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
3808 result in a zero length in the ASN1_INTEGER structure which was
3809 not consistent with the structure when d2i_ASN1_INTEGER() was used
3810 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
3811 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
3812 where it did not print out a minus for negative ASN1_INTEGER.
3815 *) Add summary printout to ocsp utility. The various functions which
3816 convert status values to strings have been renamed to:
3817 OCSP_response_status_str(), OCSP_cert_status_str() and
3818 OCSP_crl_reason_str() and are no longer static. New options
3819 to verify nonce values and to disable verification. OCSP response
3820 printout format cleaned up.
3823 *) Add additional OCSP certificate checks. These are those specified
3824 in RFC2560. This consists of two separate checks: the CA of the
3825 certificate being checked must either be the OCSP signer certificate
3826 or the issuer of the OCSP signer certificate. In the latter case the
3827 OCSP signer certificate must contain the OCSP signing extended key
3828 usage. This check is performed by attempting to match the OCSP
3829 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
3830 in the OCSP_CERTID structures of the response.
3833 *) Initial OCSP certificate verification added to OCSP_basic_verify()
3834 and related routines. This uses the standard OpenSSL certificate
3835 verify routines to perform initial checks (just CA validity) and
3836 to obtain the certificate chain. Then additional checks will be
3837 performed on the chain. Currently the root CA is checked to see
3838 if it is explicitly trusted for OCSP signing. This is used to set
3839 a root CA as a global signing root: that is any certificate that
3840 chains to that CA is an acceptable OCSP signing certificate.
3843 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
3844 extensions from a separate configuration file.
3845 As when reading extensions from the main configuration file,
3846 the '-extensions ...' option may be used for specifying the
3848 [Massimiliano Pala <madwolf@comune.modena.it>]
3850 *) New OCSP utility. Allows OCSP requests to be generated or
3851 read. The request can be sent to a responder and the output
3852 parsed, outputed or printed in text form. Not complete yet:
3853 still needs to check the OCSP response validity.
3856 *) New subcommands for 'openssl ca':
3857 'openssl ca -status <serial>' prints the status of the cert with
3858 the given serial number (according to the index file).
3859 'openssl ca -updatedb' updates the expiry status of certificates
3861 [Massimiliano Pala <madwolf@comune.modena.it>]
3863 *) New '-newreq-nodes' command option to CA.pl. This is like
3864 '-newreq', but calls 'openssl req' with the '-nodes' option
3865 so that the resulting key is not encrypted.
3866 [Damien Miller <djm@mindrot.org>]
3868 *) New configuration for the GNU Hurd.
3869 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
3871 *) Initial code to implement OCSP basic response verify. This
3872 is currently incomplete. Currently just finds the signer's
3873 certificate and verifies the signature on the response.
3876 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
3877 value of OPENSSLDIR. This is available via the new '-d' option
3878 to 'openssl version', and is also included in 'openssl version -a'.
3881 *) Allowing defining memory allocation callbacks that will be given
3882 file name and line number information in additional arguments
3883 (a const char* and an int). The basic functionality remains, as
3884 well as the original possibility to just replace malloc(),
3885 realloc() and free() by functions that do not know about these
3886 additional arguments. To register and find out the current
3887 settings for extended allocation functions, the following
3888 functions are provided:
3890 CRYPTO_set_mem_ex_functions
3891 CRYPTO_set_locked_mem_ex_functions
3892 CRYPTO_get_mem_ex_functions
3893 CRYPTO_get_locked_mem_ex_functions
3895 These work the same way as CRYPTO_set_mem_functions and friends.
3896 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
3897 extended allocation function is enabled.
3898 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
3899 a conventional allocation function is enabled.
3900 [Richard Levitte, Bodo Moeller]
3902 *) Finish off removing the remaining LHASH function pointer casts.
3903 There should no longer be any prototype-casting required when using
3904 the LHASH abstraction, and any casts that remain are "bugs". See
3905 the callback types and macros at the head of lhash.h for details
3906 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
3909 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
3910 If /dev/[u]random devices are not available or do not return enough
3911 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
3913 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
3914 /etc/entropy will be queried once each in this sequence, quering stops
3915 when enough entropy was collected without querying more sockets.
3918 *) Change the Unix RAND_poll() variant to be able to poll several
3919 random devices, as specified by DEVRANDOM, until a sufficient amount
3920 of data has been collected. We spend at most 10 ms on each file
3921 (select timeout) and read in non-blocking mode. DEVRANDOM now
3922 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
3923 (previously it was just the string "/dev/urandom"), so on typical
3924 platforms the 10 ms delay will never occur.
3925 Also separate out the Unix variant to its own file, rand_unix.c.
3926 For VMS, there's a currently-empty rand_vms.c.
3929 *) Move OCSP client related routines to ocsp_cl.c. These
3930 provide utility functions which an application needing
3931 to issue a request to an OCSP responder and analyse the
3932 response will typically need: as opposed to those which an
3933 OCSP responder itself would need which will be added later.
3935 OCSP_request_sign() signs an OCSP request with an API similar
3936 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
3937 response. OCSP_response_get1_basic() extracts basic response
3938 from response. OCSP_resp_find_status(): finds and extracts status
3939 information from an OCSP_CERTID structure (which will be created
3940 when the request structure is built). These are built from lower
3941 level functions which work on OCSP_SINGLERESP structures but
3942 wont normally be used unless the application wishes to examine
3943 extensions in the OCSP response for example.
3945 Replace nonce routines with a pair of functions.
3946 OCSP_request_add1_nonce() adds a nonce value and optionally
3947 generates a random value. OCSP_check_nonce() checks the
3948 validity of the nonce in an OCSP response.
3951 *) Change function OCSP_request_add() to OCSP_request_add0_id().
3952 This doesn't copy the supplied OCSP_CERTID and avoids the
3953 need to free up the newly created id. Change return type
3954 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
3955 This can then be used to add extensions to the request.
3956 Deleted OCSP_request_new(), since most of its functionality
3957 is now in OCSP_REQUEST_new() (and the case insensitive name
3958 clash) apart from the ability to set the request name which
3959 will be added elsewhere.
3962 *) Update OCSP API. Remove obsolete extensions argument from
3963 various functions. Extensions are now handled using the new
3964 OCSP extension code. New simple OCSP HTTP function which
3965 can be used to send requests and parse the response.
3968 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
3969 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
3970 uses the special reorder version of SET OF to sort the attributes
3971 and reorder them to match the encoded order. This resolves a long
3972 standing problem: a verify on a PKCS7 structure just after signing
3973 it used to fail because the attribute order did not match the
3974 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
3975 it uses the received order. This is necessary to tolerate some broken
3976 software that does not order SET OF. This is handled by encoding
3977 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
3978 to produce the required SET OF.
3981 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
3982 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
3983 files to get correct declarations of the ASN.1 item variables.
3986 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
3987 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
3988 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
3989 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
3990 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
3991 ASN1_ITEM and no wrapper functions.
3994 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
3995 replace the old function pointer based I/O routines. Change most of
3996 the *_d2i_bio() and *_d2i_fp() functions to use these.
3999 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
4000 lines, recognice more "algorithms" that can be deselected, and make
4001 it complain about algorithm deselection that isn't recognised.
4004 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
4005 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
4006 to use new functions. Add NO_ASN1_OLD which can be set to remove
4007 some old style ASN1 functions: this can be used to determine if old
4008 code will still work when these eventually go away.
4011 *) New extension functions for OCSP structures, these follow the
4012 same conventions as certificates and CRLs.
4015 *) New function X509V3_add1_i2d(). This automatically encodes and
4016 adds an extension. Its behaviour can be customised with various
4017 flags to append, replace or delete. Various wrappers added for
4018 certifcates and CRLs.
4021 *) Fix to avoid calling the underlying ASN1 print routine when
4022 an extension cannot be parsed. Correct a typo in the
4023 OCSP_SERVICELOC extension. Tidy up print OCSP format.
4026 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
4027 entries for variables.
4030 *) Add functionality to apps/openssl.c for detecting locking
4031 problems: As the program is single-threaded, all we have
4032 to do is register a locking callback using an array for
4033 storing which locks are currently held by the program.
4036 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
4037 SSL_get_ex_data_X509_STORE_idx(), which is used in
4038 ssl_verify_cert_chain() and thus can be called at any time
4039 during TLS/SSL handshakes so that thread-safety is essential.
4040 Unfortunately, the ex_data design is not at all suited
4041 for multi-threaded use, so it probably should be abolished.
4044 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
4045 [Broadcom, tweaked and integrated by Geoff Thorpe]
4047 *) Move common extension printing code to new function
4048 X509V3_print_extensions(). Reorganise OCSP print routines and
4049 implement some needed OCSP ASN1 functions. Add OCSP extensions.
4052 *) New function X509_signature_print() to remove duplication in some
4056 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
4057 set (this was treated exactly the same as SET OF previously). This
4058 is used to reorder the STACK representing the structure to match the
4059 encoding. This will be used to get round a problem where a PKCS7
4060 structure which was signed could not be verified because the STACK
4061 order did not reflect the encoded order.
4064 *) Reimplement the OCSP ASN1 module using the new code.
4067 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
4068 for its ASN1 operations. The old style function pointers still exist
4069 for now but they will eventually go away.
4072 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
4073 completely replaces the old ASN1 functionality with a table driven
4074 encoder and decoder which interprets an ASN1_ITEM structure describing
4075 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
4076 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
4077 has also been converted to the new form.
4080 *) Change BN_mod_exp_recp so that negative moduli are tolerated
4081 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
4082 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
4083 for negative moduli.
4086 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
4087 of not touching the result's sign bit.
4090 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
4094 *) Changed the LHASH code to use prototypes for callbacks, and created
4095 macros to declare and implement thin (optionally static) functions
4096 that provide type-safety and avoid function pointer casting for the
4097 type-specific callbacks.
4100 *) Added Kerberos Cipher Suites to be used with TLS, as written in
4102 [Veers Staats <staatsvr@asc.hpc.mil>,
4103 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
4105 *) Reformat the FAQ so the different questions and answers can be divided
4106 in sections depending on the subject.
4109 *) Have the zlib compression code load ZLIB.DLL dynamically under
4113 *) New function BN_mod_sqrt for computing square roots modulo a prime
4114 (using the probabilistic Tonelli-Shanks algorithm unless
4115 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
4116 be handled deterministically).
4117 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
4119 *) Make BN_mod_inverse faster by explicitly handling small quotients
4120 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
4121 512 bits], about 30% for larger ones [1024 or 2048 bits].)
4124 *) New function BN_kronecker.
4127 *) Fix BN_gcd so that it works on negative inputs; the result is
4128 positive unless both parameters are zero.
4129 Previously something reasonably close to an infinite loop was
4130 possible because numbers could be growing instead of shrinking
4131 in the implementation of Euclid's algorithm.
4134 *) Fix BN_is_word() and BN_is_one() macros to take into account the
4135 sign of the number in question.
4137 Fix BN_is_word(a,w) to work correctly for w == 0.
4139 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
4140 because its test if the absolute value of 'a' equals 'w'.
4141 Note that BN_abs_is_word does *not* handle w == 0 reliably;
4142 it exists mostly for use in the implementations of BN_is_zero(),
4143 BN_is_one(), and BN_is_word().
4146 *) New function BN_swap.
4149 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
4150 the exponentiation functions are more likely to produce reasonable
4151 results on negative inputs.
4154 *) Change BN_mod_mul so that the result is always non-negative.
4155 Previously, it could be negative if one of the factors was negative;
4156 I don't think anyone really wanted that behaviour.
4159 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
4160 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
4161 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
4162 and add new functions:
4171 BN_mod_lshift1_quick
4175 These functions always generate non-negative results.
4177 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
4178 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
4180 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
4181 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
4182 be reduced modulo m.
4183 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
4186 The following entry accidentily appeared in the CHANGES file
4187 distributed with OpenSSL 0.9.7. The modifications described in
4188 it do *not* apply to OpenSSL 0.9.7.
4190 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4191 was actually never needed) and in BN_mul(). The removal in BN_mul()
4192 required a small change in bn_mul_part_recursive() and the addition
4193 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4194 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4195 bn_sub_words() and bn_add_words() except they take arrays with
4200 *) In 'openssl passwd', verify passwords read from the terminal
4201 unless the '-salt' option is used (which usually means that
4202 verification would just waste user's time since the resulting
4203 hash is going to be compared with some given password hash)
4204 or the new '-noverify' option is used.
4206 This is an incompatible change, but it does not affect
4207 non-interactive use of 'openssl passwd' (passwords on the command
4208 line, '-stdin' option, '-in ...' option) and thus should not
4212 *) Remove all references to RSAref, since there's no more need for it.
4215 *) Make DSO load along a path given through an environment variable
4216 (SHLIB_PATH) with shl_load().
4219 *) Constify the ENGINE code as a result of BIGNUM constification.
4220 Also constify the RSA code and most things related to it. In a
4221 few places, most notable in the depth of the ASN.1 code, ugly
4222 casts back to non-const were required (to be solved at a later
4226 *) Make it so the openssl application has all engines loaded by default.
4229 *) Constify the BIGNUM routines a little more.
4232 *) Add the following functions:
4234 ENGINE_load_cswift()
4236 ENGINE_load_atalla()
4238 ENGINE_load_builtin_engines()
4240 That way, an application can itself choose if external engines that
4241 are built-in in OpenSSL shall ever be used or not. The benefit is
4242 that applications won't have to be linked with libdl or other dso
4243 libraries unless it's really needed.
4245 Changed 'openssl engine' to load all engines on demand.
4246 Changed the engine header files to avoid the duplication of some
4247 declarations (they differed!).
4250 *) 'openssl engine' can now list capabilities.
4253 *) Better error reporting in 'openssl engine'.
4256 *) Never call load_dh_param(NULL) in s_server.
4259 *) Add engine application. It can currently list engines by name and
4260 identity, and test if they are actually available.
4263 *) Improve RPM specification file by forcing symbolic linking and making
4264 sure the installed documentation is also owned by root.root.
4265 [Damien Miller <djm@mindrot.org>]
4267 *) Give the OpenSSL applications more possibilities to make use of
4268 keys (public as well as private) handled by engines.
4271 *) Add OCSP code that comes from CertCo.
4274 *) Add VMS support for the Rijndael code.
4277 *) Added untested support for Nuron crypto accelerator.
4280 *) Add support for external cryptographic devices. This code was
4281 previously distributed separately as the "engine" branch.
4282 [Geoff Thorpe, Richard Levitte]
4284 *) Rework the filename-translation in the DSO code. It is now possible to
4285 have far greater control over how a "name" is turned into a filename
4286 depending on the operating environment and any oddities about the
4287 different shared library filenames on each system.
4290 *) Support threads on FreeBSD-elf in Configure.
4293 *) Fix for SHA1 assembly problem with MASM: it produces
4294 warnings about corrupt line number information when assembling
4295 with debugging information. This is caused by the overlapping
4297 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
4300 NCONF_get_number() has no error checking at all. As a replacement,
4301 NCONF_get_number_e() is defined (_e for "error checking") and is
4302 promoted strongly. The old NCONF_get_number is kept around for
4303 binary backward compatibility.
4304 Make it possible for methods to load from something other than a BIO,
4305 by providing a function pointer that is given a name instead of a BIO.
4306 For example, this could be used to load configuration data from an
4310 *) Fix for non blocking accept BIOs. Added new I/O special reason
4311 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
4312 with non blocking I/O was not possible because no retry code was
4313 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
4317 *) Added the beginnings of Rijndael support.
4320 *) Fix for bug in DirectoryString mask setting. Add support for
4321 X509_NAME_print_ex() in 'req' and X509_print_ex() function
4322 to allow certificate printing to more controllable, additional
4323 'certopt' option to 'x509' to allow new printing options to be
4327 *) Clean old EAY MD5 hack from e_os.h.
4330 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
4332 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4333 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4334 [Joe Orton, Steve Henson]
4336 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
4338 *) Fix additional bug revealed by the NISCC test suite:
4340 Stop bug triggering large recursion when presented with
4341 certain ASN.1 tags (CVE-2003-0851)
4344 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
4346 *) Fix various bugs revealed by running the NISCC test suite:
4348 Stop out of bounds reads in the ASN1 code when presented with
4349 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4351 If verify callback ignores invalid public key errors don't try to check
4352 certificate signature with the NULL public key.
4356 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4357 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4361 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4362 extra data after the compression methods not only for TLS 1.0
4363 but also for SSL 3.0 (as required by the specification).
4364 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4366 *) Change X509_certificate_type() to mark the key as exported/exportable
4367 when it's 512 *bits* long, not 512 bytes.
4370 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
4372 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4373 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4374 a protocol version number mismatch like a decryption error
4375 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4378 *) Turn on RSA blinding by default in the default implementation
4379 to avoid a timing attack. Applications that don't want it can call
4380 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4381 They would be ill-advised to do so in most cases.
4382 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4384 *) Change RSA blinding code so that it works when the PRNG is not
4385 seeded (in this case, the secret RSA exponent is abused as
4386 an unpredictable seed -- if it is not unpredictable, there
4387 is no point in blinding anyway). Make RSA blinding thread-safe
4388 by remembering the creator's thread ID in rsa->blinding and
4389 having all other threads use local one-time blinding factors
4390 (this requires more computation than sharing rsa->blinding, but
4391 avoids excessive locking; and if an RSA object is not shared
4392 between threads, blinding will still be very fast).
4395 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
4397 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4398 via timing by performing a MAC computation even if incorrrect
4399 block cipher padding has been found. This is a countermeasure
4400 against active attacks where the attacker has to distinguish
4401 between bad padding and a MAC verification error. (CVE-2003-0078)
4403 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4404 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4405 Martin Vuagnoux (EPFL, Ilion)]
4407 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
4409 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
4410 memory from it's contents. This is done with a counter that will
4411 place alternating values in each byte. This can be used to solve
4412 two issues: 1) the removal of calls to memset() by highly optimizing
4413 compilers, and 2) cleansing with other values than 0, since those can
4414 be read through on certain media, for example a swap space on disk.
4417 *) Bugfix: client side session caching did not work with external caching,
4418 because the session->cipher setting was not restored when reloading
4419 from the external cache. This problem was masked, when
4420 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
4421 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
4424 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
4425 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
4426 [Zeev Lieber <zeev-l@yahoo.com>]
4428 *) Undo an undocumented change introduced in 0.9.6e which caused
4429 repeated calls to OpenSSL_add_all_ciphers() and
4430 OpenSSL_add_all_digests() to be ignored, even after calling
4434 *) Change the default configuration reader to deal with last line not
4435 being properly terminated.
4438 *) Change X509_NAME_cmp() so it applies the special rules on handling
4439 DN values that are of type PrintableString, as well as RDNs of type
4440 emailAddress where the value has the type ia5String.
4441 [stefank@valicert.com via Richard Levitte]
4443 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
4444 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
4445 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
4446 the bitwise-OR of the two for use by the majority of applications
4447 wanting this behaviour, and update the docs. The documented
4448 behaviour and actual behaviour were inconsistent and had been
4449 changing anyway, so this is more a bug-fix than a behavioural
4451 [Geoff Thorpe, diagnosed by Nadav Har'El]
4453 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
4454 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
4457 *) Fix initialization code race conditions in
4458 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
4459 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
4460 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
4461 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
4462 ssl2_get_cipher_by_char(),
4463 ssl3_get_cipher_by_char().
4464 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
4466 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
4467 the cached sessions are flushed, as the remove_cb() might use ex_data
4468 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
4469 (see [openssl.org #212]).
4470 [Geoff Thorpe, Lutz Jaenicke]
4472 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
4473 length, instead of the encoding length to d2i_ASN1_OBJECT.
4476 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
4478 *) [In 0.9.6g-engine release:]
4479 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
4480 [Lynn Gazis <lgazis@rainbow.com>]
4482 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
4484 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
4485 and get fix the header length calculation.
4486 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
4487 Alon Kantor <alonk@checkpoint.com> (and others),
4490 *) Use proper error handling instead of 'assertions' in buffer
4491 overflow checks added in 0.9.6e. This prevents DoS (the
4492 assertions could call abort()).
4493 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
4495 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
4497 *) Add various sanity checks to asn1_get_length() to reject
4498 the ASN1 length bytes if they exceed sizeof(long), will appear
4499 negative or the content length exceeds the length of the
4501 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4503 *) Fix cipher selection routines: ciphers without encryption had no flags
4504 for the cipher strength set and where therefore not handled correctly
4505 by the selection routines (PR #130).
4508 *) Fix EVP_dsa_sha macro.
4512 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
4513 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
4514 that was added in OpenSSL 0.9.6d.
4516 As the countermeasure turned out to be incompatible with some
4517 broken SSL implementations, the new option is part of SSL_OP_ALL.
4518 SSL_OP_ALL is usually employed when compatibility with weird SSL
4519 implementations is desired (e.g. '-bugs' option to 's_client' and
4520 's_server'), so the new option is automatically set in many
4524 *) Changes in security patch:
4526 Changes marked "(CHATS)" were sponsored by the Defense Advanced
4527 Research Projects Agency (DARPA) and Air Force Research Laboratory,
4528 Air Force Materiel Command, USAF, under agreement number
4531 *) Add various sanity checks to asn1_get_length() to reject
4532 the ASN1 length bytes if they exceed sizeof(long), will appear
4533 negative or the content length exceeds the length of the
4534 supplied buffer. (CVE-2002-0659)
4535 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4537 *) Assertions for various potential buffer overflows, not known to
4539 [Ben Laurie (CHATS)]
4541 *) Various temporary buffers to hold ASCII versions of integers were
4542 too small for 64 bit platforms. (CVE-2002-0655)
4543 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
4545 *) Remote buffer overflow in SSL3 protocol - an attacker could
4546 supply an oversized session ID to a client. (CVE-2002-0656)
4547 [Ben Laurie (CHATS)]
4549 *) Remote buffer overflow in SSL2 protocol - an attacker could
4550 supply an oversized client master key. (CVE-2002-0656)
4551 [Ben Laurie (CHATS)]
4553 Changes between 0.9.6c and 0.9.6d [9 May 2002]
4555 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
4556 encoded as NULL) with id-dsa-with-sha1.
4557 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
4559 *) Check various X509_...() return values in apps/req.c.
4560 [Nils Larsch <nla@trustcenter.de>]
4562 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
4563 an end-of-file condition would erronously be flagged, when the CRLF
4564 was just at the end of a processed block. The bug was discovered when
4565 processing data through a buffering memory BIO handing the data to a
4566 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
4567 <ptsekov@syntrex.com> and Nedelcho Stanev.
4570 *) Implement a countermeasure against a vulnerability recently found
4571 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
4572 before application data chunks to avoid the use of known IVs
4573 with data potentially chosen by the attacker.
4576 *) Fix length checks in ssl3_get_client_hello().
4579 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
4580 to prevent ssl3_read_internal() from incorrectly assuming that
4581 ssl3_read_bytes() found application data while handshake
4582 processing was enabled when in fact s->s3->in_read_app_data was
4583 merely automatically cleared during the initial handshake.
4584 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
4586 *) Fix object definitions for Private and Enterprise: they were not
4587 recognized in their shortname (=lowercase) representation. Extend
4588 obj_dat.pl to issue an error when using undefined keywords instead
4589 of silently ignoring the problem (Svenning Sorensen
4590 <sss@sss.dnsalias.net>).
4593 *) Fix DH_generate_parameters() so that it works for 'non-standard'
4594 generators, i.e. generators other than 2 and 5. (Previously, the
4595 code did not properly initialise the 'add' and 'rem' values to
4596 BN_generate_prime().)
4598 In the new general case, we do not insist that 'generator' is
4599 actually a primitive root: This requirement is rather pointless;
4600 a generator of the order-q subgroup is just as good, if not
4604 *) Map new X509 verification errors to alerts. Discovered and submitted by
4605 Tom Wu <tom@arcot.com>.
4608 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
4609 returning non-zero before the data has been completely received
4610 when using non-blocking I/O.
4611 [Bodo Moeller; problem pointed out by John Hughes]
4613 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
4614 [Ben Laurie, Lutz Jaenicke]
4616 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
4617 Yoram Zahavi <YoramZ@gilian.com>).
4620 *) Add information about CygWin 1.3 and on, and preserve proper
4621 configuration for the versions before that.
4622 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4624 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
4625 check whether we deal with a copy of a session and do not delete from
4626 the cache in this case. Problem reported by "Izhar Shoshani Levi"
4627 <izhar@checkpoint.com>.
4630 *) Do not store session data into the internal session cache, if it
4631 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
4632 flag is set). Proposed by Aslam <aslam@funk.com>.
4635 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
4639 *) [In 0.9.6d-engine release:]
4640 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4641 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4643 *) Add the configuration target linux-s390x.
4644 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
4646 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
4647 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
4648 variable as an indication that a ClientHello message has been
4649 received. As the flag value will be lost between multiple
4650 invocations of ssl3_accept when using non-blocking I/O, the
4651 function may not be aware that a handshake has actually taken
4652 place, thus preventing a new session from being added to the
4655 To avoid this problem, we now set s->new_session to 2 instead of
4656 using a local variable.
4657 [Lutz Jaenicke, Bodo Moeller]
4659 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
4660 if the SSL_R_LENGTH_MISMATCH error is detected.
4661 [Geoff Thorpe, Bodo Moeller]
4663 *) New 'shared_ldflag' column in Configure platform table.
4666 *) Fix EVP_CIPHER_mode macro.
4667 ["Dan S. Camper" <dan@bti.net>]
4669 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
4670 type, we must throw them away by setting rr->length to 0.
4671 [D P Chang <dpc@qualys.com>]
4673 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
4675 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
4676 <Dominikus.Scherkl@biodata.com>. (The previous implementation
4677 worked incorrectly for those cases where range = 10..._2 and
4678 3*range is two bits longer than range.)
4681 *) Only add signing time to PKCS7 structures if it is not already
4685 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
4686 OBJ_ld_ce should be OBJ_id_ce.
4687 Also some ip-pda OIDs in crypto/objects/objects.txt were
4688 incorrect (cf. RFC 3039).
4689 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
4691 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
4692 returns early because it has nothing to do.
4693 [Andy Schneider <andy.schneider@bjss.co.uk>]
4695 *) [In 0.9.6c-engine release:]
4696 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
4697 [Andy Schneider <andy.schneider@bjss.co.uk>]
4699 *) [In 0.9.6c-engine release:]
4700 Add support for Cryptographic Appliance's keyserver technology.
4701 (Use engine 'keyclient')
4702 [Cryptographic Appliances and Geoff Thorpe]
4704 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
4705 is called via tools/c89.sh because arguments have to be
4706 rearranged (all '-L' options must appear before the first object
4708 [Richard Shapiro <rshapiro@abinitio.com>]
4710 *) [In 0.9.6c-engine release:]
4711 Add support for Broadcom crypto accelerator cards, backported
4713 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
4715 *) [In 0.9.6c-engine release:]
4716 Add support for SureWare crypto accelerator cards from
4717 Baltimore Technologies. (Use engine 'sureware')
4718 [Baltimore Technologies and Mark Cox]
4720 *) [In 0.9.6c-engine release:]
4721 Add support for crypto accelerator cards from Accelerated
4722 Encryption Processing, www.aep.ie. (Use engine 'aep')
4723 [AEP Inc. and Mark Cox]
4725 *) Add a configuration entry for gcc on UnixWare.
4726 [Gary Benson <gbenson@redhat.com>]
4728 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
4729 messages are stored in a single piece (fixed-length part and
4730 variable-length part combined) and fix various bugs found on the way.
4733 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
4734 instead. BIO_gethostbyname() does not know what timeouts are
4735 appropriate, so entries would stay in cache even when they have
4737 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
4739 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
4740 faced with a pathologically small ClientHello fragment that does
4741 not contain client_version: Instead of aborting with an error,
4742 simply choose the highest available protocol version (i.e.,
4743 TLS 1.0 unless it is disabled). In practice, ClientHello
4744 messages are never sent like this, but this change gives us
4745 strictly correct behaviour at least for TLS.
4748 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
4749 never resets s->method to s->ctx->method when called from within
4750 one of the SSL handshake functions.
4751 [Bodo Moeller; problem pointed out by Niko Baric]
4753 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
4754 (sent using the client's version number) if client_version is
4755 smaller than the protocol version in use. Also change
4756 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
4757 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
4758 the client will at least see that alert.
4761 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
4765 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
4766 client receives HelloRequest while in a handshake.
4767 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
4769 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
4770 should end in 'break', not 'goto end' which circuments various
4771 cleanups done in state SSL_ST_OK. But session related stuff
4772 must be disabled for SSL_ST_OK in the case that we just sent a
4775 Also avoid some overhead by not calling ssl_init_wbio_buffer()
4776 before just sending a HelloRequest.
4777 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
4779 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
4780 reveal whether illegal block cipher padding was found or a MAC
4781 verification error occured. (Neither SSLerr() codes nor alerts
4782 are directly visible to potential attackers, but the information
4783 may leak via logfiles.)
4785 Similar changes are not required for the SSL 2.0 implementation
4786 because the number of padding bytes is sent in clear for SSL 2.0,
4787 and the extra bytes are just ignored. However ssl/s2_pkt.c
4788 failed to verify that the purported number of padding bytes is in
4792 *) Add OpenUNIX-8 support including shared libraries
4793 (Boyd Lynn Gerber <gerberb@zenez.com>).
4796 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
4797 'wristwatch attack' using huge encoding parameters (cf.
4798 James H. Manger's CRYPTO 2001 paper). Note that the
4799 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
4800 encoding parameters and hence was not vulnerable.
4803 *) BN_sqr() bug fix.
4804 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
4806 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
4807 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
4808 followed by modular reduction.
4809 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
4811 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
4812 equivalent based on BN_pseudo_rand() instead of BN_rand().
4815 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
4816 This function was broken, as the check for a new client hello message
4817 to handle SGC did not allow these large messages.
4818 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
4821 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
4824 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
4825 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
4828 *) Rework the configuration and shared library support for Tru64 Unix.
4829 The configuration part makes use of modern compiler features and
4830 still retains old compiler behavior for those that run older versions
4831 of the OS. The shared library support part includes a variant that
4832 uses the RPATH feature, and is available through the special
4833 configuration target "alpha-cc-rpath", which will never be selected
4835 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
4837 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
4838 with the same message size as in ssl3_get_certificate_request().
4839 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
4840 messages might inadvertently be reject as too long.
4841 [Petr Lampa <lampa@fee.vutbr.cz>]
4843 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
4846 *) Modified SSL library such that the verify_callback that has been set
4847 specificly for an SSL object with SSL_set_verify() is actually being
4848 used. Before the change, a verify_callback set with this function was
4849 ignored and the verify_callback() set in the SSL_CTX at the time of
4850 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
4851 to allow the necessary settings.
4854 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
4855 explicitly to NULL, as at least on Solaris 8 this seems not always to be
4856 done automatically (in contradiction to the requirements of the C
4857 standard). This made problems when used from OpenSSH.
4860 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
4861 dh->length and always used
4863 BN_rand_range(priv_key, dh->p).
4865 BN_rand_range() is not necessary for Diffie-Hellman, and this
4866 specific range makes Diffie-Hellman unnecessarily inefficient if
4867 dh->length (recommended exponent length) is much smaller than the
4868 length of dh->p. We could use BN_rand_range() if the order of
4869 the subgroup was stored in the DH structure, but we only have
4874 BN_rand(priv_key, l, ...)
4876 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
4882 RSA_eay_public_encrypt
4883 RSA_eay_private_decrypt
4884 RSA_eay_private_encrypt (signing)
4885 RSA_eay_public_decrypt (signature verification)
4887 (default implementations for RSA_public_encrypt,
4888 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
4889 always reject numbers >= n.
4892 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
4893 to synchronize access to 'locking_thread'. This is necessary on
4894 systems where access to 'locking_thread' (an 'unsigned long'
4895 variable) is not atomic.
4898 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
4899 *before* setting the 'crypto_lock_rand' flag. The previous code had
4900 a race condition if 0 is a valid thread ID.
4901 [Travis Vitek <vitek@roguewave.com>]
4903 *) Add support for shared libraries under Irix.
4904 [Albert Chin-A-Young <china@thewrittenword.com>]
4906 *) Add configuration option to build on Linux on both big-endian and
4908 [Ralf Baechle <ralf@uni-koblenz.de>]
4910 *) Add the possibility to create shared libraries on HP-UX.
4913 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
4915 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
4916 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
4917 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
4918 PRNG state recovery was possible based on the output of
4919 one PRNG request appropriately sized to gain knowledge on
4920 'md' followed by enough consecutive 1-byte PRNG requests
4921 to traverse all of 'state'.
4923 1. When updating 'md_local' (the current thread's copy of 'md')
4924 during PRNG output generation, hash all of the previous
4925 'md_local' value, not just the half used for PRNG output.
4927 2. Make the number of bytes from 'state' included into the hash
4928 independent from the number of PRNG bytes requested.
4930 The first measure alone would be sufficient to avoid
4931 Markku-Juhani's attack. (Actually it had never occurred
4932 to me that the half of 'md_local' used for chaining was the
4933 half from which PRNG output bytes were taken -- I had always
4934 assumed that the secret half would be used.) The second
4935 measure makes sure that additional data from 'state' is never
4936 mixed into 'md_local' in small portions; this heuristically
4937 further strengthens the PRNG.
4940 *) Fix crypto/bn/asm/mips3.s.
4943 *) When only the key is given to "enc", the IV is undefined. Print out
4944 an error message in this case.
4947 *) Handle special case when X509_NAME is empty in X509 printing routines.
4950 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
4951 positive and less than q.
4954 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
4955 used: it isn't thread safe and the add_lock_callback should handle
4957 [Paul Rose <Paul.Rose@bridge.com>]
4959 *) Verify that incoming data obeys the block size in
4960 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
4964 [Ulf Möller, Bodo Möller]
4966 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
4967 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
4968 when fixing the server behaviour for backwards-compatible 'client
4969 hello' messages. (Note that the attack is impractical against
4970 SSL 3.0 and TLS 1.0 anyway because length and version checking
4971 means that the probability of guessing a valid ciphertext is
4972 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
4975 Before 0.9.5, the countermeasure (hide the error by generating a
4976 random 'decryption result') did not work properly because
4977 ERR_clear_error() was missing, meaning that SSL_get_error() would
4978 detect the supposedly ignored error.
4980 Both problems are now fixed.
4983 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
4984 (previously it was 1024).
4987 *) Fix for compatibility mode trust settings: ignore trust settings
4988 unless some valid trust or reject settings are present.
4991 *) Fix for blowfish EVP: its a variable length cipher.
4994 *) Fix various bugs related to DSA S/MIME verification. Handle missing
4995 parameters in DSA public key structures and return an error in the
4996 DSA routines if parameters are absent.
4999 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
5000 in the current directory if neither $RANDFILE nor $HOME was set.
5001 RAND_file_name() in 0.9.6a returned NULL in this case. This has
5002 caused some confusion to Windows users who haven't defined $HOME.
5003 Thus RAND_file_name() is changed again: e_os.h can define a
5004 DEFAULT_HOME, which will be used if $HOME is not set.
5005 For Windows, we use "C:"; on other platforms, we still require
5006 environment variables.
5008 *) Move 'if (!initialized) RAND_poll()' into regions protected by
5009 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
5010 having multiple threads call RAND_poll() concurrently.
5013 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
5014 combination of a flag and a thread ID variable.
5015 Otherwise while one thread is in ssleay_rand_bytes (which sets the
5016 flag), *other* threads can enter ssleay_add_bytes without obeying
5017 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
5018 that they do not hold after the first thread unsets add_do_not_lock).
5021 *) Change bctest again: '-x' expressions are not available in all
5025 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
5027 *) Fix a couple of memory leaks in PKCS7_dataDecode()
5028 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
5030 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
5031 the default extension for executables, if any. Also, make the perl
5032 scripts that use symlink() to test if it really exists and use "cp"
5033 if it doesn't. All this made OpenSSL compilable and installable in
5037 *) Fix for asn1_GetSequence() for indefinite length constructed data.
5038 If SEQUENCE is length is indefinite just set c->slen to the total
5039 amount of data available.
5040 [Steve Henson, reported by shige@FreeBSD.org]
5041 [This change does not apply to 0.9.7.]
5043 *) Change bctest to avoid here-documents inside command substitution
5044 (workaround for FreeBSD /bin/sh bug).
5045 For compatibility with Ultrix, avoid shell functions (introduced
5046 in the bctest version that searches along $PATH).
5049 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
5050 with des_encrypt() defined on some operating systems, like Solaris
5054 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
5055 On the Importance of Eliminating Errors in Cryptographic
5056 Computations, J. Cryptology 14 (2001) 2, 101-119,
5057 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
5060 *) MIPS assembler BIGNUM division bug fix.
5063 *) Disabled incorrect Alpha assembler code.
5066 *) Fix PKCS#7 decode routines so they correctly update the length
5067 after reading an EOC for the EXPLICIT tag.
5069 [This change does not apply to 0.9.7.]
5071 *) Fix bug in PKCS#12 key generation routines. This was triggered
5072 if a 3DES key was generated with a 0 initial byte. Include
5073 PKCS12_BROKEN_KEYGEN compilation option to retain the old
5074 (but broken) behaviour.
5077 *) Enhance bctest to search for a working bc along $PATH and print
5079 [Tim Rice <tim@multitalents.net> via Richard Levitte]
5081 *) Fix memory leaks in err.c: free err_data string if necessary;
5082 don't write to the wrong index in ERR_set_error_data.
5085 *) Implement ssl23_peek (analogous to ssl23_read), which previously
5089 *) Replace rdtsc with _emit statements for VC++ version 5.
5090 [Jeremy Cooper <jeremy@baymoo.org>]
5092 *) Make it possible to reuse SSLv2 sessions.
5095 *) In copy_email() check for >= 0 as a return value for
5096 X509_NAME_get_index_by_NID() since 0 is a valid index.
5097 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
5099 *) Avoid coredump with unsupported or invalid public keys by checking if
5100 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
5101 PKCS7_verify() fails with non detached data.
5104 *) Don't use getenv in library functions when run as setuid/setgid.
5105 New function OPENSSL_issetugid().
5108 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
5109 due to incorrect handling of multi-threading:
5111 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
5113 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
5115 3. Count how many times MemCheck_off() has been called so that
5116 nested use can be treated correctly. This also avoids
5117 inband-signalling in the previous code (which relied on the
5118 assumption that thread ID 0 is impossible).
5121 *) Add "-rand" option also to s_client and s_server.
5124 *) Fix CPU detection on Irix 6.x.
5125 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
5126 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
5128 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
5131 [This change does not apply to 0.9.7.]
5133 *) Use the cached encoding of an X509_NAME structure rather than
5134 copying it. This is apparently the reason for the libsafe "errors"
5135 but the code is actually correct.
5138 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
5139 Bleichenbacher's DSA attack.
5140 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
5141 to be set and top=0 forces the highest bit to be set; top=-1 is new
5142 and leaves the highest bit random.
5143 [Ulf Moeller, Bodo Moeller]
5145 *) In the NCONF_...-based implementations for CONF_... queries
5146 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
5147 a temporary CONF structure with the data component set to NULL
5148 (which gives segmentation faults in lh_retrieve).
5149 Instead, use NULL for the CONF pointer in CONF_get_string and
5150 CONF_get_number (which may use environment variables) and directly
5151 return NULL from CONF_get_section.
5154 *) Fix potential buffer overrun for EBCDIC.
5157 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
5158 keyUsage if basicConstraints absent for a CA.
5161 *) Make SMIME_write_PKCS7() write mail header values with a format that
5162 is more generally accepted (no spaces before the semicolon), since
5163 some programs can't parse those values properly otherwise. Also make
5164 sure BIO's that break lines after each write do not create invalid
5168 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
5169 macros previously used would not encode an empty SEQUENCE OF
5170 and break the signature.
5172 [This change does not apply to 0.9.7.]
5174 *) Zero the premaster secret after deriving the master secret in
5178 *) Add some EVP_add_digest_alias registrations (as found in
5179 OpenSSL_add_all_digests()) to SSL_library_init()
5180 aka OpenSSL_add_ssl_algorithms(). This provides improved
5181 compatibility with peers using X.509 certificates
5182 with unconventional AlgorithmIdentifier OIDs.
5185 *) Fix for Irix with NO_ASM.
5186 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
5188 *) ./config script fixes.
5189 [Ulf Moeller, Richard Levitte]
5191 *) Fix 'openssl passwd -1'.
5194 *) Change PKCS12_key_gen_asc() so it can cope with non null
5195 terminated strings whose length is passed in the passlen
5196 parameter, for example from PEM callbacks. This was done
5197 by adding an extra length parameter to asc2uni().
5198 [Steve Henson, reported by <oddissey@samsung.co.kr>]
5200 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
5201 call failed, free the DSA structure.
5204 *) Fix to uni2asc() to cope with zero length Unicode strings.
5205 These are present in some PKCS#12 files.
5208 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
5209 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
5210 when writing a 32767 byte record.
5211 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
5213 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
5214 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
5216 (RSA objects have a reference count access to which is protected
5217 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
5218 so they are meant to be shared between threads.)
5219 [Bodo Moeller, Geoff Thorpe; original patch submitted by
5220 "Reddie, Steven" <Steven.Reddie@ca.com>]
5222 *) Fix a deadlock in CRYPTO_mem_leaks().
5225 *) Use better test patterns in bntest.
5228 *) rand_win.c fix for Borland C.
5231 *) BN_rshift bugfix for n == 0.
5234 *) Add a 'bctest' script that checks for some known 'bc' bugs
5235 so that 'make test' does not abort just because 'bc' is broken.
5238 *) Store verify_result within SSL_SESSION also for client side to
5239 avoid potential security hole. (Re-used sessions on the client side
5240 always resulted in verify_result==X509_V_OK, not using the original
5241 result of the server certificate verification.)
5244 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
5245 SSL3_RT_APPLICATION_DATA, return 0.
5246 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
5250 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
5251 releases, have been re-implemented by renaming the previous
5252 implementations of ssl2_read and ssl3_read to ssl2_read_internal
5253 and ssl3_read_internal, respectively, and adding 'peek' parameters
5254 to them. The new ssl[23]_{read,peek} functions are calls to
5255 ssl[23]_read_internal with the 'peek' flag set appropriately.
5256 A 'peek' parameter has also been added to ssl3_read_bytes, which
5257 does the actual work for ssl3_read_internal.
5260 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
5261 the method-specific "init()" handler. Also clean up ex_data after
5262 calling the method-specific "finish()" handler. Previously, this was
5263 happening the other way round.
5266 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
5267 The previous value, 12, was not always sufficient for BN_mod_exp().
5270 *) Make sure that shared libraries get the internal name engine with
5271 the full version number and not just 0. This should mark the
5272 shared libraries as not backward compatible. Of course, this should
5273 be changed again when we can guarantee backward binary compatibility.
5276 *) Fix typo in get_cert_by_subject() in by_dir.c
5277 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
5279 *) Rework the system to generate shared libraries:
5281 - Make note of the expected extension for the shared libraries and
5282 if there is a need for symbolic links from for example libcrypto.so.0
5283 to libcrypto.so.0.9.7. There is extended info in Configure for
5286 - Make as few rebuilds of the shared libraries as possible.
5288 - Still avoid linking the OpenSSL programs with the shared libraries.
5290 - When installing, install the shared libraries separately from the
5294 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
5296 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
5297 and not in SSL_clear because the latter is also used by the
5298 accept/connect functions; previously, the settings made by
5299 SSL_set_read_ahead would be lost during the handshake.
5300 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
5302 *) Correct util/mkdef.pl to be selective about disabled algorithms.
5303 Previously, it would create entries for disableed algorithms no
5307 *) Added several new manual pages for SSL_* function.
5310 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
5312 *) In ssl23_get_client_hello, generate an error message when faced
5313 with an initial SSL 3.0/TLS record that is too small to contain the
5314 first two bytes of the ClientHello message, i.e. client_version.
5315 (Note that this is a pathologic case that probably has never happened
5316 in real life.) The previous approach was to use the version number
5317 from the record header as a substitute; but our protocol choice
5318 should not depend on that one because it is not authenticated
5319 by the Finished messages.
5322 *) More robust randomness gathering functions for Windows.
5323 [Jeffrey Altman <jaltman@columbia.edu>]
5325 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
5326 not set then we don't setup the error code for issuer check errors
5327 to avoid possibly overwriting other errors which the callback does
5328 handle. If an application does set the flag then we assume it knows
5329 what it is doing and can handle the new informational codes
5333 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
5334 a general "ANY" type, as such it should be able to decode anything
5335 including tagged types. However it didn't check the class so it would
5336 wrongly interpret tagged types in the same way as their universal
5337 counterpart and unknown types were just rejected. Changed so that the
5338 tagged and unknown types are handled in the same way as a SEQUENCE:
5339 that is the encoding is stored intact. There is also a new type
5340 "V_ASN1_OTHER" which is used when the class is not universal, in this
5341 case we have no idea what the actual type is so we just lump them all
5345 *) On VMS, stdout may very well lead to a file that is written to
5346 in a record-oriented fashion. That means that every write() will
5347 write a separate record, which will be read separately by the
5348 programs trying to read from it. This can be very confusing.
5350 The solution is to put a BIO filter in the way that will buffer
5351 text until a linefeed is reached, and then write everything a
5352 line at a time, so every record written will be an actual line,
5353 not chunks of lines and not (usually doesn't happen, but I've
5354 seen it once) several lines in one record. BIO_f_linebuffer() is
5357 Currently, it's a VMS-only method, because that's where it has
5358 been tested well enough.
5361 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
5362 it can return incorrect results.
5363 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
5364 but it was in 0.9.6-beta[12].)
5367 *) Disable the check for content being present when verifying detached
5368 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
5369 include zero length content when signing messages.
5372 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
5373 BIO_ctrl (for BIO pairs).
5376 *) Add DSO method for VMS.
5379 *) Bug fix: Montgomery multiplication could produce results with the
5383 *) Add RPM specification openssl.spec and modify it to build three
5384 packages. The default package contains applications, application
5385 documentation and run-time libraries. The devel package contains
5386 include files, static libraries and function documentation. The
5387 doc package contains the contents of the doc directory. The original
5388 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
5391 *) Add a large number of documentation files for many SSL routines.
5392 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5394 *) Add a configuration entry for Sony News 4.
5395 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
5397 *) Don't set the two most significant bits to one when generating a
5398 random number < q in the DSA library.
5401 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
5402 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
5403 the underlying transport is blocking) if a handshake took place.
5404 (The default behaviour is needed by applications such as s_client
5405 and s_server that use select() to determine when to use SSL_read;
5406 but for applications that know in advance when to expect data, it
5407 just makes things more complicated.)
5410 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
5414 *) Add a few more EBCDIC conditionals that make `req' and `x509'
5415 work better on such systems.
5416 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5418 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
5419 Update PKCS12_parse() so it copies the friendlyName and the
5420 keyid to the certificates aux info.
5423 *) Fix bug in PKCS7_verify() which caused an infinite loop
5424 if there was more than one signature.
5425 [Sven Uszpelkat <su@celocom.de>]
5427 *) Major change in util/mkdef.pl to include extra information
5428 about each symbol, as well as presentig variables as well
5429 as functions. This change means that there's n more need
5430 to rebuild the .num files when some algorithms are excluded.
5433 *) Allow the verify time to be set by an application,
5434 rather than always using the current time.
5437 *) Phase 2 verify code reorganisation. The certificate
5438 verify code now looks up an issuer certificate by a
5439 number of criteria: subject name, authority key id
5440 and key usage. It also verifies self signed certificates
5441 by the same criteria. The main comparison function is
5442 X509_check_issued() which performs these checks.
5444 Lot of changes were necessary in order to support this
5445 without completely rewriting the lookup code.
5447 Authority and subject key identifier are now cached.
5449 The LHASH 'certs' is X509_STORE has now been replaced
5450 by a STACK_OF(X509_OBJECT). This is mainly because an
5451 LHASH can't store or retrieve multiple objects with
5452 the same hash value.
5454 As a result various functions (which were all internal
5455 use only) have changed to handle the new X509_STORE
5456 structure. This will break anything that messed round
5457 with X509_STORE internally.
5459 The functions X509_STORE_add_cert() now checks for an
5460 exact match, rather than just subject name.
5462 The X509_STORE API doesn't directly support the retrieval
5463 of multiple certificates matching a given criteria, however
5464 this can be worked round by performing a lookup first
5465 (which will fill the cache with candidate certificates)
5466 and then examining the cache for matches. This is probably
5467 the best we can do without throwing out X509_LOOKUP
5468 entirely (maybe later...).
5470 The X509_VERIFY_CTX structure has been enhanced considerably.
5472 All certificate lookup operations now go via a get_issuer()
5473 callback. Although this currently uses an X509_STORE it
5474 can be replaced by custom lookups. This is a simple way
5475 to bypass the X509_STORE hackery necessary to make this
5476 work and makes it possible to use more efficient techniques
5477 in future. A very simple version which uses a simple
5478 STACK for its trusted certificate store is also provided
5479 using X509_STORE_CTX_trusted_stack().
5481 The verify_cb() and verify() callbacks now have equivalents
5482 in the X509_STORE_CTX structure.
5484 X509_STORE_CTX also has a 'flags' field which can be used
5485 to customise the verify behaviour.
5488 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
5489 excludes S/MIME capabilities.
5492 *) When a certificate request is read in keep a copy of the
5493 original encoding of the signed data and use it when outputing
5494 again. Signatures then use the original encoding rather than
5495 a decoded, encoded version which may cause problems if the
5496 request is improperly encoded.
5499 *) For consistency with other BIO_puts implementations, call
5500 buffer_write(b, ...) directly in buffer_puts instead of calling
5503 In BIO_puts, increment b->num_write as in BIO_write.
5504 [Peter.Sylvester@EdelWeb.fr]
5506 *) Fix BN_mul_word for the case where the word is 0. (We have to use
5507 BN_zero, we may not return a BIGNUM with an array consisting of
5511 *) Avoid calling abort() from within the library when problems are
5512 detected, except if preprocessor symbols have been defined
5513 (such as REF_CHECK, BN_DEBUG etc.).
5516 *) New openssl application 'rsautl'. This utility can be
5517 used for low level RSA operations. DER public key
5518 BIO/fp routines also added.
5521 *) New Configure entry and patches for compiling on QNX 4.
5522 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
5524 *) A demo state-machine implementation was sponsored by
5525 Nuron (http://www.nuron.com/) and is now available in
5526 demos/state_machine.
5529 *) New options added to the 'dgst' utility for signature
5530 generation and verification.
5533 *) Unrecognized PKCS#7 content types are now handled via a
5534 catch all ASN1_TYPE structure. This allows unsupported
5535 types to be stored as a "blob" and an application can
5536 encode and decode it manually.
5539 *) Fix various signed/unsigned issues to make a_strex.c
5541 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
5543 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
5544 length if passed a buffer. ASN1_INTEGER_to_BN failed
5545 if passed a NULL BN and its argument was negative.
5546 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
5548 *) Modification to PKCS#7 encoding routines to output definite
5549 length encoding. Since currently the whole structures are in
5550 memory there's not real point in using indefinite length
5551 constructed encoding. However if OpenSSL is compiled with
5552 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
5555 *) Added BIO_vprintf() and BIO_vsnprintf().
5558 *) Added more prefixes to parse for in the the strings written
5559 through a logging bio, to cover all the levels that are available
5560 through syslog. The prefixes are now:
5562 PANIC, EMERG, EMR => LOG_EMERG
5563 ALERT, ALR => LOG_ALERT
5564 CRIT, CRI => LOG_CRIT
5565 ERROR, ERR => LOG_ERR
5566 WARNING, WARN, WAR => LOG_WARNING
5567 NOTICE, NOTE, NOT => LOG_NOTICE
5568 INFO, INF => LOG_INFO
5569 DEBUG, DBG => LOG_DEBUG
5571 and as before, if none of those prefixes are present at the
5572 beginning of the string, LOG_ERR is chosen.
5574 On Win32, the LOG_* levels are mapped according to this:
5576 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
5577 LOG_WARNING => EVENTLOG_WARNING_TYPE
5578 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
5582 *) Made it possible to reconfigure with just the configuration
5583 argument "reconf" or "reconfigure". The command line arguments
5584 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
5585 and are retrieved from there when reconfiguring.
5589 [Assar Westerlund <assar@sics.se>, Richard Levitte]
5591 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
5594 *) The obj_dat.pl script was messing up the sorting of object
5595 names. The reason was that it compared the quoted version
5596 of strings as a result "OCSP" > "OCSP Signing" because
5597 " > SPACE. Changed script to store unquoted versions of
5598 names and add quotes on output. It was also omitting some
5599 names from the lookup table if they were given a default
5600 value (that is if SN is missing it is given the same
5601 value as LN and vice versa), these are now added on the
5602 grounds that if an object has a name we should be able to
5603 look it up. Finally added warning output when duplicate
5604 short or long names are found.
5607 *) Changes needed for Tandem NSK.
5608 [Scott Uroff <scott@xypro.com>]
5610 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
5611 RSA_padding_check_SSLv23(), special padding was never detected
5612 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
5613 version rollback attacks was not effective.
5615 In s23_clnt.c, don't use special rollback-attack detection padding
5616 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
5617 client; similarly, in s23_srvr.c, don't do the rollback check if
5618 SSL 2.0 is the only protocol enabled in the server.
5621 *) Make it possible to get hexdumps of unprintable data with 'openssl
5622 asn1parse'. By implication, the functions ASN1_parse_dump() and
5623 BIO_dump_indent() are added.
5626 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
5627 these print out strings and name structures based on various
5628 flags including RFC2253 support and proper handling of
5629 multibyte characters. Added options to the 'x509' utility
5630 to allow the various flags to be set.
5633 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
5634 Also change the functions X509_cmp_current_time() and
5635 X509_gmtime_adj() work with an ASN1_TIME structure,
5636 this will enable certificates using GeneralizedTime in validity
5637 dates to be checked.
5640 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
5641 negative public key encodings) on by default,
5642 NO_NEG_PUBKEY_BUG can be set to disable it.
5645 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
5646 content octets. An i2c_ASN1_OBJECT is unnecessary because
5647 the encoding can be trivially obtained from the structure.
5650 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
5651 not read locks (CRYPTO_r_[un]lock).
5654 *) A first attempt at creating official support for shared
5655 libraries through configuration. I've kept it so the
5656 default is static libraries only, and the OpenSSL programs
5657 are always statically linked for now, but there are
5658 preparations for dynamic linking in place.
5659 This has been tested on Linux and Tru64.
5662 *) Randomness polling function for Win9x, as described in:
5663 Peter Gutmann, Software Generation of Practically Strong
5667 *) Fix so PRNG is seeded in req if using an already existing
5671 *) New options to smime application. -inform and -outform
5672 allow alternative formats for the S/MIME message including
5673 PEM and DER. The -content option allows the content to be
5674 specified separately. This should allow things like Netscape
5675 form signing output easier to verify.
5678 *) Fix the ASN1 encoding of tags using the 'long form'.
5681 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
5682 STRING types. These convert content octets to and from the
5683 underlying type. The actual tag and length octets are
5684 already assumed to have been read in and checked. These
5685 are needed because all other string types have virtually
5686 identical handling apart from the tag. By having versions
5687 of the ASN1 functions that just operate on content octets
5688 IMPLICIT tagging can be handled properly. It also allows
5689 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
5690 and ASN1_INTEGER are identical apart from the tag.
5693 *) Change the handling of OID objects as follows:
5695 - New object identifiers are inserted in objects.txt, following
5696 the syntax given in objects.README.
5697 - objects.pl is used to process obj_mac.num and create a new
5699 - obj_dat.pl is used to create a new obj_dat.h, using the data in
5702 This is currently kind of a hack, and the perl code in objects.pl
5703 isn't very elegant, but it works as I intended. The simplest way
5704 to check that it worked correctly is to look in obj_dat.h and
5705 check the array nid_objs and make sure the objects haven't moved
5706 around (this is important!). Additions are OK, as well as
5707 consistent name changes.
5710 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
5713 *) Addition of the command line parameter '-rand file' to 'openssl req'.
5714 The given file adds to whatever has already been seeded into the
5715 random pool through the RANDFILE configuration file option or
5716 environment variable, or the default random state file.
5719 *) mkstack.pl now sorts each macro group into lexical order.
5720 Previously the output order depended on the order the files
5721 appeared in the directory, resulting in needless rewriting
5725 *) Patches to make OpenSSL compile under Win32 again. Mostly
5726 work arounds for the VC++ problem that it treats func() as
5727 func(void). Also stripped out the parts of mkdef.pl that
5728 added extra typesafe functions: these no longer exist.
5731 *) Reorganisation of the stack code. The macros are now all
5732 collected in safestack.h . Each macro is defined in terms of
5733 a "stack macro" of the form SKM_<name>(type, a, b). The
5734 DEBUG_SAFESTACK is now handled in terms of function casts,
5735 this has the advantage of retaining type safety without the
5736 use of additional functions. If DEBUG_SAFESTACK is not defined
5737 then the non typesafe macros are used instead. Also modified the
5738 mkstack.pl script to handle the new form. Needs testing to see
5739 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
5740 the default if no major problems. Similar behaviour for ASN1_SET_OF
5741 and PKCS12_STACK_OF.
5744 *) When some versions of IIS use the 'NET' form of private key the
5745 key derivation algorithm is different. Normally MD5(password) is
5746 used as a 128 bit RC4 key. In the modified case
5747 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
5748 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
5749 as the old Netscape_RSA functions except they have an additional
5750 'sgckey' parameter which uses the modified algorithm. Also added
5751 an -sgckey command line option to the rsa utility. Thanks to
5752 Adrian Peck <bertie@ncipher.com> for posting details of the modified
5753 algorithm to openssl-dev.
5756 *) The evp_local.h macros were using 'c.##kname' which resulted in
5757 invalid expansion on some systems (SCO 5.0.5 for example).
5758 Corrected to 'c.kname'.
5759 [Phillip Porch <root@theporch.com>]
5761 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
5762 a STACK of email addresses from a certificate or request, these look
5763 in the subject name and the subject alternative name extensions and
5764 omit any duplicate addresses.
5767 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
5768 This makes DSA verification about 2 % faster.
5771 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
5772 (meaning that now 2^5 values will be precomputed, which is only 4 KB
5773 plus overhead for 1024 bit moduli).
5774 This makes exponentiations about 0.5 % faster for 1024 bit
5775 exponents (as measured by "openssl speed rsa2048").
5778 *) Rename memory handling macros to avoid conflicts with other
5780 Malloc => OPENSSL_malloc
5781 Malloc_locked => OPENSSL_malloc_locked
5782 Realloc => OPENSSL_realloc
5783 Free => OPENSSL_free
5786 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
5787 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
5790 *) CygWin32 support.
5791 [John Jarvie <jjarvie@newsguy.com>]
5793 *) The type-safe stack code has been rejigged. It is now only compiled
5794 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
5795 by default all type-specific stack functions are "#define"d back to
5796 standard stack functions. This results in more streamlined output
5797 but retains the type-safety checking possibilities of the original
5801 *) The STACK code has been cleaned up, and certain type declarations
5802 that didn't make a lot of sense have been brought in line. This has
5803 also involved a cleanup of sorts in safestack.h to more correctly
5804 map type-safe stack functions onto their plain stack counterparts.
5805 This work has also resulted in a variety of "const"ifications of
5806 lots of the code, especially "_cmp" operations which should normally
5807 be prototyped with "const" parameters anyway.
5810 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
5811 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
5812 (The PRNG state consists of two parts, the large pool 'state' and 'md',
5813 where all of 'md' is used each time the PRNG is used, but 'state'
5814 is used only indexed by a cyclic counter. As entropy may not be
5815 well distributed from the beginning, 'md' is important as a
5816 chaining variable. However, the output function chains only half
5817 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
5818 all of 'md', and seeding with STATE_SIZE dummy bytes will result
5819 in all of 'state' being rewritten, with the new values depending
5820 on virtually all of 'md'. This overcomes the 80 bit limitation.)
5823 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
5824 the handshake is continued after ssl_verify_cert_chain();
5825 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
5826 can lead to 'unexplainable' connection aborts later.
5827 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
5829 *) Major EVP API cipher revision.
5830 Add hooks for extra EVP features. This allows various cipher
5831 parameters to be set in the EVP interface. Support added for variable
5832 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
5833 setting of RC2 and RC5 parameters.
5835 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
5838 Remove lots of duplicated code from the EVP library. For example *every*
5839 cipher init() function handles the 'iv' in the same way according to the
5840 cipher mode. They also all do nothing if the 'key' parameter is NULL and
5841 for CFB and OFB modes they zero ctx->num.
5843 New functionality allows removal of S/MIME code RC2 hack.
5845 Most of the routines have the same form and so can be declared in terms
5848 By shifting this to the top level EVP_CipherInit() it can be removed from
5849 all individual ciphers. If the cipher wants to handle IVs or keys
5850 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
5853 Change lots of functions like EVP_EncryptUpdate() to now return a
5854 value: although software versions of the algorithms cannot fail
5855 any installed hardware versions can.
5858 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
5859 this option is set, tolerate broken clients that send the negotiated
5860 protocol version number instead of the requested protocol version
5864 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
5865 i.e. non-zero for export ciphersuites, zero otherwise.
5866 Previous versions had this flag inverted, inconsistent with
5867 rsa_tmp_cb (..._TMP_RSA_CB).
5868 [Bodo Moeller; problem reported by Amit Chopra]
5870 *) Add missing DSA library text string. Work around for some IIS
5871 key files with invalid SEQUENCE encoding.
5874 *) Add a document (doc/standards.txt) that list all kinds of standards
5875 and so on that are implemented in OpenSSL.
5878 *) Enhance c_rehash script. Old version would mishandle certificates
5879 with the same subject name hash and wouldn't handle CRLs at all.
5880 Added -fingerprint option to crl utility, to support new c_rehash
5884 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
5887 *) Fix for SSL server purpose checking. Server checking was
5888 rejecting certificates which had extended key usage present
5889 but no ssl client purpose.
5890 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
5892 *) Make PKCS#12 code work with no password. The PKCS#12 spec
5893 is a little unclear about how a blank password is handled.
5894 Since the password in encoded as a BMPString with terminating
5895 double NULL a zero length password would end up as just the
5896 double NULL. However no password at all is different and is
5897 handled differently in the PKCS#12 key generation code. NS
5898 treats a blank password as zero length. MSIE treats it as no
5899 password on export: but it will try both on import. We now do
5900 the same: PKCS12_parse() tries zero length and no password if
5901 the password is set to "" or NULL (NULL is now a valid password:
5902 it wasn't before) as does the pkcs12 application.
5905 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
5906 perror when PEM_read_bio_X509_REQ fails, the error message must
5907 be obtained from the error queue.
5910 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
5911 it in ERR_remove_state if appropriate, and change ERR_get_state
5912 accordingly to avoid race conditions (this is necessary because
5913 thread_hash is no longer constant once set).
5916 *) Bugfix for linux-elf makefile.one.
5919 *) RSA_get_default_method() will now cause a default
5920 RSA_METHOD to be chosen if one doesn't exist already.
5921 Previously this was only set during a call to RSA_new()
5922 or RSA_new_method(NULL) meaning it was possible for
5923 RSA_get_default_method() to return NULL.
5926 *) Added native name translation to the existing DSO code
5927 that will convert (if the flag to do so is set) filenames
5928 that are sufficiently small and have no path information
5929 into a canonical native form. Eg. "blah" converted to
5930 "libblah.so" or "blah.dll" etc.
5933 *) New function ERR_error_string_n(e, buf, len) which is like
5934 ERR_error_string(e, buf), but writes at most 'len' bytes
5935 including the 0 terminator. For ERR_error_string_n, 'buf'
5937 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
5939 *) CONF library reworked to become more general. A new CONF
5940 configuration file reader "class" is implemented as well as a
5941 new functions (NCONF_*, for "New CONF") to handle it. The now
5942 old CONF_* functions are still there, but are reimplemented to
5943 work in terms of the new functions. Also, a set of functions
5944 to handle the internal storage of the configuration data is
5945 provided to make it easier to write new configuration file
5946 reader "classes" (I can definitely see something reading a
5947 configuration file in XML format, for example), called _CONF_*,
5948 or "the configuration storage API"...
5950 The new configuration file reading functions are:
5952 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
5953 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
5955 NCONF_default, NCONF_WIN32
5957 NCONF_dump_fp, NCONF_dump_bio
5959 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
5960 NCONF_new creates a new CONF object. This works in the same way
5961 as other interfaces in OpenSSL, like the BIO interface.
5962 NCONF_dump_* dump the internal storage of the configuration file,
5963 which is useful for debugging. All other functions take the same
5964 arguments as the old CONF_* functions wth the exception of the
5965 first that must be a `CONF *' instead of a `LHASH *'.
5967 To make it easer to use the new classes with the old CONF_* functions,
5968 the function CONF_set_default_method is provided.
5971 *) Add '-tls1' option to 'openssl ciphers', which was already
5972 mentioned in the documentation but had not been implemented.
5973 (This option is not yet really useful because even the additional
5974 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
5977 *) Initial DSO code added into libcrypto for letting OpenSSL (and
5978 OpenSSL-based applications) load shared libraries and bind to
5979 them in a portable way.
5980 [Geoff Thorpe, with contributions from Richard Levitte]
5982 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
5984 *) Make sure _lrotl and _lrotr are only used with MSVC.
5986 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
5987 (the default implementation of RAND_status).
5989 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
5990 to '-clrext' (= clear extensions), as intended and documented.
5991 [Bodo Moeller; inconsistency pointed out by Michael Attili
5994 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
5995 was larger than the MD block size.
5996 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
5998 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
5999 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
6000 using the passed key: if the passed key was a private key the result
6001 of X509_print(), for example, would be to print out all the private key
6005 *) des_quad_cksum() byte order bug fix.
6006 [Ulf Möller, using the problem description in krb4-0.9.7, where
6007 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
6009 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
6011 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
6013 *) For easily testing in shell scripts whether some command
6014 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
6015 returns with exit code 0 iff no command of the given name is available.
6016 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
6017 the output goes to stdout and nothing is printed to stderr.
6018 Additional arguments are always ignored.
6020 Since for each cipher there is a command of the same name,
6021 the 'no-cipher' compilation switches can be tested this way.
6023 ('openssl no-XXX' is not able to detect pseudo-commands such
6024 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
6027 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
6030 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
6031 is set; it will be thrown away anyway because each handshake creates
6033 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
6034 to parameters -- in previous versions (since OpenSSL 0.9.3) the
6035 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
6036 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
6039 *) New s_client option -ign_eof: EOF at stdin is ignored, and
6040 'Q' and 'R' lose their special meanings (quit/renegotiate).
6041 This is part of what -quiet does; unlike -quiet, -ign_eof
6042 does not suppress any output.
6045 *) Add compatibility options to the purpose and trust code. The
6046 purpose X509_PURPOSE_ANY is "any purpose" which automatically
6047 accepts a certificate or CA, this was the previous behaviour,
6048 with all the associated security issues.
6050 X509_TRUST_COMPAT is the old trust behaviour: only and
6051 automatically trust self signed roots in certificate store. A
6052 new trust setting X509_TRUST_DEFAULT is used to specify that
6053 a purpose has no associated trust setting and it should instead
6054 use the value in the default purpose.
6057 *) Fix the PKCS#8 DSA private key code so it decodes keys again
6058 and fix a memory leak.
6061 *) In util/mkerr.pl (which implements 'make errors'), preserve
6062 reason strings from the previous version of the .c file, as
6063 the default to have only downcase letters (and digits) in
6064 automatically generated reasons codes is not always appropriate.
6067 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
6068 using strerror. Previously, ERR_reason_error_string() returned
6069 library names as reason strings for SYSerr; but SYSerr is a special
6070 case where small numbers are errno values, not library numbers.
6073 *) Add '-dsaparam' option to 'openssl dhparam' application. This
6074 converts DSA parameters into DH parameters. (When creating parameters,
6075 DSA_generate_parameters is used.)
6078 *) Include 'length' (recommended exponent length) in C code generated
6079 by 'openssl dhparam -C'.
6082 *) The second argument to set_label in perlasm was already being used
6083 so couldn't be used as a "file scope" flag. Moved to third argument
6087 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
6088 instead of RAND_bytes for encryption IVs and salts.
6091 *) Include RAND_status() into RAND_METHOD instead of implementing
6092 it only for md_rand.c Otherwise replacing the PRNG by calling
6093 RAND_set_rand_method would be impossible.
6096 *) Don't let DSA_generate_key() enter an infinite loop if the random
6097 number generation fails.
6100 *) New 'rand' application for creating pseudo-random output.
6103 *) Added configuration support for Linux/IA64
6104 [Rolf Haberrecker <rolf@suse.de>]
6106 *) Assembler module support for Mingw32.
6109 *) Shared library support for HPUX (in shlib/).
6110 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
6112 *) Shared library support for Solaris gcc.
6113 [Lutz Behnke <behnke@trustcenter.de>]
6115 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
6117 *) PKCS7_encrypt() was adding text MIME headers twice because they
6118 were added manually and by SMIME_crlf_copy().
6121 *) In bntest.c don't call BN_rand with zero bits argument.
6122 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
6124 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
6125 case was implemented. This caused BN_div_recp() to fail occasionally.
6128 *) Add an optional second argument to the set_label() in the perl
6129 assembly language builder. If this argument exists and is set
6130 to 1 it signals that the assembler should use a symbol whose
6131 scope is the entire file, not just the current function. This
6132 is needed with MASM which uses the format label:: for this scope.
6133 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
6135 *) Change the ASN1 types so they are typedefs by default. Before
6136 almost all types were #define'd to ASN1_STRING which was causing
6137 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
6141 *) Change names of new functions to the new get1/get0 naming
6142 convention: After 'get1', the caller owns a reference count
6143 and has to call ..._free; 'get0' returns a pointer to some
6144 data structure without incrementing reference counters.
6145 (Some of the existing 'get' functions increment a reference
6146 counter, some don't.)
6147 Similarly, 'set1' and 'add1' functions increase reference
6148 counters or duplicate objects.
6151 *) Allow for the possibility of temp RSA key generation failure:
6152 the code used to assume it always worked and crashed on failure.
6155 *) Fix potential buffer overrun problem in BIO_printf().
6156 [Ulf Möller, using public domain code by Patrick Powell; problem
6157 pointed out by David Sacerdote <das33@cornell.edu>]
6159 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
6160 RAND_egd() and RAND_status(). In the command line application,
6161 the EGD socket can be specified like a seed file using RANDFILE
6165 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
6166 Some CAs (e.g. Verisign) distribute certificates in this form.
6169 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
6170 list to exclude them. This means that no special compilation option
6171 is needed to use anonymous DH: it just needs to be included in the
6175 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
6176 EVP_MD_type. The old functionality is available in a new macro called
6177 EVP_MD_md(). Change code that uses it and update docs.
6180 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
6181 where the 'void *' argument is replaced by a function pointer argument.
6182 Previously 'void *' was abused to point to functions, which works on
6183 many platforms, but is not correct. As these functions are usually
6184 called by macros defined in OpenSSL header files, most source code
6185 should work without changes.
6188 *) <openssl/opensslconf.h> (which is created by Configure) now contains
6189 sections with information on -D... compiler switches used for
6190 compiling the library so that applications can see them. To enable
6191 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
6192 must be defined. E.g.,
6193 #define OPENSSL_ALGORITHM_DEFINES
6194 #include <openssl/opensslconf.h>
6195 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
6196 [Richard Levitte, Ulf and Bodo Möller]
6198 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
6202 *) Change the 'other' type in certificate aux info to a STACK_OF
6203 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
6204 the required ASN1 format: arbitrary types determined by an OID.
6207 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
6208 argument to 'req'. This is not because the function is newer or
6209 better than others it just uses the work 'NEW' in the certificate
6210 request header lines. Some software needs this.
6213 *) Reorganise password command line arguments: now passwords can be
6214 obtained from various sources. Delete the PEM_cb function and make
6215 it the default behaviour: i.e. if the callback is NULL and the
6216 usrdata argument is not NULL interpret it as a null terminated pass
6217 phrase. If usrdata and the callback are NULL then the pass phrase
6218 is prompted for as usual.
6221 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
6222 the support is automatically enabled. The resulting binaries will
6223 autodetect the card and use it if present.
6224 [Ben Laurie and Compaq Inc.]
6226 *) Work around for Netscape hang bug. This sends certificate request
6227 and server done in one record. Since this is perfectly legal in the
6228 SSL/TLS protocol it isn't a "bug" option and is on by default. See
6229 the bugs/SSLv3 entry for more info.
6232 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
6235 *) Add -rand argument to smime and pkcs12 applications and read/write
6239 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
6242 *) Add command line password options to the remaining applications.
6245 *) Bug fix for BN_div_recp() for numerators with an even number of
6249 *) More tests in bntest.c, and changed test_bn output.
6252 *) ./config recognizes MacOS X now.
6255 *) Bug fix for BN_div() when the first words of num and divsor are
6256 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
6259 *) Add support for various broken PKCS#8 formats, and command line
6260 options to produce them.
6263 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
6264 get temporary BIGNUMs from a BN_CTX.
6267 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
6271 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
6272 include a #define from the old name to the new. The original intent
6273 was that statically linked binaries could for example just call
6274 SSLeay_add_all_ciphers() to just add ciphers to the table and not
6275 link with digests. This never worked becayse SSLeay_add_all_digests()
6276 and SSLeay_add_all_ciphers() were in the same source file so calling
6277 one would link with the other. They are now in separate source files.
6280 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
6283 *) Use a less unusual form of the Miller-Rabin primality test (it used
6284 a binary algorithm for exponentiation integrated into the Miller-Rabin
6285 loop, our standard modexp algorithms are faster).
6288 *) Support for the EBCDIC character set completed.
6289 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
6291 *) Source code cleanups: use const where appropriate, eliminate casts,
6292 use void * instead of char * in lhash.
6295 *) Bugfix: ssl3_send_server_key_exchange was not restartable
6296 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
6297 this the server could overwrite ephemeral keys that the client
6301 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
6302 using 50 iterations of the Rabin-Miller test.
6304 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
6305 iterations of the Rabin-Miller test as required by the appendix
6306 to FIPS PUB 186[-1]) instead of DSA_is_prime.
6307 As BN_is_prime_fasttest includes trial division, DSA parameter
6308 generation becomes much faster.
6310 This implies a change for the callback functions in DSA_is_prime
6311 and DSA_generate_parameters: The callback function is called once
6312 for each positive witness in the Rabin-Miller test, not just
6313 occasionally in the inner loop; and the parameters to the
6314 callback function now provide an iteration count for the outer
6315 loop rather than for the current invocation of the inner loop.
6316 DSA_generate_parameters additionally can call the callback
6317 function with an 'iteration count' of -1, meaning that a
6318 candidate has passed the trial division test (when q is generated
6319 from an application-provided seed, trial division is skipped).
6322 *) New function BN_is_prime_fasttest that optionally does trial
6323 division before starting the Rabin-Miller test and has
6324 an additional BN_CTX * argument (whereas BN_is_prime always
6325 has to allocate at least one BN_CTX).
6326 'callback(1, -1, cb_arg)' is called when a number has passed the
6327 trial division stage.
6330 *) Fix for bug in CRL encoding. The validity dates weren't being handled
6334 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
6337 *) New function BN_pseudo_rand().
6340 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
6341 bignum version of BN_from_montgomery() with the working code from
6342 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
6346 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
6347 made it impossible to use the same SSL_SESSION data structure in
6348 SSL2 clients in multiple threads.
6351 *) The return value of RAND_load_file() no longer counts bytes obtained
6352 by stat(). RAND_load_file(..., -1) is new and uses the complete file
6353 to seed the PRNG (previously an explicit byte count was required).
6354 [Ulf Möller, Bodo Möller]
6356 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
6357 used (char *) instead of (void *) and had casts all over the place.
6360 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
6363 *) Retain source code compatibility for BN_prime_checks macro:
6364 BN_is_prime(..., BN_prime_checks, ...) now uses
6365 BN_prime_checks_for_size to determine the appropriate number of
6366 Rabin-Miller iterations.
6369 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
6370 DH_CHECK_P_NOT_SAFE_PRIME.
6371 (Check if this is true? OpenPGP calls them "strong".)
6374 *) Merge the functionality of "dh" and "gendh" programs into a new program
6375 "dhparam". The old programs are retained for now but will handle DH keys
6376 (instead of parameters) in future.
6379 *) Make the ciphers, s_server and s_client programs check the return values
6380 when a new cipher list is set.
6383 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
6384 ciphers. Before when the 56bit ciphers were enabled the sorting was
6387 The syntax for the cipher sorting has been extended to support sorting by
6388 cipher-strength (using the strength_bits hard coded in the tables).
6389 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
6391 Fix a bug in the cipher-command parser: when supplying a cipher command
6392 string with an "undefined" symbol (neither command nor alphanumeric
6393 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
6394 an error is flagged.
6396 Due to the strength-sorting extension, the code of the
6397 ssl_create_cipher_list() function was completely rearranged. I hope that
6398 the readability was also increased :-)
6399 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6401 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
6402 for the first serial number and places 2 in the serial number file. This
6403 avoids problems when the root CA is created with serial number zero and
6404 the first user certificate has the same issuer name and serial number
6408 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
6409 the new code. Add documentation for this stuff.
6412 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
6413 X509_*() to X509at_*() on the grounds that they don't handle X509
6414 structures and behave in an analagous way to the X509v3 functions:
6415 they shouldn't be called directly but wrapper functions should be used
6418 So we also now have some wrapper functions that call the X509at functions
6419 when passed certificate requests. (TO DO: similar things can be done with
6420 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
6421 things. Some of these need some d2i or i2d and print functionality
6422 because they handle more complex structures.)
6425 *) Add missing #ifndefs that caused missing symbols when building libssl
6426 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
6427 NO_RSA in ssl/s2*.c.
6428 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
6430 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
6431 has a return value which indicates the quality of the random data
6432 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
6433 error queue. New function RAND_pseudo_bytes() generates output that is
6434 guaranteed to be unique but not unpredictable. RAND_add is like
6435 RAND_seed, but takes an extra argument for an entropy estimate
6436 (RAND_seed always assumes full entropy).
6439 *) Do more iterations of Rabin-Miller probable prime test (specifically,
6440 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
6441 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
6442 in crypto/bn/bn_prime.c for the complete table). This guarantees a
6443 false-positive rate of at most 2^-80 for random input.
6446 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
6449 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
6450 in the 0.9.5 release), this returns the chain
6451 from an X509_CTX structure with a dup of the stack and all
6452 the X509 reference counts upped: so the stack will exist
6453 after X509_CTX_cleanup() has been called. Modify pkcs12.c
6456 Also make SSL_SESSION_print() print out the verify return
6460 *) Add manpage for the pkcs12 command. Also change the default
6461 behaviour so MAC iteration counts are used unless the new
6462 -nomaciter option is used. This improves file security and
6463 only older versions of MSIE (4.0 for example) need it.
6466 *) Honor the no-xxx Configure options when creating .DEF files.
6469 *) Add PKCS#10 attributes to field table: challengePassword,
6470 unstructuredName and unstructuredAddress. These are taken from
6471 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
6472 international characters are used.
6474 More changes to X509_ATTRIBUTE code: allow the setting of types
6475 based on strings. Remove the 'loc' parameter when adding
6476 attributes because these will be a SET OF encoding which is sorted
6480 *) Initial changes to the 'req' utility to allow request generation
6481 automation. This will allow an application to just generate a template
6482 file containing all the field values and have req construct the
6485 Initial support for X509_ATTRIBUTE handling. Stacks of these are
6486 used all over the place including certificate requests and PKCS#7
6487 structures. They are currently handled manually where necessary with
6488 some primitive wrappers for PKCS#7. The new functions behave in a
6489 manner analogous to the X509 extension functions: they allow
6490 attributes to be looked up by NID and added.
6492 Later something similar to the X509V3 code would be desirable to
6493 automatically handle the encoding, decoding and printing of the
6494 more complex types. The string types like challengePassword can
6495 be handled by the string table functions.
6497 Also modified the multi byte string table handling. Now there is
6498 a 'global mask' which masks out certain types. The table itself
6499 can use the flag STABLE_NO_MASK to ignore the mask setting: this
6500 is useful when for example there is only one permissible type
6501 (as in countryName) and using the mask might result in no valid
6505 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
6506 SSL_get_peer_finished to allow applications to obtain the latest
6507 Finished messages sent to the peer or expected from the peer,
6508 respectively. (SSL_get_peer_finished is usually the Finished message
6509 actually received from the peer, otherwise the protocol will be aborted.)
6511 As the Finished message are message digests of the complete handshake
6512 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
6513 be used for external authentication procedures when the authentication
6514 provided by SSL/TLS is not desired or is not enough.
6517 *) Enhanced support for Alpha Linux is added. Now ./config checks if
6518 the host supports BWX extension and if Compaq C is present on the
6519 $PATH. Just exploiting of the BWX extension results in 20-30%
6520 performance kick for some algorithms, e.g. DES and RC4 to mention
6521 a couple. Compaq C in turn generates ~20% faster code for MD5 and
6525 *) Add support for MS "fast SGC". This is arguably a violation of the
6526 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
6527 weak crypto and after checking the certificate is SGC a second one
6528 with strong crypto. MS SGC stops the first handshake after receiving
6529 the server certificate message and sends a second client hello. Since
6530 a server will typically do all the time consuming operations before
6531 expecting any further messages from the client (server key exchange
6532 is the most expensive) there is little difference between the two.
6534 To get OpenSSL to support MS SGC we have to permit a second client
6535 hello message after we have sent server done. In addition we have to
6536 reset the MAC if we do get this second client hello.
6539 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
6540 if a DER encoded private key is RSA or DSA traditional format. Changed
6541 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
6542 format DER encoded private key. Newer code should use PKCS#8 format which
6543 has the key type encoded in the ASN1 structure. Added DER private key
6544 support to pkcs8 application.
6547 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
6548 ciphersuites has been selected (as required by the SSL 3/TLS 1
6549 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
6550 is set, we interpret this as a request to violate the specification
6551 (the worst that can happen is a handshake failure, and 'correct'
6552 behaviour would result in a handshake failure anyway).
6555 *) In SSL_CTX_add_session, take into account that there might be multiple
6556 SSL_SESSION structures with the same session ID (e.g. when two threads
6557 concurrently obtain them from an external cache).
6558 The internal cache can handle only one SSL_SESSION with a given ID,
6559 so if there's a conflict, we now throw out the old one to achieve
6563 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
6564 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
6565 some routines that use cipher OIDs: some ciphers do not have OIDs
6566 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
6570 *) Simplify the trust setting structure and code. Now we just have
6571 two sequences of OIDs for trusted and rejected settings. These will
6572 typically have values the same as the extended key usage extension
6573 and any application specific purposes.
6575 The trust checking code now has a default behaviour: it will just
6576 check for an object with the same NID as the passed id. Functions can
6577 be provided to override either the default behaviour or the behaviour
6578 for a given id. SSL client, server and email already have functions
6579 in place for compatibility: they check the NID and also return "trusted"
6580 if the certificate is self signed.
6583 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
6584 traditional format into an EVP_PKEY structure.
6587 *) Add a password callback function PEM_cb() which either prompts for
6588 a password if usr_data is NULL or otherwise assumes it is a null
6589 terminated password. Allow passwords to be passed on command line
6590 environment or config files in a few more utilities.
6593 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
6594 keys. Add some short names for PKCS#8 PBE algorithms and allow them
6595 to be specified on the command line for the pkcs8 and pkcs12 utilities.
6596 Update documentation.
6599 *) Support for ASN1 "NULL" type. This could be handled before by using
6600 ASN1_TYPE but there wasn't any function that would try to read a NULL
6601 and produce an error if it couldn't. For compatibility we also have
6602 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
6603 don't allocate anything because they don't need to.
6606 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
6608 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
6610 *) Rebuild of the memory allocation routines used by OpenSSL code and
6611 possibly others as well. The purpose is to make an interface that
6612 provide hooks so anyone can build a separate set of allocation and
6613 deallocation routines to be used by OpenSSL, for example memory
6614 pool implementations, or something else, which was previously hard
6615 since Malloc(), Realloc() and Free() were defined as macros having
6616 the values malloc, realloc and free, respectively (except for Win32
6617 compilations). The same is provided for memory debugging code.
6618 OpenSSL already comes with functionality to find memory leaks, but
6619 this gives people a chance to debug other memory problems.
6621 With these changes, a new set of functions and macros have appeared:
6623 CRYPTO_set_mem_debug_functions() [F]
6624 CRYPTO_get_mem_debug_functions() [F]
6625 CRYPTO_dbg_set_options() [F]
6626 CRYPTO_dbg_get_options() [F]
6627 CRYPTO_malloc_debug_init() [M]
6629 The memory debug functions are NULL by default, unless the library
6630 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
6631 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
6632 gives the standard debugging functions that come with OpenSSL) or
6633 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
6634 provided by the library user) must be used. When the standard
6635 debugging functions are used, CRYPTO_dbg_set_options can be used to
6636 request additional information:
6637 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
6638 the CRYPTO_MDEBUG_xxx macro when compiling the library.
6640 Also, things like CRYPTO_set_mem_functions will always give the
6641 expected result (the new set of functions is used for allocation
6642 and deallocation) at all times, regardless of platform and compiler
6645 To finish it up, some functions that were never use in any other
6646 way than through macros have a new API and new semantic:
6649 CRYPTO_dbg_realloc()
6652 All macros of value have retained their old syntax.
6653 [Richard Levitte and Bodo Moeller]
6655 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
6656 ordering of SMIMECapabilities wasn't in "strength order" and there
6657 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
6661 *) Some ASN1 types with illegal zero length encoding (INTEGER,
6662 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
6663 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
6665 *) Merge in my S/MIME library for OpenSSL. This provides a simple
6666 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
6667 functionality to handle multipart/signed properly) and a utility
6668 called 'smime' to call all this stuff. This is based on code I
6669 originally wrote for Celo who have kindly allowed it to be
6670 included in OpenSSL.
6673 *) Add variants des_set_key_checked and des_set_key_unchecked of
6674 des_set_key (aka des_key_sched). Global variable des_check_key
6675 decides which of these is called by des_set_key; this way
6676 des_check_key behaves as it always did, but applications and
6677 the library itself, which was buggy for des_check_key == 1,
6678 have a cleaner way to pick the version they need.
6681 *) New function PKCS12_newpass() which changes the password of a
6685 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
6686 dynamic mix. In both cases the ids can be used as an index into the
6687 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
6688 functions so they accept a list of the field values and the
6689 application doesn't need to directly manipulate the X509_TRUST
6693 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
6697 *) Modify the way the V3 extension code looks up extensions. This now
6698 works in a similar way to the object code: we have some "standard"
6699 extensions in a static table which is searched with OBJ_bsearch()
6700 and the application can add dynamic ones if needed. The file
6701 crypto/x509v3/ext_dat.h now has the info: this file needs to be
6702 updated whenever a new extension is added to the core code and kept
6703 in ext_nid order. There is a simple program 'tabtest.c' which checks
6704 this. New extensions are not added too often so this file can readily
6705 be maintained manually.
6707 There are two big advantages in doing things this way. The extensions
6708 can be looked up immediately and no longer need to be "added" using
6709 X509V3_add_standard_extensions(): this function now does nothing.
6710 [Side note: I get *lots* of email saying the extension code doesn't
6711 work because people forget to call this function]
6712 Also no dynamic allocation is done unless new extensions are added:
6713 so if we don't add custom extensions there is no need to call
6714 X509V3_EXT_cleanup().
6717 *) Modify enc utility's salting as follows: make salting the default. Add a
6718 magic header, so unsalted files fail gracefully instead of just decrypting
6719 to garbage. This is because not salting is a big security hole, so people
6720 should be discouraged from doing it.
6723 *) Fixes and enhancements to the 'x509' utility. It allowed a message
6724 digest to be passed on the command line but it only used this
6725 parameter when signing a certificate. Modified so all relevant
6726 operations are affected by the digest parameter including the
6727 -fingerprint and -x509toreq options. Also -x509toreq choked if a
6728 DSA key was used because it didn't fix the digest.
6731 *) Initial certificate chain verify code. Currently tests the untrusted
6732 certificates for consistency with the verify purpose (which is set
6733 when the X509_STORE_CTX structure is set up) and checks the pathlength.
6735 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
6736 this is because it will reject chains with invalid extensions whereas
6737 every previous version of OpenSSL and SSLeay made no checks at all.
6739 Trust code: checks the root CA for the relevant trust settings. Trust
6740 settings have an initial value consistent with the verify purpose: e.g.
6741 if the verify purpose is for SSL client use it expects the CA to be
6742 trusted for SSL client use. However the default value can be changed to
6743 permit custom trust settings: one example of this would be to only trust
6744 certificates from a specific "secure" set of CAs.
6746 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
6747 which should be used for version portability: especially since the
6748 verify structure is likely to change more often now.
6750 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
6751 to set them. If not set then assume SSL clients will verify SSL servers
6754 Two new options to the verify program: -untrusted allows a set of
6755 untrusted certificates to be passed in and -purpose which sets the
6756 intended purpose of the certificate. If a purpose is set then the
6757 new chain verify code is used to check extension consistency.
6760 *) Support for the authority information access extension.
6763 *) Modify RSA and DSA PEM read routines to transparently handle
6764 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
6765 public keys in a format compatible with certificate
6766 SubjectPublicKeyInfo structures. Unfortunately there were already
6767 functions called *_PublicKey_* which used various odd formats so
6768 these are retained for compatibility: however the DSA variants were
6769 never in a public release so they have been deleted. Changed dsa/rsa
6770 utilities to handle the new format: note no releases ever handled public
6771 keys so we should be OK.
6773 The primary motivation for this change is to avoid the same fiasco
6774 that dogs private keys: there are several incompatible private key
6775 formats some of which are standard and some OpenSSL specific and
6776 require various evil hacks to allow partial transparent handling and
6777 even then it doesn't work with DER formats. Given the option anything
6778 other than PKCS#8 should be dumped: but the other formats have to
6779 stay in the name of compatibility.
6781 With public keys and the benefit of hindsight one standard format
6782 is used which works with EVP_PKEY, RSA or DSA structures: though
6783 it clearly returns an error if you try to read the wrong kind of key.
6785 Added a -pubkey option to the 'x509' utility to output the public key.
6786 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
6787 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
6788 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
6789 that do the same as the EVP_PKEY_assign_*() except they up the
6790 reference count of the added key (they don't "swallow" the
6794 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
6795 CRLs would fail if the file contained no certificates or no CRLs:
6796 added a new function to read in both types and return the number
6797 read: this means that if none are read it will be an error. The
6798 DER versions of the certificate and CRL reader would always fail
6799 because it isn't possible to mix certificates and CRLs in DER format
6800 without choking one or the other routine. Changed this to just read
6801 a certificate: this is the best we can do. Also modified the code
6802 in apps/verify.c to take notice of return codes: it was previously
6803 attempting to read in certificates from NULL pointers and ignoring
6804 any errors: this is one reason why the cert and CRL reader seemed
6805 to work. It doesn't check return codes from the default certificate
6806 routines: these may well fail if the certificates aren't installed.
6809 *) Code to support otherName option in GeneralName.
6812 *) First update to verify code. Change the verify utility
6813 so it warns if it is passed a self signed certificate:
6814 for consistency with the normal behaviour. X509_verify
6815 has been modified to it will now verify a self signed
6816 certificate if *exactly* the same certificate appears
6817 in the store: it was previously impossible to trust a
6818 single self signed certificate. This means that:
6819 openssl verify ss.pem
6820 now gives a warning about a self signed certificate but
6821 openssl verify -CAfile ss.pem ss.pem
6825 *) For servers, store verify_result in SSL_SESSION data structure
6826 (and add it to external session representation).
6827 This is needed when client certificate verifications fails,
6828 but an application-provided verification callback (set by
6829 SSL_CTX_set_cert_verify_callback) allows accepting the session
6830 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
6831 but returns 1): When the session is reused, we have to set
6832 ssl->verify_result to the appropriate error code to avoid
6834 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
6836 *) Fix a bug in the new PKCS#7 code: it didn't consider the
6837 case in PKCS7_dataInit() where the signed PKCS7 structure
6838 didn't contain any existing data because it was being created.
6839 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
6841 *) Add a salt to the key derivation routines in enc.c. This
6842 forms the first 8 bytes of the encrypted file. Also add a
6843 -S option to allow a salt to be input on the command line.
6846 *) New function X509_cmp(). Oddly enough there wasn't a function
6847 to compare two certificates. We do this by working out the SHA1
6848 hash and comparing that. X509_cmp() will be needed by the trust
6852 *) SSL_get1_session() is like SSL_get_session(), but increments
6853 the reference count in the SSL_SESSION returned.
6854 [Geoff Thorpe <geoff@eu.c2.net>]
6856 *) Fix for 'req': it was adding a null to request attributes.
6857 Also change the X509_LOOKUP and X509_INFO code to handle
6858 certificate auxiliary information.
6861 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
6865 *) Add the possibility to add extra information to the memory leak
6866 detecting output, to form tracebacks, showing from where each
6867 allocation was originated: CRYPTO_push_info("constant string") adds
6868 the string plus current file name and line number to a per-thread
6869 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
6870 is like calling CYRPTO_pop_info() until the stack is empty.
6871 Also updated memory leak detection code to be multi-thread-safe.
6874 *) Add options -text and -noout to pkcs7 utility and delete the
6875 encryption options which never did anything. Update docs.
6878 *) Add options to some of the utilities to allow the pass phrase
6879 to be included on either the command line (not recommended on
6880 OSes like Unix) or read from the environment. Update the
6881 manpages and fix a few bugs.
6884 *) Add a few manpages for some of the openssl commands.
6887 *) Fix the -revoke option in ca. It was freeing up memory twice,
6888 leaking and not finding already revoked certificates.
6891 *) Extensive changes to support certificate auxiliary information.
6892 This involves the use of X509_CERT_AUX structure and X509_AUX
6893 functions. An X509_AUX function such as PEM_read_X509_AUX()
6894 can still read in a certificate file in the usual way but it
6895 will also read in any additional "auxiliary information". By
6896 doing things this way a fair degree of compatibility can be
6897 retained: existing certificates can have this information added
6898 using the new 'x509' options.
6900 Current auxiliary information includes an "alias" and some trust
6901 settings. The trust settings will ultimately be used in enhanced
6902 certificate chain verification routines: currently a certificate
6903 can only be trusted if it is self signed and then it is trusted
6907 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
6908 The problem was that one of the replacement routines had not been working
6909 since SSLeay releases. For now the offending routine has been replaced
6910 with non-optimised assembler. Even so, this now gives around 95%
6911 performance improvement for 1024 bit RSA signs.
6914 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
6915 handling. Most clients have the effective key size in bits equal to
6916 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
6917 A few however don't do this and instead use the size of the decrypted key
6918 to determine the RC2 key length and the AlgorithmIdentifier to determine
6919 the effective key length. In this case the effective key length can still
6920 be 40 bits but the key length can be 168 bits for example. This is fixed
6921 by manually forcing an RC2 key into the EVP_PKEY structure because the
6922 EVP code can't currently handle unusual RC2 key sizes: it always assumes
6923 the key length and effective key length are equal.
6926 *) Add a bunch of functions that should simplify the creation of
6927 X509_NAME structures. Now you should be able to do:
6928 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
6929 and have it automatically work out the correct field type and fill in
6930 the structures. The more adventurous can try:
6931 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
6932 and it will (hopefully) work out the correct multibyte encoding.
6935 *) Change the 'req' utility to use the new field handling and multibyte
6936 copy routines. Before the DN field creation was handled in an ad hoc
6937 way in req, ca, and x509 which was rather broken and didn't support
6938 BMPStrings or UTF8Strings. Since some software doesn't implement
6939 BMPStrings or UTF8Strings yet, they can be enabled using the config file
6940 using the dirstring_type option. See the new comment in the default
6941 openssl.cnf for more info.
6944 *) Make crypto/rand/md_rand.c more robust:
6945 - Assure unique random numbers after fork().
6946 - Make sure that concurrent threads access the global counter and
6947 md serializably so that we never lose entropy in them
6948 or use exactly the same state in multiple threads.
6949 Access to the large state is not always serializable because
6950 the additional locking could be a performance killer, and
6951 md should be large enough anyway.
6954 *) New file apps/app_rand.c with commonly needed functionality
6955 for handling the random seed file.
6957 Use the random seed file in some applications that previously did not:
6959 dsaparam -genkey (which also ignored its '-rand' option),
6962 x509 (when signing).
6963 Except on systems with /dev/urandom, it is crucial to have a random
6964 seed file at least for key creation, DSA signing, and for DH exchanges;
6965 for RSA signatures we could do without one.
6967 gendh and gendsa (unlike genrsa) used to read only the first byte
6968 of each file listed in the '-rand' option. The function as previously
6969 found in genrsa is now in app_rand.c and is used by all programs
6970 that support '-rand'.
6973 *) In RAND_write_file, use mode 0600 for creating files;
6974 don't just chmod when it may be too late.
6977 *) Report an error from X509_STORE_load_locations
6978 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
6981 *) New function ASN1_mbstring_copy() this copies a string in either
6982 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
6983 into an ASN1_STRING type. A mask of permissible types is passed
6984 and it chooses the "minimal" type to use or an error if not type
6988 *) Add function equivalents to the various macros in asn1.h. The old
6989 macros are retained with an M_ prefix. Code inside the library can
6990 use the M_ macros. External code (including the openssl utility)
6991 should *NOT* in order to be "shared library friendly".
6994 *) Add various functions that can check a certificate's extensions
6995 to see if it usable for various purposes such as SSL client,
6996 server or S/MIME and CAs of these types. This is currently
6997 VERY EXPERIMENTAL but will ultimately be used for certificate chain
6998 verification. Also added a -purpose flag to x509 utility to
6999 print out all the purposes.
7002 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
7006 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
7007 for, obtain and decode and extension and obtain its critical flag.
7008 This allows all the necessary extension code to be handled in a
7009 single function call.
7012 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
7013 platforms. See crypto/rc4/rc4_enc.c for further details.
7016 *) New -noout option to asn1parse. This causes no output to be produced
7017 its main use is when combined with -strparse and -out to extract data
7018 from a file (which may not be in ASN.1 format).
7021 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
7022 when producing the local key id.
7023 [Richard Levitte <levitte@stacken.kth.se>]
7025 *) New option -dhparam in s_server. This allows a DH parameter file to be
7026 stated explicitly. If it is not stated then it tries the first server
7027 certificate file. The previous behaviour hard coded the filename
7031 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
7032 a public key to be input or output. For example:
7033 openssl rsa -in key.pem -pubout -out pubkey.pem
7034 Also added necessary DSA public key functions to handle this.
7037 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
7038 in the message. This was handled by allowing
7039 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
7040 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
7042 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
7043 to the end of the strings whereas this didn't. This would cause problems
7044 if strings read with d2i_ASN1_bytes() were later modified.
7045 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
7047 *) Fix for base64 decode bug. When a base64 bio reads only one line of
7048 data and it contains EOF it will end up returning an error. This is
7049 caused by input 46 bytes long. The cause is due to the way base64
7050 BIOs find the start of base64 encoded data. They do this by trying a
7051 trial decode on each line until they find one that works. When they
7052 do a flag is set and it starts again knowing it can pass all the
7053 data directly through the decoder. Unfortunately it doesn't reset
7054 the context it uses. This means that if EOF is reached an attempt
7055 is made to pass two EOFs through the context and this causes the
7056 resulting error. This can also cause other problems as well. As is
7057 usual with these problems it takes *ages* to find and the fix is
7058 trivial: move one line.
7059 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
7061 *) Ugly workaround to get s_client and s_server working under Windows. The
7062 old code wouldn't work because it needed to select() on sockets and the
7063 tty (for keypresses and to see if data could be written). Win32 only
7064 supports select() on sockets so we select() with a 1s timeout on the
7065 sockets and then see if any characters are waiting to be read, if none
7066 are present then we retry, we also assume we can always write data to
7067 the tty. This isn't nice because the code then blocks until we've
7068 received a complete line of data and it is effectively polling the
7069 keyboard at 1s intervals: however it's quite a bit better than not
7070 working at all :-) A dedicated Windows application might handle this
7071 with an event loop for example.
7074 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
7075 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
7076 will be called when RSA_sign() and RSA_verify() are used. This is useful
7077 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
7078 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
7079 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
7080 This necessitated the support of an extra signature type NID_md5_sha1
7081 for SSL signatures and modifications to the SSL library to use it instead
7082 of calling RSA_public_decrypt() and RSA_private_encrypt().
7085 *) Add new -verify -CAfile and -CApath options to the crl program, these
7086 will lookup a CRL issuers certificate and verify the signature in a
7087 similar way to the verify program. Tidy up the crl program so it
7088 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
7089 less strict. It will now permit CRL extensions even if it is not
7090 a V2 CRL: this will allow it to tolerate some broken CRLs.
7093 *) Initialize all non-automatic variables each time one of the openssl
7094 sub-programs is started (this is necessary as they may be started
7095 multiple times from the "OpenSSL>" prompt).
7096 [Lennart Bang, Bodo Moeller]
7098 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
7099 removing all other RSA functionality (this is what NO_RSA does). This
7100 is so (for example) those in the US can disable those operations covered
7101 by the RSA patent while allowing storage and parsing of RSA keys and RSA
7105 *) Non-copying interface to BIO pairs.
7106 (still largely untested)
7109 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
7110 ASCII string. This was handled independently in various places before.
7113 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
7114 UTF8 strings a character at a time.
7117 *) Use client_version from client hello to select the protocol
7118 (s23_srvr.c) and for RSA client key exchange verification
7119 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
7122 *) Add various utility functions to handle SPKACs, these were previously
7123 handled by poking round in the structure internals. Added new function
7124 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
7125 print, verify and generate SPKACs. Based on an original idea from
7126 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
7129 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
7132 *) Allow the config file extension section to be overwritten on the
7133 command line. Based on an original idea from Massimiliano Pala
7134 <madwolf@comune.modena.it>. The new option is called -extensions
7135 and can be applied to ca, req and x509. Also -reqexts to override
7136 the request extensions in req and -crlexts to override the crl extensions
7140 *) Add new feature to the SPKAC handling in ca. Now you can include
7141 the same field multiple times by preceding it by "XXXX." for example:
7144 this is the same syntax as used in the req config file.
7147 *) Allow certificate extensions to be added to certificate requests. These
7148 are specified in a 'req_extensions' option of the req section of the
7149 config file. They can be printed out with the -text option to req but
7150 are otherwise ignored at present.
7153 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
7154 data read consists of only the final block it would not decrypted because
7155 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
7156 A misplaced 'break' also meant the decrypted final block might not be
7157 copied until the next read.
7160 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
7161 a few extra parameters to the DH structure: these will be useful if
7162 for example we want the value of 'q' or implement X9.42 DH.
7165 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
7166 provides hooks that allow the default DSA functions or functions on a
7167 "per key" basis to be replaced. This allows hardware acceleration and
7168 hardware key storage to be handled without major modification to the
7169 library. Also added low level modexp hooks and CRYPTO_EX structure and
7170 associated functions.
7173 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
7174 as "read only": it can't be written to and the buffer it points to will
7175 not be freed. Reading from a read only BIO is much more efficient than
7176 a normal memory BIO. This was added because there are several times when
7177 an area of memory needs to be read from a BIO. The previous method was
7178 to create a memory BIO and write the data to it, this results in two
7179 copies of the data and an O(n^2) reading algorithm. There is a new
7180 function BIO_new_mem_buf() which creates a read only memory BIO from
7181 an area of memory. Also modified the PKCS#7 routines to use read only
7185 *) Bugfix: ssl23_get_client_hello did not work properly when called in
7186 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
7187 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
7188 but a retry condition occured while trying to read the rest.
7191 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
7192 NID_pkcs7_encrypted by default: this was wrong since this should almost
7193 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
7194 the encrypted data type: this is a more sensible place to put it and it
7195 allows the PKCS#12 code to be tidied up that duplicated this
7199 *) Changed obj_dat.pl script so it takes its input and output files on
7200 the command line. This should avoid shell escape redirection problems
7204 *) Initial support for certificate extension requests, these are included
7205 in things like Xenroll certificate requests. Included functions to allow
7206 extensions to be obtained and added.
7209 *) -crlf option to s_client and s_server for sending newlines as
7210 CRLF (as required by many protocols).
7213 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
7215 *) Install libRSAglue.a when OpenSSL is built with RSAref.
7216 [Ralf S. Engelschall]
7218 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
7219 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
7221 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
7225 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
7226 DH parameters/keys (q is lost during that conversion, but the resulting
7227 DH parameters contain its length).
7229 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
7230 much faster than DH_generate_parameters (which creates parameters
7231 where p = 2*q + 1), and also the smaller q makes DH computations
7232 much more efficient (160-bit exponentiation instead of 1024-bit
7233 exponentiation); so this provides a convenient way to support DHE
7234 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
7235 utter importance to use
7236 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7238 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7239 when such DH parameters are used, because otherwise small subgroup
7240 attacks may become possible!
7243 *) Avoid memory leak in i2d_DHparams.
7246 *) Allow the -k option to be used more than once in the enc program:
7247 this allows the same encrypted message to be read by multiple recipients.
7250 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
7251 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
7252 it will always use the numerical form of the OID, even if it has a short
7256 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
7257 method only got called if p,q,dmp1,dmq1,iqmp components were present,
7258 otherwise bn_mod_exp was called. In the case of hardware keys for example
7259 no private key components need be present and it might store extra data
7260 in the RSA structure, which cannot be accessed from bn_mod_exp.
7261 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
7262 private key operations.
7265 *) Added support for SPARC Linux.
7268 *) pem_password_cb function type incompatibly changed from
7269 typedef int pem_password_cb(char *buf, int size, int rwflag);
7271 ....(char *buf, int size, int rwflag, void *userdata);
7272 so that applications can pass data to their callbacks:
7273 The PEM[_ASN1]_{read,write}... functions and macros now take an
7274 additional void * argument, which is just handed through whenever
7275 the password callback is called.
7276 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
7278 New function SSL_CTX_set_default_passwd_cb_userdata.
7280 Compatibility note: As many C implementations push function arguments
7281 onto the stack in reverse order, the new library version is likely to
7282 interoperate with programs that have been compiled with the old
7283 pem_password_cb definition (PEM_whatever takes some data that
7284 happens to be on the stack as its last argument, and the callback
7285 just ignores this garbage); but there is no guarantee whatsoever that
7288 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
7289 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
7290 problems not only on Windows, but also on some Unix platforms.
7291 To avoid problematic command lines, these definitions are now in an
7292 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
7293 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
7296 *) MIPS III/IV assembler module is reimplemented.
7299 *) More DES library cleanups: remove references to srand/rand and
7300 delete an unused file.
7303 *) Add support for the the free Netwide assembler (NASM) under Win32,
7304 since not many people have MASM (ml) and it can be hard to obtain.
7305 This is currently experimental but it seems to work OK and pass all
7306 the tests. Check out INSTALL.W32 for info.
7309 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
7310 without temporary keys kept an extra copy of the server key,
7311 and connections with temporary keys did not free everything in case
7315 *) New function RSA_check_key and new openssl rsa option -check
7316 for verifying the consistency of RSA keys.
7317 [Ulf Moeller, Bodo Moeller]
7319 *) Various changes to make Win32 compile work:
7320 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
7321 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
7322 comparison" warnings.
7323 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
7326 *) Add a debugging option to PKCS#5 v2 key generation function: when
7327 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
7328 derived keys are printed to stderr.
7331 *) Copy the flags in ASN1_STRING_dup().
7332 [Roman E. Pavlov <pre@mo.msk.ru>]
7334 *) The x509 application mishandled signing requests containing DSA
7335 keys when the signing key was also DSA and the parameters didn't match.
7337 It was supposed to omit the parameters when they matched the signing key:
7338 the verifying software was then supposed to automatically use the CA's
7339 parameters if they were absent from the end user certificate.
7341 Omitting parameters is no longer recommended. The test was also
7342 the wrong way round! This was probably due to unusual behaviour in
7343 EVP_cmp_parameters() which returns 1 if the parameters match.
7344 This meant that parameters were omitted when they *didn't* match and
7345 the certificate was useless. Certificates signed with 'ca' didn't have
7347 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
7349 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
7350 The interface is as follows:
7351 Applications can use
7352 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
7353 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
7354 "off" is now the default.
7355 The library internally uses
7356 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
7357 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
7358 to disable memory-checking temporarily.
7360 Some inconsistent states that previously were possible (and were
7361 even the default) are now avoided.
7363 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
7364 with each memory chunk allocated; this is occasionally more helpful
7365 than just having a counter.
7367 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
7369 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
7373 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
7374 which largely parallels "options", but is for changing API behaviour,
7375 whereas "options" are about protocol behaviour.
7376 Initial "mode" flags are:
7378 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
7379 a single record has been written.
7380 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
7381 retries use the same buffer location.
7382 (But all of the contents must be
7386 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
7389 *) Fix problems with no-hmac etc.
7390 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
7392 *) New functions RSA_get_default_method(), RSA_set_method() and
7393 RSA_get_method(). These allows replacement of RSA_METHODs without having
7394 to mess around with the internals of an RSA structure.
7397 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
7398 Also really enable memory leak checks in openssl.c and in some
7400 [Chad C. Mulligan, Bodo Moeller]
7402 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
7403 up the length of negative integers. This has now been simplified to just
7404 store the length when it is first determined and use it later, rather
7405 than trying to keep track of where data is copied and updating it to
7407 [Steve Henson, reported by Brien Wheeler
7408 <bwheeler@authentica-security.com>]
7410 *) Add a new function PKCS7_signatureVerify. This allows the verification
7411 of a PKCS#7 signature but with the signing certificate passed to the
7412 function itself. This contrasts with PKCS7_dataVerify which assumes the
7413 certificate is present in the PKCS#7 structure. This isn't always the
7414 case: certificates can be omitted from a PKCS#7 structure and be
7415 distributed by "out of band" means (such as a certificate database).
7418 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
7419 function prototypes in pem.h, also change util/mkdef.pl to add the
7420 necessary function names.
7423 *) mk1mf.pl (used by Windows builds) did not properly read the
7424 options set by Configure in the top level Makefile, and Configure
7425 was not even able to write more than one option correctly.
7426 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
7429 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
7430 file to be loaded from a BIO or FILE pointer. The BIO version will
7431 for example allow memory BIOs to contain config info.
7434 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
7435 Whoever hopes to achieve shared-library compatibility across versions
7436 must use this, not the compile-time macro.
7437 (Exercise 0.9.4: Which is the minimum library version required by
7439 Note: All this applies only to multi-threaded programs, others don't
7443 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
7444 through a BIO pair triggered the default case, i.e.
7445 SSLerr(...,SSL_R_UNKNOWN_STATE).
7448 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
7449 can use the SSL library even if none of the specific BIOs is
7453 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
7454 for the encoded length.
7455 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
7457 *) Add initial documentation of the X509V3 functions.
7460 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
7461 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
7462 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
7463 secure PKCS#8 private key format with a high iteration count.
7466 *) Fix determination of Perl interpreter: A perl or perl5
7467 _directory_ in $PATH was also accepted as the interpreter.
7468 [Ralf S. Engelschall]
7470 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
7471 wrong with it but it was very old and did things like calling
7472 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
7476 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
7477 to use the new extension code.
7480 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
7481 with macros. This should make it easier to change their form, add extra
7482 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
7486 *) Add to configuration table a new entry that can specify an alternative
7487 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
7488 according to Mark Crispin <MRC@Panda.COM>.
7492 *) DES CBC did not update the IV. Weird.
7495 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
7496 Changing the behaviour of the former might break existing programs --
7497 where IV updating is needed, des_ncbc_encrypt can be used.
7500 *) When bntest is run from "make test" it drives bc to check its
7501 calculations, as well as internally checking them. If an internal check
7502 fails, it needs to cause bc to give a non-zero result or make test carries
7503 on without noticing the failure. Fixed.
7506 *) DES library cleanups.
7509 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
7510 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
7511 ciphers. NOTE: although the key derivation function has been verified
7512 against some published test vectors it has not been extensively tested
7513 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
7517 *) Instead of "mkdir -p", which is not fully portable, use new
7518 Perl script "util/mkdir-p.pl".
7521 *) Rewrite the way password based encryption (PBE) is handled. It used to
7522 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
7523 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
7524 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
7525 the 'parameter' field of the AlgorithmIdentifier is passed to the
7526 underlying key generation function so it must do its own ASN1 parsing.
7527 This has also changed the EVP_PBE_CipherInit() function which now has a
7528 'parameter' argument instead of literal salt and iteration count values
7529 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
7532 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
7533 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
7534 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
7535 KEY" because this clashed with PKCS#8 unencrypted string. Since this
7536 value was just used as a "magic string" and not used directly its
7537 value doesn't matter.
7540 *) Introduce some semblance of const correctness to BN. Shame C doesn't
7544 *) "linux-sparc64" configuration (ultrapenguin).
7545 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
7546 "linux-sparc" configuration.
7547 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
7549 *) config now generates no-xxx options for missing ciphers.
7552 *) Support the EBCDIC character set (work in progress).
7553 File ebcdic.c not yet included because it has a different license.
7554 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7556 *) Support BS2000/OSD-POSIX.
7557 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7559 *) Make callbacks for key generation use void * instead of char *.
7562 *) Make S/MIME samples compile (not yet tested).
7565 *) Additional typesafe stacks.
7568 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
7572 Changes between 0.9.3 and 0.9.3a [29 May 1999]
7574 *) New configuration variant "sco5-gcc".
7576 *) Updated some demos.
7577 [Sean O Riordain, Wade Scholine]
7579 *) Add missing BIO_free at exit of pkcs12 application.
7582 *) Fix memory leak in conf.c.
7585 *) Updates for Win32 to assembler version of MD5.
7588 *) Set #! path to perl in apps/der_chop to where we found it
7589 instead of using a fixed path.
7592 *) SHA library changes for irix64-mips4-cc.
7595 *) Improvements for VMS support.
7599 Changes between 0.9.2b and 0.9.3 [24 May 1999]
7601 *) Bignum library bug fix. IRIX 6 passes "make test" now!
7602 This also avoids the problems with SC4.2 and unpatched SC5.
7603 [Andy Polyakov <appro@fy.chalmers.se>]
7605 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
7606 These are required because of the typesafe stack would otherwise break
7607 existing code. If old code used a structure member which used to be STACK
7608 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
7609 sk_num or sk_value it would produce an error because the num, data members
7610 are not present in STACK_OF. Now it just produces a warning. sk_set
7611 replaces the old method of assigning a value to sk_value
7612 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
7613 that does this will no longer work (and should use sk_set instead) but
7614 this could be regarded as a "questionable" behaviour anyway.
7617 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
7618 correctly handle encrypted S/MIME data.
7621 *) Change type of various DES function arguments from des_cblock
7622 (which means, in function argument declarations, pointer to char)
7623 to des_cblock * (meaning pointer to array with 8 char elements),
7624 which allows the compiler to do more typechecking; it was like
7625 that back in SSLeay, but with lots of ugly casts.
7627 Introduce new type const_des_cblock.
7630 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
7631 problems: find RecipientInfo structure that matches recipient certificate
7632 and initialise the ASN1 structures properly based on passed cipher.
7635 *) Belatedly make the BN tests actually check the results.
7638 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
7639 to and from BNs: it was completely broken. New compilation option
7640 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
7641 key elements as negative integers.
7644 *) Reorganize and speed up MD5.
7645 [Andy Polyakov <appro@fy.chalmers.se>]
7648 [Richard Levitte <richard@levitte.org>]
7650 *) New option -out to asn1parse to allow the parsed structure to be
7651 output to a file. This is most useful when combined with the -strparse
7652 option to examine the output of things like OCTET STRINGS.
7655 *) Make SSL library a little more fool-proof by not requiring any longer
7656 that SSL_set_{accept,connect}_state be called before
7657 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
7658 in many applications because usually everything *appeared* to work as
7659 intended anyway -- now it really works as intended).
7662 *) Move openssl.cnf out of lib/.
7665 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
7666 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
7667 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
7668 [Ralf S. Engelschall]
7670 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
7671 handle PKCS#7 enveloped data properly.
7672 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
7674 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
7675 copying pointers. The cert_st handling is changed by this in
7676 various ways (and thus what used to be known as ctx->default_cert
7677 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
7678 any longer when s->cert does not give us what we need).
7679 ssl_cert_instantiate becomes obsolete by this change.
7680 As soon as we've got the new code right (possibly it already is?),
7681 we have solved a couple of bugs of the earlier code where s->cert
7682 was used as if it could not have been shared with other SSL structures.
7684 Note that using the SSL API in certain dirty ways now will result
7685 in different behaviour than observed with earlier library versions:
7686 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
7687 does not influence s as it used to.
7689 In order to clean up things more thoroughly, inside SSL_SESSION
7690 we don't use CERT any longer, but a new structure SESS_CERT
7691 that holds per-session data (if available); currently, this is
7692 the peer's certificate chain and, for clients, the server's certificate
7693 and temporary key. CERT holds only those values that can have
7694 meaningful defaults in an SSL_CTX.
7697 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
7698 from the internal representation. Various PKCS#7 fixes: remove some
7699 evil casts and set the enc_dig_alg field properly based on the signing
7703 *) Allow PKCS#12 password to be set from the command line or the
7704 environment. Let 'ca' get its config file name from the environment
7705 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
7709 *) Allow certificate policies extension to use an IA5STRING for the
7710 organization field. This is contrary to the PKIX definition but
7711 VeriSign uses it and IE5 only recognises this form. Document 'x509'
7715 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
7716 without disallowing inline assembler and the like for non-pedantic builds.
7719 *) Support Borland C++ builder.
7720 [Janez Jere <jj@void.si>, modified by Ulf Möller]
7725 *) SHA-1 cleanups and performance enhancements.
7726 [Andy Polyakov <appro@fy.chalmers.se>]
7728 *) Sparc v8plus assembler for the bignum library.
7729 [Andy Polyakov <appro@fy.chalmers.se>]
7731 *) Accept any -xxx and +xxx compiler options in Configure.
7734 *) Update HPUX configuration.
7737 *) Add missing sk_<type>_unshift() function to safestack.h
7738 [Ralf S. Engelschall]
7740 *) New function SSL_CTX_use_certificate_chain_file that sets the
7741 "extra_cert"s in addition to the certificate. (This makes sense
7742 only for "PEM" format files, as chains as a whole are not
7746 *) Support verify_depth from the SSL API.
7747 x509_vfy.c had what can be considered an off-by-one-error:
7748 Its depth (which was not part of the external interface)
7749 was actually counting the number of certificates in a chain;
7750 now it really counts the depth.
7753 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
7754 instead of X509err, which often resulted in confusing error
7755 messages since the error codes are not globally unique
7756 (e.g. an alleged error in ssl3_accept when a certificate
7757 didn't match the private key).
7759 *) New function SSL_CTX_set_session_id_context that allows to set a default
7760 value (so that you don't need SSL_set_session_id_context for each
7761 connection using the SSL_CTX).
7764 *) OAEP decoding bug fix.
7767 *) Support INSTALL_PREFIX for package builders, as proposed by
7771 *) New Configure options "threads" and "no-threads". For systems
7772 where the proper compiler options are known (currently Solaris
7773 and Linux), "threads" is the default.
7776 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
7779 *) Install various scripts to $(OPENSSLDIR)/misc, not to
7780 $(INSTALLTOP)/bin -- they shouldn't clutter directories
7781 such as /usr/local/bin.
7784 *) "make linux-shared" to build shared libraries.
7785 [Niels Poppe <niels@netbox.org>]
7787 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
7790 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
7791 extension adding in x509 utility.
7794 *) Remove NOPROTO sections and error code comments.
7797 *) Partial rewrite of the DEF file generator to now parse the ANSI
7801 *) New Configure options --prefix=DIR and --openssldir=DIR.
7804 *) Complete rewrite of the error code script(s). It is all now handled
7805 by one script at the top level which handles error code gathering,
7806 header rewriting and C source file generation. It should be much better
7807 than the old method: it now uses a modified version of Ulf's parser to
7808 read the ANSI prototypes in all header files (thus the old K&R definitions
7809 aren't needed for error creation any more) and do a better job of
7810 translating function codes into names. The old 'ASN1 error code imbedded
7811 in a comment' is no longer necessary and it doesn't use .err files which
7812 have now been deleted. Also the error code call doesn't have to appear all
7813 on one line (which resulted in some large lines...).
7816 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
7819 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
7820 0 (which usually indicates a closed connection), but continue reading.
7823 *) Fix some race conditions.
7826 *) Add support for CRL distribution points extension. Add Certificate
7827 Policies and CRL distribution points documentation.
7830 *) Move the autogenerated header file parts to crypto/opensslconf.h.
7833 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
7834 8 of keying material. Merlin has also confirmed interop with this fix
7835 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
7836 [Merlin Hughes <merlin@baltimore.ie>]
7838 *) Fix lots of warnings.
7839 [Richard Levitte <levitte@stacken.kth.se>]
7841 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
7842 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
7843 [Richard Levitte <levitte@stacken.kth.se>]
7845 *) Fix problems with sizeof(long) == 8.
7846 [Andy Polyakov <appro@fy.chalmers.se>]
7848 *) Change functions to ANSI C.
7851 *) Fix typos in error codes.
7852 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
7854 *) Remove defunct assembler files from Configure.
7857 *) SPARC v8 assembler BIGNUM implementation.
7858 [Andy Polyakov <appro@fy.chalmers.se>]
7860 *) Support for Certificate Policies extension: both print and set.
7861 Various additions to support the r2i method this uses.
7864 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
7865 return a const string when you are expecting an allocated buffer.
7868 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
7869 types DirectoryString and DisplayText.
7872 *) Add code to allow r2i extensions to access the configuration database,
7873 add an LHASH database driver and add several ctx helper functions.
7876 *) Fix an evil bug in bn_expand2() which caused various BN functions to
7877 fail when they extended the size of a BIGNUM.
7880 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
7881 support typesafe stack.
7884 *) Fix typo in SSL_[gs]et_options().
7885 [Nils Frostberg <nils@medcom.se>]
7887 *) Delete various functions and files that belonged to the (now obsolete)
7888 old X509V3 handling code.
7891 *) New Configure option "rsaref".
7894 *) Don't auto-generate pem.h.
7897 *) Introduce type-safe ASN.1 SETs.
7900 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
7901 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
7903 *) Introduce type-safe STACKs. This will almost certainly break lots of code
7904 that links with OpenSSL (well at least cause lots of warnings), but fear
7905 not: the conversion is trivial, and it eliminates loads of evil casts. A
7906 few STACKed things have been converted already. Feel free to convert more.
7907 In the fullness of time, I'll do away with the STACK type altogether.
7910 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
7911 specified in <certfile> by updating the entry in the index.txt file.
7912 This way one no longer has to edit the index.txt file manually for
7913 revoking a certificate. The -revoke option does the gory details now.
7914 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
7916 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
7917 `-text' option at all and this way the `-noout -text' combination was
7918 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
7919 [Ralf S. Engelschall]
7921 *) Make sure a corresponding plain text error message exists for the
7922 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
7923 verify callback function determined that a certificate was revoked.
7924 [Ralf S. Engelschall]
7926 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
7927 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
7928 all available cipers including rc5, which was forgotten until now.
7929 In order to let the testing shell script know which algorithms
7930 are available, a new (up to now undocumented) command
7931 "openssl list-cipher-commands" is used.
7934 *) Bugfix: s_client occasionally would sleep in select() when
7935 it should have checked SSL_pending() first.
7938 *) New functions DSA_do_sign and DSA_do_verify to provide access to
7939 the raw DSA values prior to ASN.1 encoding.
7942 *) Tweaks to Configure
7943 [Niels Poppe <niels@netbox.org>]
7945 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
7949 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
7952 *) New config option to avoid instructions that are illegal on the 80386.
7953 The default code is faster, but requires at least a 486.
7956 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
7957 SSL2_SERVER_VERSION (not used at all) macros, which are now the
7958 same as SSL2_VERSION anyway.
7961 *) New "-showcerts" option for s_client.
7964 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
7965 application. Various cleanups and fixes.
7968 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
7969 modify error routines to work internally. Add error codes and PBE init
7970 to library startup routines.
7973 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
7974 packing functions to asn1 and evp. Changed function names and error
7975 codes along the way.
7978 *) PKCS12 integration: and so it begins... First of several patches to
7979 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
7980 objects to objects.h
7983 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
7984 and display support for Thawte strong extranet extension.
7987 *) Add LinuxPPC support.
7988 [Jeff Dubrule <igor@pobox.org>]
7990 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
7991 bn_div_words in alpha.s.
7992 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
7994 *) Make sure the RSA OAEP test is skipped under -DRSAref because
7995 OAEP isn't supported when OpenSSL is built with RSAref.
7996 [Ulf Moeller <ulf@fitug.de>]
7998 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
7999 so they no longer are missing under -DNOPROTO.
8000 [Soren S. Jorvang <soren@t.dk>]
8003 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
8005 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
8006 doesn't work when the session is reused. Coming soon!
8009 *) Fix a security hole, that allows sessions to be reused in the wrong
8010 context thus bypassing client cert protection! All software that uses
8011 client certs and session caches in multiple contexts NEEDS PATCHING to
8012 allow session reuse! A fuller solution is in the works.
8013 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
8015 *) Some more source tree cleanups (removed obsolete files
8016 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
8017 permission on "config" script to be executable) and a fix for the INSTALL
8019 [Ulf Moeller <ulf@fitug.de>]
8021 *) Remove some legacy and erroneous uses of malloc, free instead of
8023 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
8025 *) Make rsa_oaep_test return non-zero on error.
8026 [Ulf Moeller <ulf@fitug.de>]
8028 *) Add support for native Solaris shared libraries. Configure
8029 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
8030 if someone would make that last step automatic.
8031 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
8033 *) ctx_size was not built with the right compiler during "make links". Fixed.
8036 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
8037 except NULL ciphers". This means the default cipher list will no longer
8038 enable NULL ciphers. They need to be specifically enabled e.g. with
8039 the string "DEFAULT:eNULL".
8042 *) Fix to RSA private encryption routines: if p < q then it would
8043 occasionally produce an invalid result. This will only happen with
8044 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
8047 *) Be less restrictive and allow also `perl util/perlpath.pl
8048 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
8049 because this way one can also use an interpreter named `perl5' (which is
8050 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
8051 installed as `perl').
8052 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8054 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
8055 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8057 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
8058 advapi32.lib to Win32 build and change the pem test comparision
8059 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
8060 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
8061 and crypto/des/ede_cbcm_enc.c.
8064 *) DES quad checksum was broken on big-endian architectures. Fixed.
8067 *) Comment out two functions in bio.h that aren't implemented. Fix up the
8068 Win32 test batch file so it (might) work again. The Win32 test batch file
8069 is horrible: I feel ill....
8072 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
8073 in e_os.h. Audit of header files to check ANSI and non ANSI
8074 sections: 10 functions were absent from non ANSI section and not exported
8075 from Windows DLLs. Fixed up libeay.num for new functions.
8078 *) Make `openssl version' output lines consistent.
8079 [Ralf S. Engelschall]
8081 *) Fix Win32 symbol export lists for BIO functions: Added
8082 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
8083 to ms/libeay{16,32}.def.
8084 [Ralf S. Engelschall]
8086 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
8087 fine under Unix and passes some trivial tests I've now added. But the
8088 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
8089 added to make sure no one expects that this stuff really works in the
8090 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
8091 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
8093 [Ralf S. Engelschall]
8095 *) Fix the generation of two part addresses in perl.
8096 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
8098 *) Add config entry for Linux on MIPS.
8099 [John Tobey <jtobey@channel1.com>]
8101 *) Make links whenever Configure is run, unless we are on Windoze.
8104 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
8105 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
8109 *) Add a useful kludge to allow package maintainers to specify compiler and
8110 other platforms details on the command line without having to patch the
8111 Configure script everytime: One now can use ``perl Configure
8112 <id>:<details>'', i.e. platform ids are allowed to have details appended
8113 to them (seperated by colons). This is treated as there would be a static
8114 pre-configured entry in Configure's %table under key <id> with value
8115 <details> and ``perl Configure <id>'' is called. So, when you want to
8116 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
8117 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
8118 now, which overrides the FreeBSD-elf entry on-the-fly.
8119 [Ralf S. Engelschall]
8121 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
8124 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
8125 on the `perl Configure ...' command line. This way one can compile
8126 OpenSSL libraries with Position Independent Code (PIC) which is needed
8127 for linking it into DSOs.
8128 [Ralf S. Engelschall]
8130 *) Remarkably, export ciphers were totally broken and no-one had noticed!
8134 *) Cleaned up the LICENSE document: The official contact for any license
8135 questions now is the OpenSSL core team under openssl-core@openssl.org.
8136 And add a paragraph about the dual-license situation to make sure people
8137 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
8138 to the OpenSSL toolkit.
8139 [Ralf S. Engelschall]
8141 *) General source tree makefile cleanups: Made `making xxx in yyy...'
8142 display consistent in the source tree and replaced `/bin/rm' by `rm'.
8143 Additonally cleaned up the `make links' target: Remove unnecessary
8144 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
8145 to speed processing and no longer clutter the display with confusing
8146 stuff. Instead only the actually done links are displayed.
8147 [Ralf S. Engelschall]
8149 *) Permit null encryption ciphersuites, used for authentication only. It used
8150 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
8151 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
8155 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
8156 signed attributes when verifying signatures (this would break them),
8157 the detached data encoding was wrong and public keys obtained using
8158 X509_get_pubkey() weren't freed.
8161 *) Add text documentation for the BUFFER functions. Also added a work around
8162 to a Win95 console bug. This was triggered by the password read stuff: the
8163 last character typed gets carried over to the next fread(). If you were
8164 generating a new cert request using 'req' for example then the last
8165 character of the passphrase would be CR which would then enter the first
8169 *) Added the new `Includes OpenSSL Cryptography Software' button as
8170 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
8171 button and can be used by applications based on OpenSSL to show the
8172 relationship to the OpenSSL project.
8173 [Ralf S. Engelschall]
8175 *) Remove confusing variables in function signatures in files
8176 ssl/ssl_lib.c and ssl/ssl.h.
8177 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8179 *) Don't install bss_file.c under PREFIX/include/
8180 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8182 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
8183 functions that return function pointers and has support for NT specific
8184 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
8185 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
8186 unsigned to signed types: this was killing the Win32 compile.
8189 *) Add new certificate file to stack functions,
8190 SSL_add_dir_cert_subjects_to_stack() and
8191 SSL_add_file_cert_subjects_to_stack(). These largely supplant
8192 SSL_load_client_CA_file(), and can be used to add multiple certs easily
8193 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
8194 This means that Apache-SSL and similar packages don't have to mess around
8195 to add as many CAs as they want to the preferred list.
8198 *) Experiment with doxygen documentation. Currently only partially applied to
8200 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
8201 openssl.doxy as the configuration file.
8204 *) Get rid of remaining C++-style comments which strict C compilers hate.
8205 [Ralf S. Engelschall, pointed out by Carlos Amengual]
8207 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
8208 compiled in by default: it has problems with large keys.
8211 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
8212 DH private keys and/or callback functions which directly correspond to
8213 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
8214 is needed for applications which have to configure certificates on a
8215 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
8217 For the RSA certificate situation is makes no difference, but
8218 for the DSA certificate situation this fixes the "no shared cipher"
8219 problem where the OpenSSL cipher selection procedure failed because the
8220 temporary keys were not overtaken from the context and the API provided
8221 no way to reconfigure them.
8222 The new functions now let applications reconfigure the stuff and they
8223 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
8224 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
8225 non-public-API function ssl_cert_instantiate() is used as a helper
8226 function and also to reduce code redundancy inside ssl_rsa.c.
8227 [Ralf S. Engelschall]
8229 *) Move s_server -dcert and -dkey options out of the undocumented feature
8230 area because they are useful for the DSA situation and should be
8231 recognized by the users.
8232 [Ralf S. Engelschall]
8234 *) Fix the cipher decision scheme for export ciphers: the export bits are
8235 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
8236 SSL_EXP_MASK. So, the original variable has to be used instead of the
8237 already masked variable.
8238 [Richard Levitte <levitte@stacken.kth.se>]
8240 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
8241 [Richard Levitte <levitte@stacken.kth.se>]
8243 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
8244 from `int' to `unsigned int' because it's a length and initialized by
8245 EVP_DigestFinal() which expects an `unsigned int *'.
8246 [Richard Levitte <levitte@stacken.kth.se>]
8248 *) Don't hard-code path to Perl interpreter on shebang line of Configure
8249 script. Instead use the usual Shell->Perl transition trick.
8250 [Ralf S. Engelschall]
8252 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
8253 (in addition to RSA certificates) to match the behaviour of `openssl dsa
8254 -noout -modulus' as it's already the case for `openssl rsa -noout
8255 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
8256 currently the public key is printed (a decision which was already done by
8257 `openssl dsa -modulus' in the past) which serves a similar purpose.
8258 Additionally the NO_RSA no longer completely removes the whole -modulus
8259 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
8261 [Ralf S. Engelschall]
8263 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
8264 BIO. See the source (crypto/evp/bio_ok.c) for more info.
8265 [Arne Ansper <arne@ats.cyber.ee>]
8267 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
8268 to be added. Now both 'req' and 'ca' can use new objects defined in the
8272 *) Add cool BIO that does syslog (or event log on NT).
8273 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
8275 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
8276 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
8277 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
8278 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
8281 *) Add preliminary config info for new extension code.
8284 *) Make RSA_NO_PADDING really use no padding.
8285 [Ulf Moeller <ulf@fitug.de>]
8287 *) Generate errors when private/public key check is done.
8290 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
8291 for some CRL extensions and new objects added.
8294 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
8295 key usage extension and fuller support for authority key id.
8298 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
8299 padding method for RSA, which is recommended for new applications in PKCS
8300 #1 v2.0 (RFC 2437, October 1998).
8301 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
8302 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
8303 against Bleichbacher's attack on RSA.
8304 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
8307 *) Updates to the new SSL compression code
8308 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8310 *) Fix so that the version number in the master secret, when passed
8311 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
8312 (because the server will not accept higher), that the version number
8313 is 0x03,0x01, not 0x03,0x00
8314 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8316 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
8317 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
8318 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
8321 *) Support for RAW extensions where an arbitrary extension can be
8322 created by including its DER encoding. See apps/openssl.cnf for
8326 *) Make sure latest Perl versions don't interpret some generated C array
8327 code as Perl array code in the crypto/err/err_genc.pl script.
8328 [Lars Weber <3weber@informatik.uni-hamburg.de>]
8330 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
8331 not many people have the assembler. Various Win32 compilation fixes and
8332 update to the INSTALL.W32 file with (hopefully) more accurate Win32
8336 *) Modify configure script 'Configure' to automatically create crypto/date.h
8337 file under Win32 and also build pem.h from pem.org. New script
8338 util/mkfiles.pl to create the MINFO file on environments that can't do a
8339 'make files': perl util/mkfiles.pl >MINFO should work.
8342 *) Major rework of DES function declarations, in the pursuit of correctness
8343 and purity. As a result, many evil casts evaporated, and some weirdness,
8344 too. You may find this causes warnings in your code. Zapping your evil
8345 casts will probably fix them. Mostly.
8348 *) Fix for a typo in asn1.h. Bug fix to object creation script
8349 obj_dat.pl. It considered a zero in an object definition to mean
8350 "end of object": none of the objects in objects.h have any zeros
8351 so it wasn't spotted.
8352 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
8354 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
8355 Masking (CBCM). In the absence of test vectors, the best I have been able
8356 to do is check that the decrypt undoes the encrypt, so far. Send me test
8357 vectors if you have them.
8360 *) Correct calculation of key length for export ciphers (too much space was
8361 allocated for null ciphers). This has not been tested!
8364 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
8365 message is now correct (it understands "crypto" and "ssl" on its
8366 command line). There is also now an "update" option. This will update
8367 the util/ssleay.num and util/libeay.num files with any new functions.
8369 perl util/mkdef.pl crypto ssl update
8370 it will update them.
8373 *) Overhauled the Perl interface (perl/*):
8374 - ported BN stuff to OpenSSL's different BN library
8375 - made the perl/ source tree CVS-aware
8376 - renamed the package from SSLeay to OpenSSL (the files still contain
8377 their history because I've copied them in the repository)
8378 - removed obsolete files (the test scripts will be replaced
8379 by better Test::Harness variants in the future)
8380 [Ralf S. Engelschall]
8382 *) First cut for a very conservative source tree cleanup:
8383 1. merge various obsolete readme texts into doc/ssleay.txt
8384 where we collect the old documents and readme texts.
8385 2. remove the first part of files where I'm already sure that we no
8386 longer need them because of three reasons: either they are just temporary
8387 files which were left by Eric or they are preserved original files where
8388 I've verified that the diff is also available in the CVS via "cvs diff
8389 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
8390 the crypto/md/ stuff).
8391 [Ralf S. Engelschall]
8393 *) More extension code. Incomplete support for subject and issuer alt
8394 name, issuer and authority key id. Change the i2v function parameters
8395 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
8396 what that's for :-) Fix to ASN1 macro which messed up
8397 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
8400 *) Preliminary support for ENUMERATED type. This is largely copied from the
8404 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
8405 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8407 *) Make sure `make rehash' target really finds the `openssl' program.
8408 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8410 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
8411 like to hear about it if this slows down other processors.
8414 *) Add CygWin32 platform information to Configure script.
8415 [Alan Batie <batie@aahz.jf.intel.com>]
8417 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
8418 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
8420 *) New program nseq to manipulate netscape certificate sequences
8423 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
8427 *) Fixes to BN code. Previously the default was to define BN_RECURSION
8428 but the BN code had some problems that would cause failures when
8429 doing certificate verification and some other functions.
8430 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8432 *) Add ASN1 and PEM code to support netscape certificate sequences.
8435 *) Add ASN1 and PEM code to support netscape certificate sequences.
8438 *) Add several PKIX and private extended key usage OIDs.
8441 *) Modify the 'ca' program to handle the new extension code. Modify
8442 openssl.cnf for new extension format, add comments.
8445 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
8446 and add a sample to openssl.cnf so req -x509 now adds appropriate
8450 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
8451 error code, add initial support to X509_print() and x509 application.
8454 *) Takes a deep breath and start addding X509 V3 extension support code. Add
8455 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
8456 stuff is currently isolated and isn't even compiled yet.
8459 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
8460 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
8461 Removed the versions check from X509 routines when loading extensions:
8462 this allows certain broken certificates that don't set the version
8463 properly to be processed.
8466 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
8467 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
8468 can still be regenerated with "make depend".
8471 *) Spelling mistake in C version of CAST-128.
8472 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
8474 *) Changes to the error generation code. The perl script err-code.pl
8475 now reads in the old error codes and retains the old numbers, only
8476 adding new ones if necessary. It also only changes the .err files if new
8477 codes are added. The makefiles have been modified to only insert errors
8478 when needed (to avoid needlessly modifying header files). This is done
8479 by only inserting errors if the .err file is newer than the auto generated
8480 C file. To rebuild all the error codes from scratch (the old behaviour)
8481 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
8482 or delete all the .err files.
8485 *) CAST-128 was incorrectly implemented for short keys. The C version has
8486 been fixed, but is untested. The assembler versions are also fixed, but
8487 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
8488 to regenerate it if needed.
8489 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
8490 Hagino <itojun@kame.net>]
8492 *) File was opened incorrectly in randfile.c.
8493 [Ulf Möller <ulf@fitug.de>]
8495 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
8496 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
8497 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
8498 al: it's just almost always a UTCTime. Note this patch adds new error
8499 codes so do a "make errors" if there are problems.
8502 *) Correct Linux 1 recognition in config.
8503 [Ulf Möller <ulf@fitug.de>]
8505 *) Remove pointless MD5 hash when using DSA keys in ca.
8506 [Anonymous <nobody@replay.com>]
8508 *) Generate an error if given an empty string as a cert directory. Also
8509 generate an error if handed NULL (previously returned 0 to indicate an
8510 error, but didn't set one).
8511 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
8513 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
8516 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
8517 parameters. This was causing a warning which killed off the Win32 compile.
8520 *) Remove C++ style comments from crypto/bn/bn_local.h.
8521 [Neil Costigan <neil.costigan@celocom.com>]
8523 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
8524 based on a text string, looking up short and long names and finally
8525 "dot" format. The "dot" format stuff didn't work. Added new function
8526 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
8527 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
8528 OID is not part of the table.
8531 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
8532 X509_LOOKUP_by_alias().
8535 *) Sort openssl functions by name.
8538 *) Get the gendsa program working (hopefully) and add it to app list. Remove
8539 encryption from sample DSA keys (in case anyone is interested the password
8543 *) Make _all_ *_free functions accept a NULL pointer.
8544 [Frans Heymans <fheymans@isaserver.be>]
8546 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
8548 [Anonymous <nobody@replay.com>]
8550 *) s_server should send the CAfile as acceptable CAs, not its own cert.
8551 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8553 *) Don't blow it for numeric -newkey arguments to apps/req.
8554 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8556 *) Temp key "for export" tests were wrong in s3_srvr.c.
8557 [Anonymous <nobody@replay.com>]
8559 *) Add prototype for temp key callback functions
8560 SSL_CTX_set_tmp_{rsa,dh}_callback().
8563 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
8564 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
8567 *) X509_name_add_entry() freed the wrong thing after an error.
8568 [Arne Ansper <arne@ats.cyber.ee>]
8570 *) rsa_eay.c would attempt to free a NULL context.
8571 [Arne Ansper <arne@ats.cyber.ee>]
8573 *) BIO_s_socket() had a broken should_retry() on Windoze.
8574 [Arne Ansper <arne@ats.cyber.ee>]
8576 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
8577 [Arne Ansper <arne@ats.cyber.ee>]
8579 *) Make sure the already existing X509_STORE->depth variable is initialized
8580 in X509_STORE_new(), but document the fact that this variable is still
8581 unused in the certificate verification process.
8582 [Ralf S. Engelschall]
8584 *) Fix the various library and apps files to free up pkeys obtained from
8585 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
8588 *) Fix reference counting in X509_PUBKEY_get(). This makes
8589 demos/maurice/example2.c work, amongst others, probably.
8590 [Steve Henson and Ben Laurie]
8592 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
8593 `openssl' and second, the shortcut symlinks for the `openssl <command>'
8594 are no longer created. This way we have a single and consistent command
8595 line interface `openssl <command>', similar to `cvs <command>'.
8596 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
8598 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
8599 BIT STRING wrapper always have zero unused bits.
8602 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
8605 *) Make the top-level INSTALL documentation easier to understand.
8608 *) Makefiles updated to exit if an error occurs in a sub-directory
8609 make (including if user presses ^C) [Paul Sutton]
8611 *) Make Montgomery context stuff explicit in RSA data structure.
8614 *) Fix build order of pem and err to allow for generated pem.h.
8617 *) Fix renumbering bug in X509_NAME_delete_entry().
8620 *) Enhanced the err-ins.pl script so it makes the error library number
8621 global and can add a library name. This is needed for external ASN1 and
8622 other error libraries.
8625 *) Fixed sk_insert which never worked properly.
8628 *) Fix ASN1 macros so they can handle indefinite length construted
8629 EXPLICIT tags. Some non standard certificates use these: they can now
8633 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
8634 into a single doc/ssleay.txt bundle. This way the information is still
8635 preserved but no longer messes up this directory. Now it's new room for
8636 the new set of documenation files.
8637 [Ralf S. Engelschall]
8639 *) SETs were incorrectly DER encoded. This was a major pain, because they
8640 shared code with SEQUENCEs, which aren't coded the same. This means that
8641 almost everything to do with SETs or SEQUENCEs has either changed name or
8642 number of arguments.
8643 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
8645 *) Fix test data to work with the above.
8648 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
8649 was already fixed by Eric for 0.9.1 it seems.
8650 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
8652 *) Autodetect FreeBSD3.
8655 *) Fix various bugs in Configure. This affects the following platforms:
8659 unixware-2.0-pentium
8663 *) Eliminate generated files from CVS. Reorder tests to regenerate files
8664 before they are needed.
8667 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
8671 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
8673 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
8674 changed SSLeay to OpenSSL in version strings.
8675 [Ralf S. Engelschall]
8677 *) Some fixups to the top-level documents.
8680 *) Fixed the nasty bug where rsaref.h was not found under compile-time
8681 because the symlink to include/ was missing.
8682 [Ralf S. Engelschall]
8684 *) Incorporated the popular no-RSA/DSA-only patches
8685 which allow to compile a RSA-free SSLeay.
8686 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
8688 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
8689 when "ssleay" is still not found.
8690 [Ralf S. Engelschall]
8692 *) Added more platforms to Configure: Cray T3E, HPUX 11,
8693 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
8695 *) Updated the README file.
8696 [Ralf S. Engelschall]
8698 *) Added various .cvsignore files in the CVS repository subdirs
8699 to make a "cvs update" really silent.
8700 [Ralf S. Engelschall]
8702 *) Recompiled the error-definition header files and added
8703 missing symbols to the Win32 linker tables.
8704 [Ralf S. Engelschall]
8706 *) Cleaned up the top-level documents;
8707 o new files: CHANGES and LICENSE
8708 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
8709 o merged COPYRIGHT into LICENSE
8710 o removed obsolete TODO file
8711 o renamed MICROSOFT to INSTALL.W32
8712 [Ralf S. Engelschall]
8714 *) Removed dummy files from the 0.9.1b source tree:
8715 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
8716 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
8717 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
8718 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
8719 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
8720 [Ralf S. Engelschall]
8722 *) Added various platform portability fixes.
8725 *) The Genesis of the OpenSSL rpject:
8726 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
8727 Young and Tim J. Hudson created while they were working for C2Net until
8729 [The OpenSSL Project]
8732 Changes between 0.9.0b and 0.9.1b [not released]
8734 *) Updated a few CA certificates under certs/
8737 *) Changed some BIGNUM api stuff.
8740 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
8741 DGUX x86, Linux Alpha, etc.
8744 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
8745 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
8749 *) Add -strparse option to asn1pars program which parses nested
8751 [Dr Stephen Henson <shenson@bigfoot.com>]
8753 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
8756 *) DSA fix for "ca" program.
8759 *) Added "-genkey" option to "dsaparam" program.
8762 *) Added RIPE MD160 (rmd160) message digest.
8765 *) Added -a (all) option to "ssleay version" command.
8768 *) Added PLATFORM define which is the id given to Configure.
8771 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
8774 *) Extended the ASN.1 parser routines.
8777 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
8780 *) Added a BN_CTX to the BN library.
8783 *) Fixed the weak key values in DES library
8786 *) Changed API in EVP library for cipher aliases.
8789 *) Added support for RC2/64bit cipher.
8792 *) Converted the lhash library to the crypto/mem.c functions.
8795 *) Added more recognized ASN.1 object ids.
8798 *) Added more RSA padding checks for SSL/TLS.
8801 *) Added BIO proxy/filter functionality.
8804 *) Added extra_certs to SSL_CTX which can be used
8805 send extra CA certificates to the client in the CA cert chain sending
8806 process. It can be configured with SSL_CTX_add_extra_chain_cert().
8809 *) Now Fortezza is denied in the authentication phase because
8810 this is key exchange mechanism is not supported by SSLeay at all.
8813 *) Additional PKCS1 checks.
8816 *) Support the string "TLSv1" for all TLS v1 ciphers.
8819 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
8820 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
8823 *) Fixed a few memory leaks.
8826 *) Fixed various code and comment typos.
8829 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
8830 bytes sent in the client random.
8831 [Edward Bishop <ebishop@spyglass.com>]
projects / FreeBSD / stable / 9.git / blob